Re: why so many frameworks? Can you just put all best ideas in 1?
"Tor Iver Wilhelmsen" <jadedgamer@hotmail.com> wrote in message
news:umz7gao5a.fsf@hotmail.com...
"Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> writes:
B) Investigate a dozen packages each of which claims to provide a
good XYZ implementation (in some cases wrongly, in others
correctly); select one and use it.
Well, people don't seem to mind doing that when it comes to buying a
car, for instance... :) In fact, people make choices all the time,
because in EVERY market there is choice unless there is a (bad)
monopoly. Why should software be any different?
Let's take CVS versus Subversion: Apperently, the makers of CVS didn't
see a problem with the annoying fact that moving a class to a
different package meant deleting it in one place and creating it anew
in another, losing revision history in the process. Plus that binary
files were stored whole for each version instead of just as a binary
diff.
Now, in "your" world, Subversion (which fixes both of those issues)
would not exist because everyone would shrug and submit to the
limitations of CVS. (In fact, why should CVS exist there? Since
everyone uses Windows, everyone should just use Visual SourceSafe from
"choice-is-bad"-favorite Microsoft.)
Actually, I think in a single-choice ideal world, it's CVS which
wouldn't exist, rather than Subversion. Ideally, you'd have exactly 1
choice, and it is the most optimal choice of all possible choices that could
have existed. So actually, you'd probably have something better than
Subversion too. Some sort of source management software crafted by God, and
handed down to us human, which is so great, no one would ever want to use
anything else.
Unfortunately, that's not how our world is, and we don't have software
written by God. We have a bunch of imperfect software, some more imperfect
than others. And now we have to spend time, money, and effort, researching
which software is the least imperfect for us. And to make matters worse, the
level of imperfectness varies from usage-scenario to usage-scenario, so it's
not like someone could do the research one, determine which software is the
best, and now we all know that for the rest of time to ignore all other
software package. Nope. What's best for you might not be what's best for me.
Which means if you spend 2 weeks researching the best XML parser for you,
I'd have to spend another 2 weeks researching the best XML parser for me; I
can't re-use your research.
- Oliver