One of the reasons we have such long, loud, and inconclusive arguments about tools, methodologies, and the like is because we can't properly measure the effectiveness of such things. At the heart of this is the fact that we can't measure the productivity of development teams. Strictly the problem is that we can't measure their output - how much stuff gets cranked out.-- Martin Fowler, Measuring Productivity
Re:Sorta
tmtm on 2003-09-03T17:52:11
When most people say that "X is the best Y for Z" they're usually giving an opinion not a measurement.
Very few people, other than in jest, can assign a number of any type to these statements.
"Well, I'll give vi a 6, emacs an 8, but vim definitely gets a 9.3"Re:Sorta
pudge on 2003-09-03T18:07:05
Ability or willingness to assign numbers is not the only measure of "measurements." My car is bigger than my bike, but I have no idea by how much. Similarly, BBEdit is better than vi. Sure, I *could* get numerical measurements of the car and bike, but there's no need to.Re:Sorta
ziggy on 2003-09-03T19:03:19
Now you're overstating the case. You started out by saying that BBEdit was the best editor for you.Similarly, BBEdit is better than vi.That's not what Martin Fowler is talking about.
If we really thought hard about it, there might be a reasonable way to measure the productivity of the little things we do on a daily basis -- using text editors or write small programs. That's doubtful, and it's even more doubtful that we could measure project productivity: OK, now do the exact same project with the same team; use XP and no UML this time and see if you can finish it faster.
These types of tests have been done in the past, and about the only thing they can demonstrate is that the most productive programmers are roughly 10x as productive than the average, and everyone has the capability of writing roughly the same number of lines of code per day regardless of the language used.
Sorry, but you can't measure productivity in software. Not in any big, meaningful way.
Re:Sorta
pudge on 2003-09-03T19:11:56
Now you're overstating the case. You started out by saying that BBEdit was the best editor for you.
It is still what I am saying, necessarily. Since the way I measure such things must necessarily be subjective, saying it is better is necessarily saying it is better for me, unless qualified otherwise.
Sorry, but you can't measure productivity in software. Not in any big, meaningful way.
Of course I can, and I do it every day. But it is only big and meaningful to me.Re:Sorta
tmtm on 2003-09-04T20:38:30
You may not know by how much your car is bigger, but the point is that it's relatively easy to calculate, and get objective numbers that everyone can agree with. Without that it's not measurement, it's just comparison.
Numbers are the entire point of measurement.Re:Sorta
pudge on 2003-09-04T20:49:59
Numbers are the entire point of measurement.
No, they aren't. It is ONE of the METHODS of measurement. There is no one point of measurement, except maybe the generic point of "comparison."
Ever hear of a "measure of someone's worth"? We measure things like "worth" that all the time, but certainly without numbers.
