Python for VB6 programmers (was Re: Complexity of standard Python data structures)

Alex Martelli aleax at aleax.it
Thu Apr 17 16:53:48 CEST 2003


Drazen Gemic wrote:

>> People have said it in the past, the Python community needs to pick ONE
>> of the Python wxWindow GUI builders currently under development and
>> really promote the hell out of it.
> 
> I don't think it will ever happen. You people with microsoft background
> do not understand some things. In Python, PHP, Perl, Linux comunities
> there are no 'winners'. People just use software they like the most.

It may or may not happen, but SOME of the time some free program
emerges as the clearly dominating one in its niche -- others often
survive (not always...) but with very minor levels of effort and
investment going into them any more.  E.g., apache's dominating the
niche of "open-source general purpose webserver" -- there may be
others, but does the amount of time and effort going into their use,
development, documentation, etc, reach to even 1/10th as that which
apache attracts?

So, there aren't _always_ "winners" (but then, the same "not always"
holds for the proprietary world -- e.g., Delphi and VB keep surviving
and flourishing side by side, one against the other), but often there
are, even in open-source worlds.


> Some use this GUI builder, other use that, other do not use GUI builder
> at all, without bothering each other.

Actually, when two or more products (opensource or not) are used in
vastly overlapping niches, flamewars between devotees of each are
quite typical - not inevitable, but quite typical.  Part of the root
cause is in human sociobiology, part is in the 'risk', from the POV
of somebody who's invested hugely in one of them, that another will
"emerge as winner" and suck interest and effort away from that one.

> People do not want to be "guided" by some kind of superior authority
> that decides what is the best for everybody.

Some do, some don't.  Standards emerge, often by voluntary action
of people who are very much in the open-source community, because
in many cases there ARE substantial advantages to most everybody in
having exactly such a "superior authority" as a standards-making
body -- those cases are roughly all of those (and only those) where
the shared advantages by "networking effects" (in the economics
sense) surpass the disadvantages of relative resulting unitormity.

So (for example) you're perfectly free to build your own Linux
distribution from scratch and completely disregard any indications
from (e.g.) the FHS about what filesystem hierarchy "is the best
for everybody" -- in practice, very few people do, because the
advantage of compatible filesystem hierarchy layouts are quite
considerable to everybody, the compensating advantages of doing
your own idiosyncratic and quirky hierarchy quite small.

So, I think your assertions are way over-broad the way you put
them.  What situation applies to the specific case of GUI builders,
well, time will tell -- personally, I have no idea how that issue
will turn out... continuing fragmentation, or what degree of
consolidation.  But as you don't limit your assertions to that
specific issue, they can be seen as false in their generality.


Alex






More information about the Python-list mailing list