>Ravi, clucking about my youthful urgency, says:
>
>>before Mozilla, and other such projects that came about in the 90s, were
>>the GNU tools, such as 'gcc'.
>
>There was a time when gcc was the center of the Free Software universe:
>compilers were otherwise a Big Business (hated by one and all, speaking of
>youthful urgency), and they all seemed to share the vice of being
>simultaneously big, slow, and buggy. gcc, like Michael Jordan in the
>early 90s, broke onto the scene and immediately dominated: it was free, it
>generated better code faster than any of the closed, for-pay compilers,
>and it really gave a lot of encouragement to the whole idea of What Was
>Possible.
more excellent reasons why it's important to read Dreaming in Code! There's a chapter on Python and resistance to it because it was seen as a mere scripting language -- created to do an end run around the conventional edit-compile-run process.
>But: it's a singular example, and was quickly overtaken by advances in
>compiler technology that made it into the for-pay compilers, and gcc has
>never really recovered. It is no longer anywhere close to the best
>compiler out there, and it is plagued by the very virtues that it
>espouses; as an example, because "anyone can" make changes to it, it has
>accumulated baggage of dubious value while growing ever more brittle,
>making a break from its past ever less likely.
And another one! He talks about this problem where you can't reign developers in. One feature after another.
another most excellent chapter was on the tendency for programmers to never get started, because they will spend all their time making the perfect tools they need to get the work done.
I see that happen *all* the time.
shag