YUMPY writes "Did you miss the panel discussion on Perl Certification at TPJ 7.0,
which ended with the audience voting strongly in favor of the development of a certification procedure for Perl programmers? Did you miss the October article called "Is it Time for Perl Certification?" in The Perl Journal?
If so, thanks to the generosity of the TPJ folks,
you can now catch up on these developments by reading the TPJ article for free.
By doing so, you'll learn how veteran educator and SPUG/Seattle.pm leader Tim Maher thinks a certification procedure could enhance the employability of JAPHs and the prestige and market share of Perl, and why it's important to have this in place by the time Perl 6 arrives."
I was surprised to read in the article that some people were actually vehemently against such an idea.
I think this is a great idea myself.
Re:Surprised
jacques on 2004-01-11T01:00:13
You can't satisfy all the people all the time. Sometimes I think having a major corporation backing Perl would reap some benefits, especially in the area of strong decision making. I also think certification is the right thing to do. Perl clearly needs this!Re:Surprised
rjbs on 2004-01-14T14:11:15
Where does Perl suffer from the lack of strong decision making? Seriously.
There's often (sometimes a bit lengthy) discussion on p5p of changes to the language, but I'm not sure I've ever seen it become a hinderance to the development of Perl.
Can you cite some examples?
I think Stonehenge's ethics are higher than that. At the moment, I don't see any purpose in creating an artificial slope for Perl Programmers to climb just to pad our bottom line. I've worked too hard over the past decade to help this community in as many free ways as I can, and get paid for the things that I have to get paid for so that I can put food on the table and pay for my net access.
Although brian doesn't speak officially for Stonehenge, I like what he said in his weblog just recently, so be sure to check it out. I especially like his take on the legal aspect with regard to liability, and anyone considering creating certifications should want to understand my particular disdain for lawyers and thus wanting to avoid any more court appearances.
Certification: an idea whose time has not come. It won't help the strong, it'll artificially promote the weak, and it will only pad the pockets of those who are selling the idea, at the risk of upsetting the industry that can already rely on experts to distinguish the good from the bad.
Re:Not surprised at pushback on certification
jacques on 2004-01-11T18:11:55
So the next time I need a doctor, I should google their reputation and not give two hoots about their degree? References and reputation go a long way, but in most respected disciplines having some form of degree/certification is essential. (Note the word 'respected') You're the exception. It's not entirely fair to the rest of us who have neither the desire nor the personality type to post thousands of times on perlmonks and moderate newsgroups!brian's argument is weak, whether you agree with him or not. He doesn't back up his points. I have never heard of someone suing Sun because a Java certified engineer screwed up on the job. Have you?
And why do you and him bring up Stonehenge so much? Do the Stonehenge experts have something to lose if certification becomes a reality?
;) Re:Not surprised at pushback on certification
merlyn on 2004-01-11T18:24:31
For a discipline like being a doctor, there is a clear formal education requirement to be able to "practice medicine", but yes, to some degree, you must rely on public reputation and personal experience to know how to pick a particular doctor.So we "certify" the formal part, and the reputation part fills in the gaps.
But when's the last time there was a formal body of education required to be a programmer in general? I've never known any two programmers who got their education the same way. Ever.
And, in Perl applications, everything is specialization. What I learned from using Perl for system administration applies very little to what I had to learn for web development. No one "certification" or even a cluster of "certifications" is going to cover that spread.
As for bad lawsuits with no merit, do I even need to bring up SCO? If a "Certified" Perl hacker released a Perl CGI app that revealed thousands of credit card numbers or federally protected medical information files, where do you think the lawyers will go? They'll look for deep pockets, including the certifying body.
I bring up Stonehenge a lot because I create policy for Stonehenge, and because brian is one of my instructors. And our reputation sells us, not a piece of paper that says we completed a test we crammed for the night before. We're full-time experts, not test-time experts.
Hmm, maybe that's my real objection to this. The activities of a person on the job are only loosely correlated with their ability to cram for a test (or pay for a "teach to the test" certification course). So by creating a "test", you're providing false data, and and an artificial industry to sell that loose correlation. Let's not make the mistake in the Perl world that Microsoft and Cisco certification have become.
Re:Not surprised at pushback on certification
RobertX on 2004-01-11T22:24:49
No. This is a rediculous statement.
"Many people may want a Perl certification, but how many people want to be the one who is legally liable for the certificant who messes up?"
The only one liable is the "certificant". Otherwise we wouldn't have any other certification process out there. There wouldn't be a CISCO, JAVA, MICROSOFT, LINUX+, A+, et al. if there were legal ramifications to the certifier.
That is a bad argument. If you don't want a certification process that is fine. Trying to throw "but we might get sued" in is fear mongering at best to quell the masses. However, I would check with those certification houses. I bet not a one has been sued.
If I were a manager, looking at resumes for someone to hire, a clear industry standard certification would definately help. Does it work all the time? Nope. But then again NOBODY expects it to.
If you think certifications are predatory scams? So be it. Those certified people are being hired because of them.
Re:Not surprised at pushback on certification
brian_d_foy on 2004-01-11T18:39:01
I bring up Stonehenge because that is a lot of my Perl experience. We do not have anything to lose because we have exactly none of the certification market now.
My argument is simply my experience as the person who actually seriously investigated doing this, seeking the advice of my attorneys and accountants. All I see anyone else doing is theorizing. I do not see anyone doing their homework.
You don't ever hear about anyone suing Sun, Microsoft, and Cisco because really big companies would lawyer the plaintiff to death. However, you do not base future risk on past performance. If you are serious about the endeavour, you have to evaluate the risk. Anyone who ignores that is asking for trouble. All I see are people ignoring the risk.Re:Not surprised at pushback on certification
yumpy on 2004-01-12T06:44:07
brian, is there any case law you can share with us that bears specifically on the subject of software testers being held liable? I've done a fair bit of homework on this subject myself, and I find the advice you're getting from your attorneys rather hard to swallow. For example, if testing agencies in general are really taking on so much liability, why wouldn't the deep-pocket Dept. of Motor Vehicles get sued every time a licensed driver causes an accident? It's simply because passing a test doesn't guarantee a particular standard of future performance. In like fashion, people will understand that a Perl programmer could earn a certification without acquiring infallability (that's "Pope-talk"; the inability to make a mistake). I think our society understands this distinction.In any case, every interesting business activity worth pursuing has some liability associated with it, but that doesn't mean that good ideas have to be dropped -- after all, this is why they make E&O insurance and various kinds of business entities that enjoy limited liability.
In other news, my homework tells me that the wildspread notion that a basic certification program is in itself wildy profitable is, guess what, a pipedream (if done conventionally). For one thing, you have to pony up a large sum in advance to open an account with a world-wide collection of testing centers, and then, at typical prices, you end up handing over 50% of testing fees to them. After figuring in several thousand person-hours of work to create the tests, there's a significant burden of expenses to overcome (assuming workers get paid -- at least eventually). And if you go whole-hog and get psychometric validation of the tests (like the big testing programs do), we're talking millions of dollars (but I think we can get the benefits of that process without the expenses).
But again, these issues don't mean it can't be done, or that we shouldn't even try. They do, however, suggest that it would be wise to do this collectively, and with corporate sponsorship, to respectively spread out the workload and secure operating expenses for the first several years.
Re:Not surprised at pushback on certification
jacques on 2004-01-12T00:10:39
Well, in my experience, one's education background doesn't mean much in this business. I don't think a certificate is essential for Perl programming to be respected. But I think certification can be very helpful, if done right. I simply don't understand all the pessimism and venom. This subject has been beaten to death, and I feel that nothing will ever happen, short of a revolt.Re:Not surprised at pushback on certification
rjbs on 2004-01-14T14:22:15
I think the best thing I've heard about certification was at TPC when someone said, "Sure it may be meaningless, but it's good to get HR to think you're not a boob."
If we could make it better than meaningless, all the better. I think it's worth thinking about. I just haven't seen any examples that I think are worth tuits.
As for fooling HR, I'll use petdance's Getting Hired tips instead and feel better about it.Re:Not surprised at pushback on certification
modred on 2004-01-12T14:58:28
I have a scope issue with your analogy. Perl is just one of the tools used by a programmer (sys admin, etc.) and certifying one tool does not really make sense to me. Continuing your analogy, does that mean you would expect your doctor to be certified in the use of a certain size and brand of scalpel?perl as "a" tool or "the" tool for the task ?
gabor on 2004-01-12T16:53:57
Look at the ads: Some say Perl knowladge is an advantage (among 30 other things that might be more important). Other say 3 years experience in Perl, mod_perl, DBI, SQL.In the former case you are right. Probaly a Perl certificate won't mean much.
In my understanding the article of Tim talks about all the others where Perl is the major tool in your craft. Or rather, he is talking about all the cases when instead of Perl they are expecting Java or C++ for the same task.Re:Not surprised at pushback on certification
pudge on 2004-01-13T00:14:45
References and reputation go a long way, but in most respected disciplines having some form of degree/certification is essential.
So you are saying that Perl programming, not having such a thing, is therefore not one of the most respected disciplines? Fine by me, I guess.
You're the exception. It's not entirely fair to the rest of us who have neither the desire nor the personality type to post thousands of times on perlmonks and moderate newsgroups!
Excuse me, but you're actually making the argument that someone who doesn't work as hard to get where they are shouldn't have an advantage in the workplace. You're arguing you should get a shortcut to a good job. Sorry, I don't buy it.
I don't agree with every point on the anti-certification side, but I do agree it doesn't help people who don't need it, and it artifically props up people who do.
I merely think it is a waste of time and money and energy. I don't think it helps, in a meaningful way, people who are trying to get jobs (as a whole) or people who are hiring.
YMMV.Re:Not surprised at pushback on certification
yumpy on 2004-01-12T19:12:21
<Randal> Certification is... usually created by those who stand the most to profit from it. <Tim>Right, and IMHO The Perl Community (and its constituent JAPHs) have more to gain from Perl Certification than anybody else; that's why I'm proposing that we do it ourselves.
<Randal> After the initial sunk cost of getting employers to believe in this artificial slope, such a corporation then gets to sit back and rake in dough based on the now artificially created demand for certifications and certification support (trainings, books, infrastructure, and so on).
<Tim>As I've detailed in a separate comment, the economics aren't really all that attractive for a Cert. Program in itself. But as you point out, and as I discussed in my TPJ article, certification would inevitably stimulate additional demand for educational services. But what's so bad about JAPHs learning more about Perl? And why should an expected increase in demand for Perl services be such an objectionable development as to cast doubt on the wisdom of developing a certification program in the first place?
After all, we haven't scrapped the laws requiring drivers to be licensed by the State just because some driving schools make a living by helping drivers acquire better skills! In fact, I'd say that's a good thing for all parties, the drivers (who are more likely to survive the trip to the mail), the instructors (who make a living doing something they're good at), and the society (which is safer as a result).
<Randal> I think Stonehenge's ethics are higher than that. At the moment, I don't see any purpose in creating an artificial slope for Perl Programmers to climb just to pad our bottom line.
<Tim>I'm glad that your corporate ethics discourage you from participating in what you view as a financially motivated scam, but I disagree that certification has to fall into that category, especially if we do it ourselves, as a community.
<Randal> I've worked too hard over the past decade to help this community in as many free ways as I can, and get paid for the things that I have to get paid for so that I can put food on the table and pay for my net access.
<Tim> Many of us work hard for the community, for free, but it's clear that NOBODY does more for JAPHdom than you do! And I not only salute you for that, but I've worked to get you some official recognition (long story...). If we do move forward on creating a certification program, we'd be bless()ed to have you lending your expertise to the project. And if things turn out well, you might even get paid for your efforts, in my conception of how it could work.
Certainly, some certification programs turn out to be practically worthless and merely "programmer taxes", but that doesn't mean that ours would have to follow suit -- for one thing, we'd have the best interests of our community at heart, unlike some other entities that might want to provide this service. Other Open-Source groups have developed their own certification programs, and none of them has a greater collection of top-class educators and writers than the Perl community, so I'm confident we could do a world-class job of it.
<Randal> Although brian doesn't speak officially for Stonehenge [stonehenge.com], I like what he said in his weblog just recently [oreillynet.com], so be sure to check it out. I especially like his take on the legal aspect with regard to liability, and anyone considering creating certifications should want to understand my particular disdain for lawyers and thus wanting to avoid any more court appearances.
<Tim> Fear of litigation is surely a powerful motivating factor in our society, but AFAIK there's no evidence that certification authorities have higher exposure to lawsuits than those operating any other modern business. (If I'm wrong about that, I invite brian to point us to some case law on the subject.) After all, you don't see the Dept. of Motor Vehicles or individual driving schools getting sued every time a driver causes an accident, on the basis that their testing procedures must have been flawed. In any case, if everybody chose to manage their lives on the principle of minimizing exposure to litigation, we'd all be home watching TV 24/7, afraid to contribute anything to society.
<Randal> Certification: an idea whose time has not come. It won't help the strong, it'll artificially promote the weak, and it will only pad the pockets of those who are selling the idea, at the risk of upsetting the industry that can already rely on experts to distinguish the good from the bad.
<Tim> Those experts you're relying on aren't doing a bit of good for the business community, who have no access to them, and who end up relying on members of their IT staffs to improvise makeshift "certification tests" of their own in attempts to qualify their potential Perl hires. Or they may get frustrated and end up gravitating toward Java programmers instead, in large part because they're much easier to certify.
It's ironic that good Java programmers are inclined to do as much of their work as possible in Perl, when nobody's looking; some unemployed SPUGsters have even resorted to getting Java certifications so they can get hired to do Perl programming in this way! I have a vision of a better world, where Perl programmers are treated like first-class citizens, given proper respect for their skills, and Perl itself is recognized by the business community as the magnificent tool that we all know it to be.
Those interested in helping other community members work toward these goals are invited to join the Perl Certification mailing list (see URL at perlcert.perlocity.org.
Re:Not surprised at pushback on certification
ziggy on 2004-01-13T05:01:34
Tim, since you started this campaign last summer, all you have offered are your personal opinions on why certification is a good thing. You have not offered any proof, nor have you have you offered a concrete proposal for Perl Certification -- just vague notions, unfounded assertions and misleading annecdotal "evidence".IMHO The Perl Community (and its constituent JAPHs) have more to gain from Perl Certification than anybody else; that's why I'm proposing that we do it ourselves.One factor you completely discount is that many of the most prolific Perl developers today are autodidacts who have never received any formal Perl training whatsoever. You portray an increased demand for Perl training services as an unqualified good thing, and refuse to see the possible downsides. Your position is understandable, since Perl trainers (like you) stand to benefit from this increased demand. That factor alone makes your support for certification quite suspect.[C]ertification would inevitably stimulate additional demand for educational services. But what's so bad about JAPHs learning more about Perl? And why should an expected increase in demand for Perl [educational] services be such an objectionable development as to cast doubt on the wisdom of developing a certification program in the first place?There are many plausible downsides to certification and an accompanying increase in demand for Perl training. First, these highly prolific autodidacts who tend to disagree with the mere idea of certification will have a huge incentive to do their work in some other language, where certification and training are not pre-requisites for finding work.
Second, ubiquitous certification and its concomitant demand for training changes the dynamic of the Perl community. Instead of writing documentation and sharing our knowledge with our peers, we have a strong financial incentive to turn our knowledge into training materials we can sell in support of a certification program. In this situation, we all lose.
Third, training is not the best way for JAPHs to hone their skills, nor is it the only way. The path to Perl mastery has always involved a healthy measure of self-sufficiency -- understanding the problem in front of you and figuring out how to solve it youself using all the tools at your disposal. Sometimes this means referring to the online documentation or the books on your shelf. Sometimes this means taking a training class or sitting in on a half-day tutorial. Sometimes this means searching through CPAN or running experiments through the debugger or Benchmark.
Focusing overmuch on training options is a disservice to Perl programmers everywhere. Training is a fine tool, but it is merely one tool among many. Our goal should be to make more self-sufficient Perl programmers, not increase training opportunities.
You completely discount the possibility that these "makeshift tests" have any value whatsoever. In the many hiring situations I've found myself in over the years, these tests are a critical part of the technical interview. They help identify which applicants possess the specific skills necessary for that shop. They help identify which applicants have the ability to learn, think on their feet and solve the kinds of problems that are likely to encounter in common situations in that shop.[B]usiness[es], [...] end up relying on members of their IT staffs to improvise makeshift "certification tests" of their own in attempts to qualify their potential Perl hires.You cite annecdotal evidence that some companies would rather rely on tools like standardized certifications than "makeshift tests" written by staff to interview programmers. That may be common in some organizations, but it is certainly not universal. I for one have never worked anywhere that relied on standardized certifications as a means to weed out resumes or rank applicants for programming jobs.
Re:Not surprised at pushback on certification
yumpy on 2004-01-13T07:01:07
<Ziggy> Tim, since you started this campaign last summer, all you have offered are your personal opinions on why certification is a good thing. You have not offered any proof, nor have you have you offered a concrete proposal for Perl Certification -- just vague notions, unfounded assertions and misleading annecdotal "evidence".<Tim>
1) To quote a famous ancient philosopher, or perhaps it was an Iron Chef, the proof is in the pudding; there would be no guarantees, as with most worthwhile endeavors. If you want guaranteed returns, you're restricted to investing in CDs! As a general rule, both in investing and other pursuits, wherever there's a possibility of a significant gain, there is also an associated degree of uncertainty about the outcome. If you want proofs, please talk to a Mathematician.
2) Why should I advance a "concrete proposal"? If/when the community decides to move in this direction, then it will be time to hammer out a plan on the basis of community input. My ideas are well documented, but that's all they are, mine; I'm just a facilitator in this effort, not a dictator.
3) Misleading anectodal evidence? As I've told you before, I have had many contacts with Hiring Managers, a few with Recruiters, and several with HR personnel (including a conversation with one at Amazon.com just last month). Also, as the current "champion" of the Perl Certification idea, many JAPHs have sought me out to tell me their stories. I'm inclined to believe what they tell me, but of course some elements might be exaggerated.
<Ziggy> One factor you completely discount is that many of the most prolific Perl developers today are autodidacts who have never received any formal Perl training whatsoever. You portray an increased demand for Perl training services as an unqualified good thing, and refuse to see the possible downsides. Your position is understandable, since Perl trainers (like you) stand to benefit from this increased demand. That factor alone makes your support for certification quite suspect.
<Tim> On the contrary, I am well aware that a great many JAPHs are self-trained, and they are among the greatest proponents of certification! That's because, unlike those with degrees in impressive tech-intensive disciplines, many of them currently lack easy ways of substantiating their Perlish knowledge. That's why I've always mentioned (e.g., in the TPC Panel Discussion and my other conference talks) the case of the guy who teaches himself Perl from the Camel, and attendance at Mongers gatherings. Unless he has a relevant college degree, he might need a certification to survive the first-level HR screening process.
Regarding your other point, that evil purveyors of training like me are just promoting certification for their own purposes, I've addressed this issue in my other writings on the subject. What it comes down to is this: I don't hear anybody accusing Larry, or Damian, or Alison, or Dan, of promoting Perl 6 just to benefit from the royalty stream, although they all stand to benefit financially from this change to Perl culture -- and I hope they are well compensated for their efforts! I'm certainly no Larry, et. al, but I submit that the principle still holds. We shouldn't refrain from taking the appropriate steps to improve our situation just to prevent somebody from possibly earning some money by providing related services.
<Ziggy>
... these highly prolific autodidacts who tend to disagree with the mere idea of certification will have a huge incentive to do their work in some other language, where certification and training are not pre-requisites for finding work. <Tim>
As you know, I'm in favor of voluntary certification, and I've even proposed that we have an official designation for those who obviously have nothing left to prove -- something like "Acknowledged Perl Guru". And IMHO most currently practicing JAPHs, especially anybody smart enough to teach himself Perl to a professional standard, wouldn't need to take a training course just to pass the certification test, if it's designed properly (although a rereading of the Camel might be in order).
<Ziggy> Second, certification and its concomitant demand for training changes the dynamic of the Perl community. Instead of writing documentation and sharing our knowledge with our peers, we have a strong financial incentive to turn our knowledge into training materials we can sell in support of a certification program. In this situation, we all lose.
<Tim> I don't see the problem. Those who want to post tutorials to PerlMonks could continue to do so. Those who are willing to undertake the (much more arduous) task of writing more comprehensive tomes, including study guides for certification tests, would get paid for their services to the community, if they were judged to be of value (which ain't a bad thing, in my book). Having the option of getting paid for doing something much more difficult than posting a 3-page tutorial isn't likely to undermine the talent pool for PerlMonks, IMHO.
Do you also worry that Alison and Dan will be unfavorably affected by the ill-gotten gains of the royalty stream from their Perl 6 book? It's good when people who do valuable work for the Perl community get rewarded for it.
<Ziggy> Third, training is not the best way for JAPHs to hone their skills, nor is it the only way. The path to Perl mastery has always involved a healthy measure of self-sufficiency -- understanding the problem in front of you and figuring out how to solve it youself using all the tools at your disposal. Sometimes this means referring to the online documentation or the books on your shelf. Sometimes this means taking a training class or sitting in on a half-day tutorial. Sometimes this means searching through CPAN or running experiments through the debugger or Benchmark. Focusing overmuch on training options is a disservice to Perl programmers everywhere.
<Tim>
I agree, and please stop doing it! You're the one obsessed with training here, not me. I've passed two (non-Perl) certification courses, on my first try, without taking any "test preparation" classes. I expect most JAPHs functioning in the workaday world would have the same experience on any of the core tests we'd devise (given some prior study time). Don't forget, as the developers of the Certification Test, we'd have the authority to engineer the tests to whatever standard we'd deem appropriate.
<Ziggy> Training is a fine tool, but it is merely one tool among many. Our goal should be to make more self-sufficient Perl programmers, not increase training opportunities.
<Tim> I quite agree! And as I've said, I feel that most JAPHs should be expected to pass the tests on the basis of the books, or on-line materials, that most of us are already familiar with. I really don't care about generating more training; I'm motivated by the intriguing challenge of designing a Perl certification program (rather like I was motivated to cobble together the first half-decent Perl beautifier), and the opportunity to lend whatever pertinent skills I might have to the effort.
<Tim> [B]usiness[es], [...] end up relying on members of their IT staffs to improvise makeshift "certification tests" of their own in attempts to qualify their potential Perl hires.
<Ziggy> You completely discount the possibility that these "makeshift tests" have any value whatsoever.
<Tim> Not so. I've always been impressed at the ingenuity of the companies that devise their own tests, and I've heard that some of them are quite good. And I want the teams that have worked so hard on those tests to share them with the Perl Certification team! But it bothers me to see individual companies burdened with the effort of testing an applicant's basic Perl knowledge, when a central authority, such as TPF::Certification, could provide this service to the business community instead -- and in so doing make it easier for businesses to hire JAPHs.
Re:Not surprised at pushback on certification
ziggy on 2004-01-13T15:07:48
I do not see how your continued histrionics about the word "proof" are the least bit meaningful here.... the proof is in the pudding ... You have offered boundless optimism at a future with certification, with no support that certification will bring about the goals you hope to realize. For example, you have failed to demonstrate that significant numbers of hiring managers are clamoring for Perl certification. You have failed to demonstrate that hiring managers who are not hiring Perl programmers today would start hiring them once certification became available. You have failed to demonstrate that JAPHs with certification will have better employment prospects than they do today without a certification program. You have also failed to demonstrate that any comparable certificate program in the industry today has realized anything close to the benefits you predict will come to us if we create our own.
No one is now, nor ever was, demanding that you prove your vision will be realized precisely as you describe it. All we are asking you to do is show your work. Yet you choose to dismiss this as a hopless pursuit of mathematical certainty.
You hide behind these word games and hold up dubious annecdotal evidence up to prove your case. You have not demonstrated that the hiring managers you describe represent anything more than the smallest fraction of hiring managers in the industry today. Furthermore, your rhetoric directly contradicts my firsthand experience and the firsthand experience of my friends and colleagues. This leads me to believe that your "data" is inaccurate at best, and intentionally misleading at worst.
Because you are advocating change.Why should I advance a "concrete proposal"?We are an open source community. You can pontificate for the next three years about the benefits of certification, or can show us the code and let the work speak for itself. Even an early alpha release (say, like a concrete proposal) would be a huge leap forward.
Re:Not surprised at pushback on certification
yumpy on 2004-01-16T20:59:51
I find the notion that any corporate entity might try to impose a certification process on Perl programmers for its own benefit, especially at this point in time, to be absurd.For one thing, it would probably take years to put such a plan into action, and any training companies willing to wait that long to see an increase in business could achieve the same result just by sitting idly by waiting for Perl 6!
Secondly, if the advent of a Perl certification program would indeed result in increased demand for Perl training, such an activity could be expected to benefit the competitors of such a company to an equal degree (because a "rising tide raises all boats"). Companies generally strive to gain an advantage over their competitors, not to make them stronger.
Although I'm not aware of any companies that are currently working toward a Perl certification program, as you know there are people asking the Perl community to consider the exploration of the pros and cons of that idea, which could lead to its acceptance, if the community at large were to approve of the idea.
I for one think the idea is worth further study.
Re:Lies, Damned Lies, Statistics...
chromatic on 2004-01-12T00:10:48
Seconded.
I don't particularly care for the "Perl won't be a serious language unless it has an industry-standard certification" histrionics, but there is a non-zero possibility that certification may be useful for hiring purposes. Andy Lester and James Duncan expressed both sides of that debate quite cogently.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-12T06:03:45
During and after OSCON, I asked Damian and others about the details of the vote, to check my recollections, and they agreed with me that the first question was "How many are against the development of a certification program", and the second, "How many are for the development of a certification program". (Damian asked the questions.) And given the context of the preceding discussion, it would have been clear that we were talking about the Perl community developing the program.All the panelists (along with Nat, who instigated the vote) were shocked at the large majority of "for" votes: approx. 100/7 (also confirmed, see details here). I understand that some people are having difficulty adjusting to this result, but we need to accept this reality and decide what to do about it! My thoughts are presented in the TPJ article.
Re:Lies, Damned Lies, Statistics...
brian_d_foy on 2004-01-12T06:30:22
People vote for all sorts of things when they get caught up in the fever of a conference. I do not have enough fingers and toes to enumerate all the good projects I was a part of at a conference, and never heard from again a month later.Re:Lies, Damned Lies, Statistics...
autrijus on 2004-01-12T07:10:17
Well, I was one of the folks who raised hands to the "for" question, but my diary notes that the question was:"How many of you are for continued discussion, and possible development, of a certification program by the Perl community?"
and I also remarked that it's hardly possible to vote against such a wording. Again, my comprehension ability for spoken English may be at fault, but I remember several people who expressed similar sentiments about the wording.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-12T07:39:13
Hi Autrijus, thanks for your contribution. I remember you were sitting in the front row, in the aisle seat, and taking lots of notes.As I indicated, I was unsure of the exact wording myself, which is why I asked others about their recollections. Anyway, I didn't remember it* quite the way your notes have it, but I'm willing to accept your version (although it's just occurred to me that Nat video-taped the whole session, so we could potentially find out for sure from an examination of that record).
Anyway, if that was the actual phrasing, that would seem to recast the vote of the 100 into both For and Maybe categories, while leaving the 7 Against voters as definitely against the idea.
So we can still conclude that most people are either still open to the idea of certification, or already in favor of it.
*NOTE: For those who weren't there, you should know how the vote came about. While Andy Lester was telling us his views on Certification, with only seconds left in the session, Nat Torkington whispered "call for a show-of-hands vote" (or words to that effect) in Damian's ear, and then Damian quickly improvised the questions and posed them to the audience. Given more lead time, a man of Damian's considerable literary talents could surely have come up with more meticulous phrasings, but I'm just thankful that Nat had the presence of mind to request a vote in the first place. 8-}
Re:Lies, Damned Lies, Statistics...
ziggy on 2004-01-13T06:37:30
No, we can't.So we can still conclude that most people are either still open to the idea of certification, or already in favor of it.First, I refute your assumption that the ~200 people in the room were a representative sample. The attendance at OSCon represents a small fraction of the community. Some would say an insignificant and completely unrepresentative fraction of the Perl community. And 200 attendees for one talk represents a minority of the attendees at OSCon. It means that there were 200 people in the audience who were either quite interested in certification, or couldn't find any other sessions that interested them at the time.
Second, you consistently and deliberately misrepresent the vote in the certification panel. The most common citation you use is found on the news story that started this discussion here on use perl: «[T]he panel discussion on Perl Certification at TP[C} 7.0, which ended with the audience voting strongly in favor of the development of a certification procedure for Perl programmers» As Autrijus points out, the vote was not in favor of certification, but something significantly more vague: continued discussion, and possible development, of a certification program.
Third, you have misrepresented the meaning of the "vote". Usually, you claim that about half of the audience voted in favor of certification. Now, you seem to recall that it was about half of the audience "voting" either yes or maybe -- effectively "not voting no" -- to "continued discussion". Sorry, but this is hardly a strong vote in favor of anything.
Fourth, you repeatedly neglect to mention how the panel voted. Care to ellucidate that point? I vaguely remember you telling me privately that about five of the seven panelists all voted against this weak, non-binding resolution. The two who voted for more discussion were you and Damian Conway. Among the five who voted against it were at least two managers who actively hire Perl programmers, possibly more.
Finally, as both schwern and brian pointed out, straw polls at a conference don't mean anything significant (except possibly that Nat should schedule a similar session next year). If, as you say, «most people are either still open to the idea of certification, or already in favor of it», then there shouldn't be anywhere near as much pushback as there has been in every venue where you've raised the issue.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-13T08:04:16
I'm shocked that you would defame me in a public forum, but just to show what a reasonable guy I am, I will nonetheless address the issues you've raised.1) You have no better insight into the demographics of the 200 people who attended the TPC Panel Discussion than I do. So my guess is as good as yours, and we may both still be wrong.
2) I've always taken care to seek additional input regarding my impressions of what happened at the conference from others who were present, and I've always done my best to present the results as accurately as I could. If these efforts did indeed result in an inaccurate statement of the exact phrasing of the questions, as the recent input from Autrijus suggests, I'm as surprised as anyone else, and I humbly apologize.
(To set the scene, I was engaged in a long-distance dialogue with Andy Lester, from a distance of over 100 ft, with my ears straining to hear him, when Damian started calling out the voting questions. I was more concerned about rapidly counting the raised hands than in memorizing the questions, so I asked others to help me reconstruct the phrasings of the questions I reported.
3) There were only two questions posed to the audience, "for" and "against". So, given the recent evidence of the fuzzy phrasing, 7 voted "against" continued discussion and possible development of a certification program, and the 100 that voted "for" should apparently be recharacterized as voting either "for" (meaning for development) or "maybe" (meaning for continued discussion). FWIW, I made this correction in a separate comment on this board immediately after I saw the post from Autrijus.
4) Regarding the votes of the panelists, we were a group of 5, and 2 were pro, and the other three undecided or against.
5) You're arguing that the opposition on use.perl.org to certification is inconsistent with the 100 voting "for/maybe" at the Panel discussion. I don't quite see your logic, unless you're arguing that the posters are all fibbing about their true feelings 9-}, but as I've indicated before my take is that although a few loquacious gurus are against the idea, the majority of JAPHs seem to be either open to the idea or for it. After all, 100/200 at the Panel Discussion either wanted to continue exploring the idea or were already in favor of proceeding. Unless you can convince me they were mostly Python programmers playing a joke on us 8-}, I'm inclined to take that evidence as the best indication we've got about how our community feels on the subject.
On a personal note, if indeed I've mischaracterized the results of the vote, as seems likely given the recent clarification of the questions, I'm truly sorry. It's clear that some of certifications foes are trying to paint me as a con-man who's trying to hoodwink innocent JAPHs into burdening themselves with what amounts to a Programmer tax, for my own selfish purposes -- and it irks me that this development could be taken as evidence supporting that position. I can understand their interest in defending against such "attacks" on the community, but believe me, I'm not that kind of guy. And I've got about 400 fans in SPUGdom who can vouch for my character -- how many of you mudslingers can match that? I rest my case. 8-}
Re:Lies, Damned Lies, Statistics...
ziggy on 2004-01-13T14:31:48
No, I am calling you accountable for the misleading statements you have made repeatedly about the "vote" certification panel at OSCon. Apparently, you agree that your representations may have been less than forthright:I'm shocked that you would defame me in a public forum [...]On the issue of the panelists:[I]f indeed I've mischaracterized the results of the vote, as seems likely given the recent clarification of the questions, I'm truly sorry.You have said repeatedly that the two panelists who voted "for" (or rather voted to not vote against) your "vote" were you and Damian. Do you recall who the other panelists were?[W]e were a group of 5, and 2 were pro, and the other three undecided or against.I'm arguing no such thing. I'm arguing that the consistent and repeated posting activity (here, on the perlcert wiki, and elsewhere) of those who disagree with you are inconsistent with your "conclusion" from your non-binding straw poll that most people are either still open to the idea of certification, or already in favor.You're arguing that the opposition on use.perl.org to certification is inconsistent with the 100 voting "for/maybe" at the Panel discussion.You have also extrapolated the voice of those 100 in your article on TPJ, where you express similar ill-founded conclusions:
No, it doesn't. At best, it's the most specific evidence that some people are in favor of exploring certification -- no more, no less. It is not now nor was it ever the mandate to "make certification a reality" as you continually propound.This vote provides the most specific and credible evidence we've ever had of community opinions on this topic, and it tells us that JAPHs are clearly in favor of making certification a reality.Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-13T15:37:52
Ziggy,You keep alluding to a presumed preponderance of negative comments on this board as evidence that Certification isn't as palatable to JAPHs as the OSCON vote suggests, but you need to take a closer look at the data.
By my reckoning, only 14 of the many millions of software professionals within easy reach of a keyboard have seen fit to weigh in on this topic.
And of these, only Ziggy, foy, and merlyn have come down in opposition to certification. Jacques, RobertX, yumpy, gabor, and chromatic seem favorably disposed, and modred, pudge, autrijus, ethan, simon, and krellis seem to be undecided. Surprised? That's because a very few individuals are making all the noise here -- lately, you and me.
That's 3 against, 5 for, and 6 undecided. Hardly the compelling evidence you seem to think it is for the idea being unpopular. In contrast, of 200 people who cared enough to show up at the TPC Panel Discussion, 7 were against the idea, and 100 undecided or in favor or it, according to our latest theory of what the questions actually signified. (I'm hoping to get a definitive answer to that question soon, from an analysis of a tape of the session.)
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-13T16:47:26
I am a lot closer to No than I am to Yes or Maybe. If I had to pick one, it would be No. I just dislike coming to firm conclusions; I prefer to leave the door open unless there is a reason to close it, and I personally have no reason to close it. But if you must count my vote, it should be a No.Some other thoughts I've had as I've read through more comments:
- I think Sun is right-on to say they do not certify that the certification actually means anything. It's true. You cannot say with any real certainty, for the majority of the people who pass your test, that they are qualified for anything. You can only say they passed your test. I just think this piece of paper is meaningless.
- Yes, it would help some people get jobs, perhaps. But some of the people it will help get jobs probably shouldn't get those jobs, and it hurts the employers who trust in the certification, and the other job seekers who were more qualified but unreasonably disregarded.
- Not to raise a stink, but you quoted the figures supposedly in favor of exploring certification at TPC as evidence that the community was in favor of it, but when presented with counterevidence in the form of detractors here and in other forums, you say that only a small percentage of the millions of developers have expressed an opinion. We can't have it both ways. You're right: we don't know how most people feel about it, and we shouldn't assume any sigfnificant level of support for either side from this discussion, or from the TPC discussion.
(Polling and votes is sortof an interest of mine, so I just had to weigh in there.
:-) - At risk of fanning the flames, but in hopes of preventing more, I think both sides have cast aspersions on the other, and as the Mom of this forum, I'm asking everyone to be more careful. This is especially true, in my judgment, in the case of presuming motivations of someone else. None of us know the hearts and minds of others, and it is fruitless to pretend that we do.
Though this message is in reply to Tim, I am not directing this particular statement at him, but to everyone. Don't make me turn this web site around!
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-13T23:25:16
Pudge,Thanks for clarifying you position as NO; it doesn't change the overal picture too much though, leaving the results as
FOR: 5, AGAINST: 4, and UNDECIDED: 5.
In contrast, the OSCON Vote shows a much greater proportion of individuals being in the composite "undecided/for" certification category (according to the latest interpretation of the results), versus those against -- 100:7.
Although this vote was not drawn from a scientifically derived sample of the population of Perlers, its still the best evidence we've got so far, being based on an N of 200, rather than the measly 14 creating all the ruckus on this board. And it's a lot easier for voters to particicpate in this "straw poll" than it was for them to visit OSCON, making one wonder why the numbers are so small.
Regarding the meaning of a Certificate; although Sun (et al) do indeed downplay the significance of the ones they issue, we all know this is just a game their lawyers are making them play. Surely they (and others) truly do believe that they are measuring something worthwhile, otherwise they wouldn't spend the tens of millions of dollars it takes to have their tests vetted through exotic psychometric procedures (as I *think* they still do), and those on the hiring side of the equation wouldn't give them any credibility.
It's just like the ace house inspector I hired to check out a house I ended up buying -- he made me agree in writing that his findings were only opinons and not necessarily indicative of the true condition of the house, and that in the event of a lawsuit his liability would be limited to the price of the inspection, $100! The sad truth is that many competent professionals are so terrified of being financially ruined by litigation that they no longer even represent their work products as being relevant to the problem you've hired them to solve! 8-{
Nonetheless, he *did* provide valuable insights about the house based on his expert experience, and I was glad to have had the benefits of his professional service. In like fashion, (some/all?) testing authorities might not guarantee that their tests actually measure anything useful, but that's only legal maneuvering because the big ones do spend multi-millions to ensure that they do precisely that.
So in summary, a Perl certification is on the face of it "meaningless", at least to a would-be plaintiff, but on the other hand, it's at least as meaningful to a Hiring Manager as the Home Examiner's Inspection Report is to a consumer.
Pudge, you commented on my comments on sample sizes and what they tell us about the underlying populations. My point was this: if indeed there is truly a mass population of JAPHs out there who are strongly against certification, as its detractors like to tell us, why have we only heard from 4 of them on this board? It's not as if they had to fly across the country and pony up a lot of dough to make their feelings known, as the OSCON audience did. I'm *not* trying to have it both ways, only one way: 200/X is a much bigger proportion than 14/X is, so we should give greater consideration to the OSCON votes than those of this forum.
I wholeheartedly endorse your call for greater civility, and apologize if you have found any of my comments to have gone over the line. You should know I've taken action on my own to make this discussion more civilized, by communicating a "Cease and Desist" message earlier today to the individual who has been libelling me in this forum.
Once again folks, for the record, I truly thought I was representing the results of the OSCON vote accurately in all my communications about it; once Autrijus presented evidence that the questions were different than I had thought, I immediately posted a correction and apology.
Nat tells me that he'll review the videotape tonight, so we'll have the definitive ruling on the exact questions posed at the OSCON vote soon. At that time, I'll post a message in reply to my original posting, so people can easily find it, that reinterprets the results for compliance with the new findings (if necessary).
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-14T06:34:20
I'm *not* trying to have it both ways, only one way: 200/X is a much bigger proportion than 14/X is, so we should give greater consideration to the OSCON votes than those of this forum.
Statistics don't work that way. If you are trying to find out what X thinks as a whole, then you must do it properly. The most common method is random sampling. The simple fact is that the conference's 200/X is not more valid than the 14/X here. Both are equally invalid as a measure of X. The fact that 200 > 14 isn't relevant, statistically speaking, because the 200 is not a sufficiently representative sample of X to be meaningful.
And actually, if the 14 were a random sample of X, it could be far more valid than the 200. Of course, it is not a random sample. And yes, larger samples are better, but only if the sampling method produces a representative sample.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-14T06:50:15
In an imperfect world, sometimes you have to make do with the hand reality deals you. Although I admire your respect for the concept of statistical significance, it's meant to be a guideline for accepting or rejecting hypotheses, not an iron-clad rule that prevents sensible consideration of imperfectly collected data.I maintain that if you've got two non-random samples from the same population, and you're interested in coming up with a provisional best guess about the characteristics of that population, you've got to go with the sample of 200 over one of 14 (a seven-fold difference!), regardless of the size of the target population.
Then when time and resources permit, by all means do a proper survey, get more reliable data, and derive sound conclusions. 8-}
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-14T07:33:30
Although I admire your respect for the concept of statistical significance, it's meant to be a guideline for accepting or rejecting hypotheses, not an iron-clad rule that prevents sensible consideration of imperfectly collected data.
No. In fact, the 200 people ARE NOT MORE VALID than the 14, as a represenative sample of X. They are not. They cannot be.
I maintain that if you've got two non-random samples from the same population, and you're interested in coming up with a provisional best guess about the characteristics of that population, you've got to go with the sample of 200 over one of 14 (a seven-fold difference!), regardless of the size of the target population.
And you would be entirely wrong. The people who went to the conference were more predisposed to certification than is the rest of the population X. Your math is completely and utterly flawed.Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-14T08:00:17
The people who went to the conference were more predisposed to certification than is the rest of the population X.Maybe -- maybe not. I don't mind you speculating, but it would seem to me most prudent to assume that they were only "more interested", not "more for" or "more against", than the target population. After all, the only thing we know for sure about them, apart from the results of their votes, is that they showed up.
Your math is completely and utterly flawed.
You're quite right about that; I meant a sample of 100, not 200 (hence the 7-fold difference re:14 remark). 8-}
I guess neither of us is going to change each other's mind on this point. Thanks for playing!
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-14T17:49:44
Maybe -- maybe not. I don't mind you speculating
I was not speculating, merely giving an example. Because you don't know what their predisposition is, you cannot say they are "more" representative just because their number is larger.
Imagine a circle that represents the population X. It is divided into sections, like a pizza, that represent the different views of the population. For all you know, the 200 are all from one or two of the many sections, and the 14 are evenly distributed. If that were the case, the 14 would be MORE representative of the X than the 200.
Or 100. It doesn't matter.
The sampling method means everything. The size of the sample means very little, and in this case, nothing at all. This is not mere theory, this is actually how it works in practical measures. As you can see from the circle visualization, without a proper sampling method, your 100 cannot mean anything interesting about X, but with a proper sampling method, the 14 can mean a great deal.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-19T01:59:17
Sure, random sampling is preferred to "haphazard" sampling, or certainly none at all. But life doesn't always give you those choices. Since we started this sub-thread, I've been looking for a way to illustrate my thoughts on this subject, and I think I've got one. Let me know what you think!Imagine that you and your colleagues have to visit a tropical island next week. It's populated by two kinds of deadly snakes, red ones and blue ones.
Each of you is allowed to take only one bottle of snake-bite antidote with you, so you've each got to pick a color in advance. (You won't meet up with other humans on the island, so "collaborative strategies" in picking antidotes won't work.)
Your colleague, Mr. A, has no other information about the island, so he flips a coin to choose the color of his antivenin.
Mr. B heard from Mr. X that there were lots more blue than red snakes there last week, so Mr. B chooses the blue stuff.
Mr. C also heard Mr. X's report, but he likes contrarian bets and the "law of averages" (aka the "gambler's fallacy"), so he chooses red.
Mr. D, fancying himself a statistician, boldly asserts that "nobody's done a proper random sampling of the population", so he rejects Mr. X's observations, flips a coin, and chooses accordingly.
What should you do, if your intent is to live -- as opposed to defending your actions in a Statistical journal?
My position is that you always pay attention to your data. If the collection procedures were imperfect, you strive to refine them and beef up the testimony of your data. But if you can't expend that extra time or effort before making an important decision, "flipping a coin" is an insult to the reality of your existing observations. I for one would be taking the blue stuff to the island!
The moral of this story is that the field of Statistical Analysis provides scientifically sound procedures for inferring the underlying characteristics of populations -- to a measurable degree of certainty. But its procedures should not be construed as the sole criterion for determining the worthlessness or credibility of data.
For example, in scientific research, it's commonly the case that adding a few more subjects to an experiment can suddenly cause the results to cross the threshold required for a particular (somewhat arbitrary) level of statistical significance (usually ,
.05 or .01). This doesn't mean that the body of data that consistently showed the same pattern suddenly went from meaningless to priceless when those additional subjects were added; it just means that you weren't in a position to offer consensually acceptable proof of their value until then. Having a "proof" is important, but it shouldn't be construed as the sole criterion for paying heed to what's happening in the world.
I'm packing the blue stuff to the island -- I can't prove that my decision is the best one, but I'm satisfied that it is, given the circumstances.
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-19T02:23:53
What should you do, if your intent is to live -- as opposed to defending your actions in a Statistical journal?
So you are saying, even if you don't have real data, sometimes you are required to guess anyway. Fine. But we are not required to guess in this case, so it is fruitless to guess, and either way, it is still incorrect to say one sample is more representative than the other.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-19T03:08:04
I essentially agree with you, but where I differ is in thinking we might never get better evidence of what the JAPHly population thinks about anything, than the hints we've already got about their views on Perl Certification. And the incomplete, tentative observations we've collected so far indicate that they're interested in exploring Certification further, and possibly "making it so". I think we should give serious consideration to how to proceed from here on the basis of this new information.By the way, how could we approach the problem of surveying the population of Perl programmers? For starters, how would be decide whom to include or exclude? It's not as if we have a "JAPH registry" we can consult. This is a non-trivial problem -- and my guess is we'd eventually end up punting by designating the TPC attendees as our sample anyway.
Perhaps we could attach some survey questions to the OSCON/TPC registration form this year . . .
Re:Lies, Damned Lies, Statistics...
pudge on 2004-01-19T03:38:42
I essentially agree with you, but where I differ is in thinking we might never get better evidence of what the JAPHly population thinks about anything, than the hints we've already got about their views on Perl Certification.
I don't think I ever implied a proper survey is impossible. It just hasn't been done. We don't know what they think.
By the way, how could we approach the problem of surveying the population of Perl programmers? For starters, how would be decide whom to include or exclude? It's not as if we have a "JAPH registry" we can consult. This is a non-trivial problem -- and my guess is we'd eventually end up punting by designating the TPC attendees as our sample anyway.
Well yes. You would need to define the population, and then identify the members of it. For "general population" we cheat and pick people who have numbers listed in the phone book. The best way I can think of for Perl would be something like "people who have accounts on use.perl.org or PerlMonks, or subscribe to one of the perl.org mailing lists." This is imperfect for a few reasons, such as:
* Not everyone interested in these things uses these resources
* Not everyone who uses these resources has accounts or subscriptions
And then there are logistical problems:
* Do we have the right to contact people via these lists? If so, would we wish to infringe on their privacy?
* What about people with mutiple accounts?
I think it can be done, though I am not sure it should be done, and the answers would necessarily be representative only of this subset of the population, and it would be up to interpretation on whether or not the subset is representative of the population.
Perhaps we could attach some survey questions to the OSCON/TPC registration form this year . . .
I greatly dislike using conferences for such things. The sample is skewed in so many ways (geographical, financial, cultural, etc.). Mailing lists and web sites are skewed too (especially in regard to language, and also in culture, as many people simply don't wish to be a part of such resources), but I think far less so than conferences are.
Of course, conference attendees would be much easier, but I think the results would be much less interesting.Re:Lies, Damned Lies, Statistics...
delegatrix on 2004-01-13T17:19:03
It's a bit hard to come down on either side. The debate has quickly become 'The sky is falling" vs. "How can we live without this". For me, it comes down to risks. What are the risks, positive and negative, what is the need, and does this ease or cause any pain. Personally, I favor certification programs like PMI that require work experience rather than simply tests, and build a professional society that is recognized and respected. I have not yet reached conclusions about the value of a perl cert, and cannot do so in this kind of environment.Re:Lies, Damned Lies, Statistics...
RobertX on 2004-01-14T02:11:18
I look at it this way. Is it "necessary" for Perl to conitnue? Nope. Is it "desirable" to have a certification process that is more controlled by Perl people with funds from such going back to Perl people? Yup. It is really that black and white for me.
Re:Lies, Damned Lies, Statistics...
rob_au on 2004-01-14T13:08:26
That's 3 against, 5 for, and 6 undecided. Hardly the compelling evidence you seem to think it is for the idea being unpopular. In contrast, of 200 people who cared enough to show up at the TPC Panel Discussion, 7 were against the idea, and 100 undecided or in favor or it, according to our latest theory of what the questions actually signified.I am quite certain that there are a number of others, such as myself, who are simply watching this discussion with interest, rathering than inflaming with the discussion further with posts furthering one side of the argument or the other. The lack of comments in this thread doesn't indicate apathy on the part of myself or others.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-15T05:20:16
I'm sure there are plenty of read-only participants to this discussion, but of course, it's only the active participants we are able to classify into the various camps.That's the downside of being a lurker, I guess.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-14T04:05:20
You need to look up "forthright" in your dictionary. I've never said that my representations were less than "candid" or "straightforward"; on the contrary, I did my very best to correctly report on the events that transpired at the OSCON discussion, and I took pains to verify my facts in advance. However, despite these measures, it seems that my informants and I may have gotten the phrasings of the questions wrong, which may cast the results in a different light. Once we get the "audio recording" of the session to confirm the actual phrasing, we'll know more.Re:Lies, Damned Lies, Statistics...
ziggy on 2004-01-13T14:34:23
brian has said it before, and I will say it again:And I've got about 400 fans in SPUGdom who can vouch for my character -- how many of you mudslingers can match that? I rest my case. 8-}Please keep the ad hominem attacks off the table. Let's keep this discussion civil and focus on the merits.
Re:Lies, Damned Lies, Statistics...
yumpy on 2004-01-13T14:43:11
You've got a lot of nerve lecturing me, who you continue to defame, about civilized behaviour in this forum. Talk about calling the kettle black! And please note "smileys" in postings -- they're intended to mark attempts at levity. As if you didn't know . . .Fuck the polls, give us something real.
schwern on 2004-01-13T21:55:54
"How many of you are for continued discussion, and possible development, of a certification program by the Perl community?"Thanks, Autrijus. That's about the wording I remember. That question is so weak, you really can't vote against it. About the only thing I'd conclude from the poll is that people aren't so rabidly against the idea that they think no further work should be done. Not exactly a rousing chourus of consent.
But the poll was just a very minor part of the show, and I'd be the first person to say that you do the work first, look for consensus later. We all know that the best projects are the ones that plow ahead to get something concrete done. The worst are the ones that have nothing but vapor and plans and grope around for approval. Fuck the polls. Write up something concrete and then you'll see a real response.
The real meat of the discussion was in what form such a certification should take and there the real problem is revealed. Most folks are horrified at the thought of some sort of multiple-choice Brainbench/MSIE/A+ style cert because it doesn't really test how good a programmer you are. The other revealed problem was the fear of a "one size fits all" test. That any cert must recognize that there's different styles of Perl programmers and administer different tests. Sysadmins, testers, big app/OO, web, GUI, etc.
Most folks have had bad experiences with the existing certifications. The most important problem in putting together a cert program is convincing us that its a credible measure of how good a programmer one is. If you can do that, then people will accept it. You can go on all day about how it will benefit businesses, generate jobs, blah blah... but the bottom line is this: do I want to be working with someone that was hired based on their cert?
But before anyone will really accept it, you have to have something real. Vapor convinces no one. Concentrate on generating a real certification program. You don't need a huge group to do this. Like most other Open Source projects, most of the work will be done by a small group of people who believe it will work. Give us something concrete, else we'll just wank all day about it posting messages like this.
Re:Fuck the polls, give us something real.
yumpy on 2004-01-14T03:43:55
Schwern,I sympathize with your interest in "seeing the code" for the vapor-ware Perl Certification program everybody's talking about, rather than enduring additional discourse on this board. FWIW, my thoughts on how to proceed are as follows:
That's my rough plan at the moment. Patches welcome!
- get people talking about the pros/cons of the Perl community developing its own certification program (DONE!)
- ask TPF to designate a Certification Team Leader, who will assemble a small team to do a preliminary study on the feasibility of the project
- conduct a TPF-sanctioned survey of HR depts. and Hiring Managers to learn if they would
- be interested in having a community-sanctioned certification for Perl programmers
- be willing to support that program with donations of money
- so kind as to share the details of their home-grown Perl cert programs with the project team
- conduct a survey of the Perl community to determine how many members are in favor of creating a voluntary certification program for Perl programmers
- on the basis of corporate donation prospects, and member interest, decide whether to proceed
- if proceeding, get to work on developing the best certification program ever created for Perl, using a far-flung team of electronically connected volunteers
So as you can see, this is not the type of project that can be judged by "looking at the code". There are important cultural and political factors to consider, and there would be a significant amount of work involved in putting together even the first test (of an eventual series). Given these realities, IMHO it's unwise to invest energy in a prototype test unless it's already known that the project has support from the Perl and business communities.
Sure, one could conceivably just whip up a test, wow the crowds with it, and make it a de facto standard -- maybe. But IMHO that "cowboy" model is better suited to entreprenurial activity than it is to efforts associated with the "Perl community", where the potential impact of decisions of wide-ranging influence on millions of people has to be a paramount consideration.
JFDI!
chromatic on 2004-01-14T06:24:03
Why do you need TPF's blessing to come up with a proposal for a certification program? If it's good, I think they'll recognize it. Likewise, why do you need the approval of "the Perl community"? If it's good, again, I think they'll recognize it.
One persistent myth about volunteer projects is that having a good idea is enough. It's not. Having an idea doesn't get the work done. Having an idea doesn't entitle you to be a leader. You have to do the work yourself or convince enough other people to do the work.
Either way, someone has to do the work. There aren't enough warm bodies looking for something to do that TPF can say "Go" and hordes of willing volunteers jump in whatever direction the appointed leader points. If that were the case, we'd have made more progress on the Perl 6 compiler, which is more important than certification.
Heck, if there were ONE volunteer available for every TEN good ideas, we'd really get stuff done.
Re:JFDI!
yumpy on 2004-01-14T06:35:41
The need I see for TPF's approval is in connection with the public activites of the team, such as contacting HR departments to determine their receptivity to the idea, and to ask for sponsorhip. Somehow I think we'd be more likely to get taken seriously and get cooperation from them if they knew they were dealing with a sanctioned community project, rather than just an individual or an ad-hoc collection of interested JAPHs ! 8-}Just like I'd refrain from developing a complex system for a client until the contract was in hand, I'd want to be pretty sure that the "customers" and "subjects" of a certification program were likely to be receptive to it before committing resources to making it happen.
It's obvious that "somebody has to do the work", but it's important, IMHO, that it not be in vain.
We don't believe you can do it.
schwern on 2004-01-15T00:38:54
Well then there's nothing to worry about. With a plan as dependent on gaining approval before doing any work as that, we'll never see certifications.
Allow me to reiterate the core problem in big, bold letters: you will never get support for a Perl certification program from the Perl community until you can convince people that your cert is a valid measure of one's Perl programming abilities. And you'll never convince anyone until you have some solid picture of what your certification program looks like. Everything up to the point where you produce a viable certification program is a big waste of time.
Once you show that you can measure the skill of a Perl programmer, everything else will fall into place. Right now, since you have vapor, when you talk about Perl certifications all anyone can do is think about the existing certification programs and how much they suck. You're not going to get folks excited about certs like that. You must show that yours will not suck or you will not get people on board. WE DON'T BELIEVE YOU CAN DO IT. You must prove us wrong before progressing anywhere.
Your testing program is the single, critical point in putting together Perl certifcations. Without it, nothing works. With it, everything becomes easier. It has no dependencies, yet everything depends on it. Nothing else in your plan is that important. You can live without TPF sponsorship, you can survive without initial business approval, but you can't go anywhere without a good way to measure a programmer.
So before you generate more heat trying to gain approval, find out if you can even do the task.
Assume there is support and skip directly to step six.
This is not a cowboy model, this is simple prioritization. Your most difficult and most important task is to produce a certification program. Seeking approval does not help this. You will not get clear approval until you can convince people that this isn't just going to be another MSCE. If you cannot produce a viable program, there is no point proceeding any further. If you can produce a viable program, that's your best selling point.
Re:We don't believe you can do it.
yumpy on 2004-01-15T01:12:39
FYI, I've never claimed that I could write a Perl certification that "didn't suck" (although I believe I could, at least for some definitions of "suck"). What I've repeatedly stated is my opinion that we in the Perl community should be able to do at least as good a job as anyone else has ever done on a software certification -- and I'd be willing to bet we could do MUCH better. Just ask yourself, have you ever encountered a group of more brilliant, giving, and literate people anywhere else in life? I sure haven't.But this isn't a software development project (at least primarily). It's partly political, partly social, it involves a lot of very difficult technical writing (i.e., generating huge numbers of test questions or coding exercises that don't suck), and, oh yeah, a little bit of coding. Then once the first test is done, it itself needs to be validated through psychometric testing (at least in the orthodox model of test construction).
And of course due consideration has to be given to a viable business model, and proper management of legal risks. As I've repeatedly said, I have no doubt whatsoever that we in the Perl community could produce a state-of-the-art certification test; in some ways, that's the easy part of this exercise, IMHO.
That's why I'm concentrating on the harder parts first, like getting people interested and testing the market, before investing large amounts of effort into the complicated and difficult, but relatively straightforward task, of actually writing a test.
But YMMV! If anybody wants to pursue a different model, please be my guest, and best of luck to you! After all, I've got other things to do, such as writing a book, instead of this nearly book-length collection of articles for U.P.O.! 8-}
Re:We don't believe you can do it.
RobertX on 2004-01-15T02:37:14
You would be incorrect. I BELIEVE IT CAN BE DONE. If I believe it can be done, then there are others who believe it can be done.Re:We don't believe you can do it.
yumpy on 2004-01-15T05:22:31
I also believe that the development of a Perl Certification program could be accomplished.For the naysayers to be correct, IMHO, one or more of the following would have to be true:
a) Perl would have to be fundamentally different from every programming language for which certifications already exist, e.g., in ways that make it's programmers more like mystics than scientists,
b) Perl Subject Matter Experts would have to be inferior to those of the other languages and software specialities, with respect to creating usable testing materials,
c) the Perl community would have to be so disjointed in its actions and neglectful of its own best interests (IMHO) as to never even attempt to create a certification test
It's "c" that I worry about . . .
Re:We don't believe you can do it.
chromatic on 2004-01-15T07:02:36
Or, for the sake of completeness:
d) Certification can't tell you what you really need to know. Does this person fit in with your team? Can he write maintainable code? Can he learn what he needs to know if he doesn't know it now?
Maybe a good certification program can alleviate that. I don't know. I've yet to be convinced. Convince me with a concrete proposal.
Re:We don't believe you can do it.
RobertX on 2004-01-15T14:11:19
I don't think any certification, whether it was premo or not, could tell you whether someone fits into your team or not or if he can write maintainable code (which is another Perl subject altogether!). That happens at another level in the recruitment process. I would say that it does answer to a degree "Can he learn what he needs to know...".
But the pattern of responses is always the same -- a few well-known JAPHs coming out against it, and the rank-and-file community members taking neutral, or strongly favorable positions, because they don't want to convert to Java/C++ etc. just to gain employment, and certification might help there (see my TPJ article). As I've argued in that article, and in postings on the perlcert.perlocity.org Wiki site, objections traditionally center around the notions that a) certification cannot be done well, and is therefore a disservice to the community, or b) it's generally a scam perpetrated by greedy vendors on a helpless population of job-seekers.
I've argued that a voluntary certification program with well-defined and realistic goals could be of benefit to some job-seekers, and to Perl's image, and that if we were to make a profit by providing the service, having those funds flow into the coffers of TPF could be used to build our community further (by funding Perl Mongers, e.g.).
It's interesting to see that the people here attacking the idea most vehemently are those who have already proven their Perl expertise most clearly. But not everybody can write a book or publish a magazine, so the less accomplished JAPHs out there deserve the option of using a standardized testing procedure to substantiate their knowledge. The existence of several crummy Perl certification programs already demonstrates the need for the service; all I'm proposing is that we create a good one, and by so doing take control over this process, and this aspect of Perl's future.
C'mon Randal and brian, you're educators! Certainly no testing regimen is perfect, but don't you believe that objective testing procedures can differentiate between the students who learned their lessons well, and those who didn't? All I'm proposing is that we help would-be employers make that same distinction, and incorporate that test result into their hiring formulas, along with other sources of information.
Why shouldn't JAPHs have the same options as Java or C++ programmers in this respect?
P.S. Please read the TPJ article before jumping to any further conclusions.
Re:Certification has Merits; Don't Panic!
brian_d_foy on 2004-01-12T06:20:48
I read your article.
There is no need to devolve this debate into personal attacks. Let's stick to arguing the mertis.
You still ignore my most important concern, and the one I have expressed for years---who is going to take on the legal laibility. No one has ever answered that. Cisco, Sun, and Microsoft know how to answer that question because they have legal departments and deep pockets.Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-12T06:50:44
Personal attacks? What attacks? I merely directed some questions to some real people, who like me, know a lot about software training, a discipline that's closely related to testing. I didn't mean any disrespect, if anybody was offended. Let's try to let the discussion heat up without getting inflamed along with it 8-}
Regarding your "liability" concern, I addressed that in detail in a separate message (probably posted after your request for comments on it).Re:Certification has Merits; Don't Panic!
ethan on 2004-01-12T11:02:24
You still ignore my most important concern, and the one I have expressed for years---who is going to take on the legal laibility.
Is there such a thing as a legal liability at all? In my understanding, no certificate ever guaranteed infallibility. And that should be easy to settle in a sort of disclaimer.
And since you mentioned them: Has anyone ever seen one of the certificates as offered by Cisco, Sun or Microsoft? I haven't, but I am pretty sure that they refuse to take on any legal liability.Re:Certification has Merits; Don't Panic!
brian_d_foy on 2004-01-12T14:00:24
Remember that a woman sued McDonald's because they served her hot coffee, and she won.
Lawsuits do not have to have merit to destroy you. They do not bother deep pockets because to them the risk is minimal. To someone like TPJ, it could be the last thing they ever do, however. I do not see it worth the risk.Re:Certification has Merits; Don't Panic!
RobertX on 2004-01-12T17:44:31
How about finding a lawsuit for certification in our field? I can pull a thousand bogus lawsuits out of the air that have no merit but I have yet to see anyone produce one based on a certification firm being sued because the "certificatee" turned out to be incompetant.
Your argument has no merit.
Re:Certification has Merits; Don't Panic!
simonflk on 2004-01-12T19:09:13
I can pull a thousand bogus lawsuits out of the air that have no merit but I have yet to see anyone produce one based on a certification firm being sued because the "certificatee" turned out to be incompetant.
Your argument has no merit.I have only ever seen white swans.
Therefore all swans are white.Your argument has no merit
;) Re:Certification has Merits; Don't Panic!
RobertX on 2004-01-13T13:13:30
No my argument is to show me where that has happened in the industry. Certainly there have been many such instances in the certification processes that are currently in place where a person just studies and passes and gets hired. Have lawsuits resulted back to the certifier?Re:Certification has Merits; Don't Panic!
krellis on 2004-01-12T20:18:12
Remember that a woman sued McDonald's because they served her hot coffee, and she won.Coffee which was hotter by a significant margin than the generally accepted / legal limit for hot beverage temperatures. That case is completely incorrect to use in this type of discussion, because the woman had a perfectly legitimate case. The coffee was considerably hotter than it should have been for her safety as a consumer. If you're going to pick examples, pick relevant ones. If you can't find relevant ones, perhaps that should tell you something.
Re:Certification has Merits; Don't Panic!
pudge on 2004-01-13T00:20:32
Coffee which was hotter by a significant margin than the generally accepted / legal limit for hot beverage temperatures. That case is completely incorrect to use in this type of discussion, because the woman had a perfectly legitimate case.
You are incorrect about the former, and the latter is an opinion (borne out by the result of the case, of course). There is no legal limit for the temperature of such beverages, and what is acceptable differs among people. The whole point was that the patrons complained the coffee got too cold too fast, which meant making it hotter. They stopped getting complaints when they did. Clearly, it was generally acceptable to make it that hot.
The case hinged on expectations: that someone would not expect to be scalded by coffee they just purchased, should it spilled. I think the case was decided incorrectly, but whatever: this case is actually quite applicable because someone could say they had the expectation that the certification is more meaningful than it actually is.Re:Certification has Merits; Don't Panic!
yDNA Barak on 2004-01-15T05:19:08
There is no legal limit for the temperature of such beverages...(I know I should keep my mouth shut...) Actually, there is a legal limit, largely dependent on the boiling point of said beverage. To cross this limit, it will no longer be a beverage (in the nominal sense of the word). It's the law, baby!
Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-13T11:35:30
My mid-90's certifications from Sun for Solaris System Administration are typical of their breed, AFAIK, and bear no disclaimers of any kind on them.However, the Sun web site has a disclaimer that applies to their certification tests, that states there are no warranties of merchantability, fitness for any purpose, or employability, etc. You can read it for yourself here at Sun's web site.
IANAL, but IMHO, this disclaimer just proves that Sun is trying to protect itself, not that a plaintiff with a compelling case would be unable to win a judgment, or get a generous settlement. But even so, AFAIK the instances of certification authorities being sued (for negligence, etc.) would seem to be few, or else we would probably have heard about some of them by now.
An idea suggested by another poster for handling the liability issue is to base the certification authority in the UK, where the judicial system is reportedly less indulgent of company-killing judgments in such cases.
Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-12T21:27:14
I did address those concerns, in this commentOn testing
brian_d_foy on 2004-01-12T06:42:59
Tim Maher asks if I beleive that objective tests can work.
I cannot agree that any test is objective.
Stephen Jay Gould's The Mismeasure of Man is an interesting read for those who have not come to despise tests yet.Re:On testing
yumpy on 2004-01-12T06:57:46
Nothing is perfect, or perfectly measurable (just ask Heisenberg). Like all things worth measuring, objectivity in testing comes in degrees. There's a whole science devoted to maximizing the objectivity of testing instruments, and measuring the degree to which that goal has been attained. It provides procedures that result in tests that are much better and fairer than the ones college professors dream up on their own -- but with some greater associated costs. (And I'm speaking from experience on both sides here - college prof and psychologist!)Testing can be done well, and with the wealth of talent and experience we've got in the Perl community, it could be done very well!
Re:On testing
RobertX on 2004-01-12T13:26:26
I fall into the category of "it would be good". Please forge ahead!Re:Certification has Merits; Don't Panic!
delegatrix on 2004-01-12T16:55:06
Why is the focus always on testing? Discussing methodolgy is so beyond topic right now. How about laying the foundations for certification instead. What is the perl body of knowlesge? What are standards of professional practice? Who is an issuing authority? Answer those questions and then maybe you can come up with a meaningful certification that can a) have recognized stature, and b) lead to a professional society for perlers.Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-12T17:35:53
The "focus is always on testing" because that's the role of a certification procedure -- to ascertain the knowledge of the testee. But of course you're correct that there would be some homework to do before sitting down and writing that test 8-} In that connection, we have lots of resources (mostly in print) that already do a pretty good job of summarizing the "body of knowledge" for Perl, and we could certainly come up with some "best practices" and "preferred coding styles" too.But here's the real answer to your question: As I've argued (against Ziggy, IIRC) over at perlcert.perlocity.org, there's little point in wasting energy cataloging the essentials of Perl knowledge unless we're already committed to developing a certification program, so IMHO doing it the other way around would be putting the "cart before the camel".
Re:Certification has Merits; Don't Panic!
ziggy on 2004-01-13T02:40:04
I strongly disagree.There's plenty of benefit to ascertaining the Perl Body of Knowledge without committing to a certification program. That body of knowledge serves as a repository of best practices on how to (and how not to) write Perl programs.
Currently, there's a vague body of knowledge floating around, built up from received wisdom, cargo cult programming, and uncritical repetition of previous practices. Building the BOK examines all of those practices and codifies what works, what doesn't, and why. I know I'd find that rather handy. Especially if I wanted to avoid the classic pitfalls in an unfamiliar domain (like the SQL injection bug when writing database backed web apps).
«Cart before the camel»? Hardly.
Whether or not the certification that could follow out of such an effort is worthwhile is a totally separate discussion. I still have yet to see a persuasive argument that a Perl Certification -- even with the prerequisite BOK -- is anything more than a waste of time, money and effort.
Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-13T04:51:50
<Ziggy> There's plenty of benefit to ascertaining the Perl Body of Knowledge without committing to a certification program.
... Building the BOK examines all of those practices and codifies what works, what doesn't, and why. I know I'd find that rather handy. Especially if I wanted to avoid the classic pitfalls in an unfamiliar domain (like the SQL injection bug when writing database backed web apps). «Cart before the camel»? Hardly.
<Tim> If the BOK is so important, how have we survived so nicely for so long without having it codified? Why aren't people (besides you) clamoring for its development?
<Ziggy> Whether or not the certification that could follow out of such an effort is worthwhile is a totally separate discussion. I still have yet to see a persuasive argument that a Perl Certification -- even with the prerequisite BOK -- is anything more than a waste of time, money and effort.
<Tim> Then you're probably not talking to the Hiring Managers, or the HR people, for whom Perl Certification is a veritable Holy Grail.
I'm in favor of starting out small with certification, by testing a basic understanding of the language, and optionally of other specialty areas (CGI, DBI, etc.). And at the base-level of testing, IMHO, programmers should be given credit for knowing any solution to a problem, rather than having to comply with somebody's notion of the "best" way to do it. The construction of other tests, requiring a much greater development effort to address more esoteric goals, could be added on later, if warranted by the demand.
Re:Certification has Merits; Don't Panic!
ziggy on 2004-01-13T05:37:08
If Perl Certification is so critical to Perl's success, how did it get as successful as it is today without a certification program?If the BOK is so important, how have we survived so nicely for so long without having it codified? Why aren't people (besides you) clamoring for its development?A Perl BOK would be a useful tool, but it is not a pre-requisite for Perl's success. A body of knowledge is a necessary pre-requisite for any meaningful certification program, anywhere -- Perl Certifications, driver's licenses, whatever.
When I was interviewing for jobs as a Perl programmer many years ago, certification was not a holy grail by any means.[Y]ou're probably not talking to the Hiring Managers, or the HR people, for whom Perl Certification is a veritable Holy Grail.When I was a hiring manager a few years ago, certification was not a holy grail by any means. For the HR people I worked with, certification was not a holy grail in any way, shape or form, either.
Now that I'm interviewing for Perl positions again, certification is still not a holy grail. The hiring managers I am speaking with today do not find certification to be a holy grail, either. My friends who are also interviewing for jobs now do not find certifications to be a holy grail, either.
I interviewed for my first programming job over 13 years ago. I am not new to this industry. I have interviewed for many jobs over the years, and interviewed dozens of candidates as well. I can say without reservation that certifications or the lack thereof have never once entered into the process. Calling them a "holy grail" is elevating them to a level they do not deserve.
I do not maintain that my experiences are in any way typical. Nor do I blindly assert that hiring managers or HR staff are uniformly ignoring certifications.
But for you to assert that hiring managers and HR people consider certifications to be a "holy grail" is patently absurd. They are merely one tool among many that some managers will use occasionally to weed out a stack of resumes. No more, no less. A few annecdotal managers crying out for a Perl Certification does not provide a mandate to create one. Especially if a good deal of hiring managers couldn't care less about Perl certification in the first place.
Re:Certification has Merits; Don't Panic!
yumpy on 2004-01-19T07:35:28
If Perl Certification is so critical to Perl's success, how did it get as successful as it is today without a certification program?As you know I've stated elsewhere, by not providing an official Certification Program for Perl Programmers, we have necessitated the creation of hundreds of private Perl certification programs in IT departments all over the world.
It's a testament to the appeal of the language that so many companies have been willing to take on this burden on their own. But can we really expect them all to upgrade their homegrown Cert Programs for Perl 6? Maybe they'll just throw in the Perl-towel and go totally to the Java side at that point. And what about the other companies that have never had their own Cert programs, and don't feel qualified to judge and hire JAPHs? Why should we let them fall by the wayside, rather than pulling them into our flock (er, herd).
I think it would be wise for the Perl Community to help companies, who are unwilling or unable to shoulder the burden of developing and/or maintaining their own programs, to easily validate the knowledge of JAPH applicants, and therefore make them easier to hire.
When I was interviewing for jobs as a Perl programmer many years ago, certification was not a holy grail by any means.
Thanks for sharing your personal experience. Given the absence of hard data on the subject, we all have to make use of the info we can get, but we're all like blind men feeling different parts of the Camel.
Just because a company already has a homegrown Perl cert doesn't mean they're looking forward to revamping it for Perl 6, and that they wouldn't prefer to delegate at least the purely-technical-knowledge part of that vetting process to the Perl Community instead. And having Perl 6 on the horizon is certain to make the Grail of Perl Certification ever Holier by the minute, as companies ponder what to do with their (somewhat) outdated homegrown programs. I think we should carefully consider the possibility that the arrival of the spectacular Perl 6 could ironically cause us to lost market share to Java, just because it's emergence will render obsolete so many homegrown cert programs, and consequently shift new development projects away from Perl.
... for you to assert that hiring managers and HR people consider certifications to be a "holy grail" is patently absurd. They are merely one tool among many that some managers will use occasionally to weed out a stack of resumes. No more, no less. Even a Swiss Army knife is by some definitions "only one tool", but as tools go, it's hard to beat on a camping trip. 8-}
If we had an entity worthy of a name like the Perl Professional Institute, we could ask them to randomly call 100 of the Fortune 500 companies and ask them if they have a Homegrown Cert and what their plans are for upgrading it for Perl 6. I'd be willing to bet a Larry-autographed copy of the Camel that the IT managers would ask why we don't get on the ball and help them in this area -- as I've been hearing in my own (less ambitious) research.
We're not likely to get the PPI anytime soon, but perhaps it would be worth creating a special TPF-sponsored email address, or Wiki page, through which we could solicit input from businesses on these matters.
Or how about assembling an "Enterprise Advisory Panel", or maybe a "Camel Trainers Club" of corporate IT execs, who could tell us what their Perlish needs are and help guide us in directions that will meet them.
Wouldn't these efforts be within our meagre capabilities?
Let's face it, Perl is not a toy in some geek's closet anymore; it's a powerful resource with a lot to offer the business world. Isn't it time we started capitalizing on that base to build up our community's reputation and stability, and pave the way for an even brighter future?
I would probably get a certificate now if it was affordable, and I knew that the fees would have some benefit to the perl community. Thats after writing an article and several perl modules on CPAN.
A side effect of certification, like any testing is that I would be able to find the weak spots in my knowledge and work on them.
Another side effect is that we could say to employers that we have $bignum certified programmers that can get the jobs done.
As for liability - base the project in a country outside the US where it is safer from meritless litigation. The UK has both sensible (on the whole) courts and a large section of the perl community.
Re:Alright for some
yumpy on 2004-01-12T23:18:54
Excellent idea, AFAIK, about basing the legal entity providing the certification services in the UK, to enjoy the benefits of a more reasonable legal system (re: liability exposure).It just shows that, with a worldwide pool of clever people at our dispposal, there won't be many obstacles that we can't overcome! 8-}
Re:The sad fact is...
yumpy on 2004-01-15T05:47:33
Ah, I see you're a conspiracy theorist. I've always liked a good one myself, but you're "barking up the wrong dude" in this case.For anyone who believes that I would purposely distort the reality of an event that transpired in front of 200 witnesses from my professional field, if indeed that's what happened, I'd like to hear the details of your latest alien abduction.
P.S. I'm still waiting to learn what a playback of the audio transcript reveals.
Re:The sad fact is...
Simon on 2004-01-15T13:34:21
For anyone who believes that I would purposely distort the reality of an event that transpired in front of 200 witnesses from my professional field, if indeed that's what happenedCalling it an "alien abduction" doesn't stop it being what happened. And less of the ad hominems. If you can't argue against the facts, please don't argue at all.
Re:The sad fact is...
yumpy on 2004-01-16T21:31:28
Huh? Please re-read my post. I didn't call the OSCON event an "alien abduction" -- although I find that notion intriguing, and I would like to hear more about that interpretation of it! 8-}
Could we somehow work this into mix? I'm certain that I would qualify as a journeyman myself, if only I could find a master of suitable breeding. Oh Captain! My Captain!
like all business there must be a demand. Is this a solution to a problem created by demand OR is it a solution looking for a problem created to make demand?
I looked at this article and it reported an audience of what 200 bods, (100 for, 7 against). Then I read
is 100/200 a representitive of the perl community? I have no problem with business but dont try to stamp a commercial product as *officially endorsed* endoresed product of the *Perl community*.
I suspect without the *official stamp* the business model I dare say a bit weak.
<rant> What does the Perl community care for *official* certification when you have source code, documentation and Internet connection and some good reading? </rant>
Tim Maher is a man whose integrity I hold in the highest regard, and I am deeply disturbed by the snide insinuations I have read here regarding ulterior commercial motives on his part, and by other attacks on Tim's character.
Like several other participants in this thread, Tim has been recognized with a White Camel for his selfless and tireless contributions to the Perl community. While I deeply respect anyone's right to vehemently disagree with his views on certification, to express genuine doubt that his enthusiasm and belief in the notion is well-founded, or to assert that he's just plain wrong, I find it highly distressing that anyone should cast aspersions on his motives in wanting to promote this discussion in the first place.
Tim is one of the best, most generous, and most honest men I know. I wish I had half his integrity. And I am deeply offended that anyone should imply that he is otherwise, or cast such unfounded aspersions on the motivations of such a significant and long-term contributor to our community.
More practically, knowing Tim as I do I find that these misguided attacks on his character and motivations only serve to significantly weaken the force of your arguments against certification. If you can be so wrong in your assessment of the man proposing the idea, how can I have any faith that you're right in your opinions about the idea itself?
So debate Tim's proposition as forcefully as you like, but treat the man himself with the respect that his long track-record of service to our community deserves. Anything less demeans us all.
Re:Distressed
merlyn on 2004-01-19T06:57:24
I agree. While I may have my professional disagreements with Tim about the value of certification as anything other than a money-making scam (grin), I hope none of my comments have been construed as a personal attack on Tim. I know him well enough to know that he is not intentionally misleading anyone. Maybe a bit misguided, but not misleading.Let's keep the comments about certification to the merits of certification, not about the proposers or any bystanders.
Re:Distressed
jacques on 2004-01-25T23:34:46
{wink, wink} Okay!
On the basis of the written record provided by Autrijus Tang, it seems likely that the questions were stated a bit differently than I and some others present had recalled; here's a pseudo-code rendition of his version:
if (outcome_of_discussion() > $threshold_for_action) {
do_development()
}
Accordingly, I've offered an updated interpretation of the results. Put simply, IMHO we should now say that the majority voted "conditionally for" Perl Certification, rather than simply "for", to reflect their expressed support for "continued discussion, and possible development, of a certification program by the Perl community".
FWIW, the panelist who actually phrased and uttered the voting questions at OSCON has doubts about the accuracy of the new report on the "yes" question, but he was very busy during the conference and admits his recollections are fuzzy. But in such cases we've got to go with the written evidence, IMHO.
Sorry for the confusion engendered by my original reports of the outcome, and thanks to all who helped sort things out -- peer review rules!
Re:New Interpretation of OSCON Vote
n1vux on 2004-01-22T16:27:30
So some unknown portion of the FOR voters were voting against prejudicially closing debate, as opposed to voting for immediate development. This adequately explains the shock & awe in the community to the early reports.As Col.McKean said in Continental Congress about the question of discussing the question of Independence, nothing is too dangerous to talk about. In an open source community, I would hope the vote would be to discuss almost anything. (I draw the line at discussing deprecation of Perl in favor of VB.:-) Certification is not as quite as dangerous a discussion as treason, so it should be talked about, and I'm glad OSCON voted for continued discussion.
I'm all for our discussing a certification process, what benefits it would bring to what portions of the community, and if it's antithetical for a language with the TIMTOWTDI motto. It will take a lot of discussion to convince me that the benefits out weigh the risks for this community. I don't want to see the community divided into the "Official Certified Style" and the Swiss Army Chainsaw TIMTOWTDI style.
I am somewhat disturbed that a vote to continue discussion has been repeatedly reported as a vote to start a process. Apparently this was an inadvertent result of natural human optimism and vagueness combined with lack of formal minuting at a non-bureaucratic event.
As to the idea of certification itself, first an analogy: In the presidential polls, President Bush is currently shown losing to an unnamed democrat (the imaginary perfect candidate) but winning against any named democrat in the race today. Likewise, an imaginary Certification may be perfect, but any real one has its problems.
A case in point: I was more than annoyed by a new certification system adopted by a volunteer group I'd been an instructor for; a means for quick-certifying in-place instructors and coordinators was omitted (promised "soon" and only slowly rolled out). I've drifted off to other things, as the organization fills up with people with paper certificates and frequently no practical field experience. Maybe some day I'll take the time to take challenge exams to get certified at the level I was operating at under the prior peer review system, and maybe I won't bother.
What's that have to do with Perl? Our Perl culture is inherently "alternate", proudly so. Perl was not released by a Vendor but developed by Larry with his community. It works because it works, not because Marketing says it's the best or only. Branded Certificates are useful in monopoly or oligopoly markets (MS, Novell, CISCO) where there is essentially a common style, whether necessary or merely assumed and enforced by market pressures (Windows GUI apps, network management). In general software engineering, a single certificate won't cover all uses (BioPerl? PDL Perl? Graphical perl? SysAdmin Perl? PWB/Toolsmith Perl?)
If a non-degreed or re-targeting job-seeker want (feels the need for) a meaningful Certificate for Human Resources, Head-hunters, Monster.com, and hiring PHB managers to admire, s/he should talk to a university. A certificate in Unix, WebDev, IS or IT from a degree-granting university will or can include a course or two of Perl, and also include that other stuff that's useful in real work that a language-based certificate doesn't usually include, that theory stuff that Damian and Larry enjoy slipping into our language, but in a more practical setting than in a full degree. example1 Example being marketed to Employers (1 or 2 Perl courses in eBiz and WebDev certs). [Disclosure -- since NEU has let me teach Certificate and Bachelor level classes before, I'm unlikely to take them there either.] Note that these certificate examples don't say Perl Certificate or Unix Certificate, they say what specialty the training good for, which should be much more valuable to the right manager.
The existence of University certificates that include Perl tells me that the lack of certified programmers is not what is blocking Perl adoption in mainstream IT hiring. I think it's a fantasy that a sudden supply of Certified Perl Mongers would create a demand for such.
Personally, I think a "Completed my course" bit of wall art from Dr.Tim, Merlyn, or Damian, is more impressive than a "passed our multiple guess test" from the Perl Certification Board (even if signed by the same Dr.Tim). (The infamous $3 certificate demonstrates attending a conference, I guess, so it's actually more impressive than the reductio ad absurdem it pretends to be.) A job that requires a multiple-guess certificate may be better than no job at all, but it's a warning of PHB's or cookie-cutter work.
Bill
Perl is my *second* swiss-army language in 25 years.
I started with PL/I.Re:New Interpretation of OSCON Vote
yumpy on 2004-01-22T18:08:31
I am somewhat disturbed that a vote to continue discussion has been repeatedly reported as a vote to start a process. Apparently this was an inadvertent result of natural human optimism and vagueness combined with lack of formal minuting at a non-bureaucratic event.
... not to mention the spontaneity of the vote being sprung on us all, and the odd wording of the questions. What's more, "optimism" wasn't required for a reaction of "shock and awe". Nat Torkington, who was both present and against the idea of Certification, told me he was "horrified" at the strong majority of those voting on the "FOR" side. I think that was one of the factors that led me, perhaps with the aid of some renegade memory-polishing processes, to believe that they had truly been voting to "make it so". What's that have to do with Perl? Our Perl culture is inherently "alternate", proudly so
... In general software engineering, a single certificate won't cover all uses (BioPerl? PDL Perl? Graphical perl? SysAdmin Perl? PWB/Toolsmith Perl?) You seem to be assuming, like many do, that a Perl programming certificate would document the testee's ability to program in different application areas. While that would be a laudable goal, it would also be a very ambitious one. My feeling is that the first type of certificate we'd want to produce would be simply one that tests familiarity with the core language itself, which would be just enough to help HR departments screen out JAPH impersonators. Judging their problem-solving ability at the higher levels of performance, or familiarity with techniques required for different specialized kinds of programming, would be outside the scope of the test. (Although I could see adding DBI and CGI sub-tests, etc., later on.)
The "grandiose" model of certifying all skills remotely Perlish (general problem-solving ability, exemplary style, etc.) is frequently assumed by those concerned about the possibility of the success of a Certifcation Program, but that presents formidable challenges to implementation. I don't think we have to go that far to make a useful advancement from the status quo.
The existence of University certificates that include Perl tells me that the lack of certified programmers is not what is blocking Perl adoption in mainstream IT hiring. I think it's a fantasy that a sudden supply of Certified Perl Mongers would create a demand for such.
AFAIK, the program I'm affiliated with at the U of WA is one of only a handful in the entire USA. So the relative shortage of such programs could indeed be a current obstacle to job-seekers.
Imagine a Hiring Manager who realizes (at some time in the near future) that Perl 6 provides the enterprise capabilities his programmers need, and is considering using JAPHs on an upcoming project. Do we really want him to choose Java because his HR department has no way of screening out applicants who don't know Perl 6?
Also, I believe the "character" and "reputation" of a language affect its opportunities in the enterprise world, along with its technical properties. I was utterly shocked when I discovered early on in my explorations of Perl that it didn't even have a beautifier! That was a deal-breaker for me, so I had to either write one, or two, or give Perl up. I'm convinced that many Hiring Managers feel the same way about the current lack of a decent Certification Program. It's true that many S&P 500 companies with big IT shops have developed homegrown ones, but I'd like to make it easier for the Russell 2000 (i.e., smaller companies) to use Perl.
Personally, I think a "Completed my course" bit of wall art from Dr.Tim, Merlyn, or Damian, is more impressive than a "passed our multiple guess test" from the Perl Certification Board (even if signed by the same Dr.Tim).
I may have agreed with you ten years ago, but not now. The difference is that every corporate training organization I know of has long ago stopped testing their students at the course's end to decide who gets a certificate (govt trainers are still an exception, AFAIK). So at this point in time, a University certificate has an edge over a corporate training class in documenting that the student actually acquired a certain amount of knowledge (assuming proper cheating-controls - a problem nowadays with "remote learning"). That's why IMHO JAPHs should have the option of voluntarily taking a standard, approved certification test, to document their knowledge (yes, even the ones who've got a Timological or Damianiacal Certificate).
Bill
Perl is my *second* swiss-army language in 25 years. I started with PL/I.<GRIN> Me too, with the PL/C (Cornell) variation, although I consider AWK the first language that freed me from the "tyranny of the computer" to code in a more "organic" style.
Re:New Interpretation of OSCON Vote
ziggy on 2004-01-26T21:03:09
I did my best to observe foy's law[*] on this discussion, but this point needs to be addressed:You seem to presume that the relative paucity of university certificates (and the lack of Perl Certification) presents an obstacle. There are many other valid hypotheses that explain this datapoint:AFAIK, the program I'm affiliated with at the U of WA is one of only a handful in the entire USA. So the relative shortage of such programs could indeed be a current obstacle to job-seekers.The existence of University certificates that include Perl tells me that the lack of certified programmers is not what is blocking Perl adoption in mainstream IT hiring. I think it's a fantasy that a sudden supply of Certified Perl Mongers would create a demand for such.But the most interesting explanation is this: these kinds of programs exists where there is a market for these skills, and a market for these continuing education programs. In a country of 291 million residents, there are only a few metropolitan areas where a significant number of programmers live who (a) are looking for jobs with these skills, (b) are seeking certification (c) in large enough numbers to start a university program such as these.
- Low demand
- High cost to produce, administer and market a program
- High cost (in time and money) to attend a program
- Low value (real or perceived) for university certifications (vs. an industry-wide certificate)
- Low value (real or perceived) for "wall art" -- either from a University or a certification program
While you may find it surprising that there are as few as (say) 12 programs in the US similar to the one at UW, someone else could easily look at the same data and be equally surprised that there are as many as 12 such programs in the US. This data point is just that -- one data point. It does not support a conclusion in either direction.
In that light, it is also quite possible that university certifications (or even Perl certifications) have no direct bearing on job-seekers finding work.
Again, you continue to presume that Perl6 will provide the capabilities that hiring managers care about, that it will provide those capabilities from day 1, and that choosing Java for any reason whatsoever is a bad thing. From as far back as the kickoff meeting in 2000, the Perl porters accepted that the Perl5 -> Perl6 migration would take years, just like the Perl4 -> Perl5 migration did. This is neither good nor bad, this is just the way the industry works. It also took years for Perl5 to mature. Yet you want managers to hire as many Perl6 programmers as soon as possible, without consideration to whether this is a good thing or a bad thing for Perl.Imagine a Hiring Manager who realizes (at some time in the near future) that Perl 6 provides the enterprise capabilities his programmers need, and is considering using JAPHs on an upcoming project. Do we really want him to choose Java because his HR department has no way of screening out applicants who don't know Perl 6?But the one point you still fail to acknowledge is that there are two kinds of hiring managers out in the world today: those who can identify talented job-seekers on their own, and those who follow the herd. The first kind of manager has already hired Perl programmers, has not had any problems identifying talented applicants in the past, does not have any problems today, and will not need any help (from, say, any form of certificate program) to identify these applicants in the future. The second kind of manager does what everyone else does. If the rest of the industry is using Java, C#, or VB, so will he.
In that light, a little thing like certification alone will make Perl6 the new darling of the tech sector once it is released. Managers need more than access to certified programmers to bless a new language as an industry standard.
Re:New Interpretation of OSCON Vote
delegatrix on 2004-01-26T23:00:32
My feeling is that the first type of certificate we'd want to produce would be simply one that tests familiarity with the core language itself, which would be just enough to help HR departments screen out JAPH impersonators.Do you really mean to use the word 'familiarity'?
As a person who hires perl programmers, I'm not looking for familiarity, I'm looking for proficiency. And not just proficiency in in knowing perl, but proficiency in the application areas, namely database interaction, secure web programming, and statistical processing, for example. If someone has demonstrated experience with those, but knows 10% of core perl (like me?
;-), then great. Tell me something meaningful about your experience. I'll weed out the players. If I'm not hiring an entry level position, I want to hear about projects, applications, live sites, etc.
I can evaluate familiarity in an interview in about 2 minutes.
But proficiency, there's the rub.
Re:New Interpretation of OSCON Vote
yumpy on 2004-01-27T17:29:45
I'll weed out the players. If I'm not hiring an entry level position, I want to hear about projects, applications, live sites, etc.I can evaluate familiarity in an interview in about 2 minutes.
I'm sure you can; but the benefit of even the most rudimentary type of Perl Certificate would be that it would let your HR department filter out those who don't have a basic grasp of the language before they ever get to the point of a scheduled interview with you (whether by phone or in person). If you get 200 applicants for one job, as some people in our community do, that would save you a total of 400 minutes of wasted time at your 2-min. rate -- or even more, if the time blocked-out of your schedule for needless interviews could not be efficiently recycled.
