From CNet:
The OpenBSD project hopes new changes to its latest release will eliminate "buffer overflows," a software issue that has been plaguing security experts for more than three decades. [...]I've been thinking a lot about innovation this weekend. There was a presentation at the MySQL Users Conference where an analyst from the Meta group asserted that open source communities are fundementally unable to innovate. There are some serious flaws in the argument, as well as some serious flaws in the obvious rebuttal.The research was funded by a $2.3 million grant from the Defense Advanced Research Projects Agency (DARPA) to the OpenBSD Project, but the latest changes go beyond the original grant request, de Raadt said.
"This really wasn't part of the DARPA grant," he said. "But it happened because the DARPA grant happened, because when you throw a bunch of...guys into a room and get them drunk, this is what you get." De Raadt was careful to point out that the group paid for its own beer.
The big problem I have refuting this all-innovation-is-proprietary argument comes from irrefutably disproving it on its own terms. I have not located the one obvious counterexample from the open source community where innovation occured devoid of government R&D spending or university support. Or is there?
Why are you precluding govt/uni supported activities? Open source includes them, includes people who are commercially employed, and also includes the occassional bit of totally freelance hobby effort.
You might be trying to distinguish between funded goals (although funding a goal with the intent of making the result open source surely counts as open source and not proprietary), but often the open source that comes was not part of the original plan but was invented to scratch an itch encountered on the path - and is the result of the developer only, regardless of the funding process.
Re:baby with the bathwater
ziggy on 2003-04-14T14:21:39
In order to irrefutably disprove the assertion that innovation can only happen within a commercial (or quasi-commercial, funded) setting. Or, rather, it cannot happen in an "open source" environment. It's a specious argument, but if I can defeat it on its own terms (i.e. with one hand tied behind my back), then it is a provably false assertion, not an arguably false assertion.Why are you precluding govt/uni supported activities? Open source includes them, includes people who are commercially employed, and also includes the occassional bit of totally freelance hobby effort.I need to get a hold of the speaker's presentation. I've heard similar lines of thought before, and I don't know if he's fundementally excluding university research in his argument. I'd rather not assume that this analyst is that stupid.
They are good enough that most other languages have been moving to adopt them, and the pcre (perl compatible regular exporession) project developed an implementation. (After Perl 6 comes along, Perl will be the only major scripting language not using perl-compatible-regular-expressions.
I don't think that any significant portion of them could be claimed to have been specifically funded to directly meet any sort of proprietary requirement. Of course, they have been used for myriad proprietary purposes, including in some cases surely, by those people who actually developed them.
Re:innovation
ziggy on 2003-04-14T16:05:45
I have to say no. To a reductionist, it's mostly mimicry.How about Perl's regular expression extensions?I'm looking for something that was truly big, fresh and new ab initio that came from open source. The big critique about open source is that it mimics true innovation that starts in the R&D lab (or university, or gov't funded project). The point I am trying to dispute is that open source may be a capable petri dish for creating commodity software, infrastructures or copying known solutions, but fundementally incapable creating something totally new from scratch.
A lot of the work behind regexes comes out of work done at AT&T Labs (UNIX), as well as from Griswold and Griswold (SNOBOL, Icon). Perhaps there was something new in trying to shove lex and yacc into a simple regex engine, but in essence, that was just shoving existing ideas into an body of code.
To be fair, this line of thought reduces all innovation to a small set of truly big ideas. It isn't unfair to open source in that respect; instead, it characterized virtually all software to mimicry to a greater or lesser degree. By this metric, there is approximately zero innovation found in any version of Mac OS (including NeXTSTEP and Mac OS X), or any version of Windows (except possibly ODBC or the 16-bit/32-bit thunking layer, but I wouldn't swear to it).
Truly innovative, and really big ideas include PostScript, Smalltalk, Ethernet, TCP/IP, FORTRAN, the UNIX philosophy, and OS (the original operating system discussed in The Mythical Man Month). All of those came out of big corporate R&D departments: Xerox PARC, Bell Labs or IBM. (Although PostScript can be reduced to a special-purpose FORTH, there is something new and innovative there. Similarly, Smalltalk is much more than a warmed over Simula.)
Re:innovation
pne on 2003-04-24T12:23:43
They are good enough that most other languages have been moving to adopt them, and the pcre (perl compatible regular exporession) project developed an implementation. (After Perl 6 comes along, Perl will be the only major scripting language not using perl-compatible-regular-expressions.
:-) Fear not; you might get perl-compatible-regular-expressions in 5.10--at least, when Hugo talked about taking a good hard look at the regular expression engine last YAPC::Europe and I asked him whether he'd consider using PCRE he said yes...
:) Re:innovation
barbie on 2003-04-29T08:55:49
Anything that could be traced to a starting point of early Unix systems shouldn't be considered, as it was all for commercial gain. The instigators of Unix & C (Ken Thompson & Dennis M. Ritchie) were both collaborators on the MULTICS project, which Bell Labs pulled out of but inspired KT to develop UNICS (later renamed to UNIX, before the authors agreed it should be Unix) and DMR to develop C. Out of each came grep, lex, sed, awk among a whole host of tools that Bell could market as the killer apps for their new OS.However, Regular Expressions go back even further. Their origins begin in the 1950s with a paper by Stephen Kleene (also) called "Representation of events in nerve nets and finite automata" (1956)
I happen to be doing a RegEx talk tomorrow at Birmingham Perl Mongers, so I've done a bit of research
;) Re:innovation
jmm on 2003-04-29T13:48:46
Commercially-supported I'll agree with, but not for commercial gain.Bell Labs was in those days a real research organization, with relatively free rein to play with developing whatever they liked. Because of anti-trust negotiations, Bell was under court order that prevented them from getting involved commercially in the computer industry. Unix was not available for sale for a long time after that - the only way it was available in those days was to visit the Labs, and take some people out to lunch while "accidentally" leaving a tape on the tape drive and later discovering that it had been over-written.
Steve Johnson was a member of the Labs at the time - he spent a year's sabbatical at University of Waterloo in 72/73 or 73/74. The regular expression design from ed was used for an editor "qed" written at Waterloo, there were also derivatives of yacc that came with him, and the language B (an alternative derivative of BCPL, whence came C).
The Labs was not a commercially-oriented organization at that time. It has changed a lot since then, of course.
Re:innovation
ziggy on 2003-04-29T14:40:20
The (misguided) assertion I am trying to refute (on its own terms) is that all innovation needs a corporate environment to occur. The core premise is that you need a sizeable profit stream to fund a research lab, which limits research labs (and therefore all innovation) to a select few giant corporations. By extension, university labs fit the bill as well. This statement is phrased in such a manner to assert that open source (1) has never and (2) will never come up with something truly innovative like the mouse, multitasking or object oriented programming. Therefore, by its very nature, open source is limited to mimicry, not innovation.Commercially-supported I'll agree with, but not for commercial gain.Bell Labs was in those days a real research organization, with relatively free rein to play with developing whatever they liked.
Back in the day, neither Bell Labs nor PARC nor IBM's T. J. Watson Research Center were run to directly turn ideas into commercially profitable products. The intent is irrelevant here -- each of these labs were funded by a hugely profitable company who decided it was beneficial to put a bunch of smart people together under one roof to think (for them).
