The Central C Archive Network?

Alias on 2007-11-28T22:47:15

Some interest has been shown by a few People That Matter at OSDC for the idea of cloning a CPAN similar system for C programmers, taking lessons from what worked and didn't work for CPAN.

I'm curious about your feedback on the creation of such a thing.


Skeptical

educated_foo on 2007-11-29T05:01:14

I'm skeptical. When I turn to CPAN, it's often because I need to quickly satisfy some short-term need mostly unconnected to my existing code. On the other hand, I would guess that most C projects involve enhancing some long-standing infrastructure. This means there's a lot more room for incompatibility between CCAN contributions. It would be an interesting experiment, but I think too many people are locked into their own string libraries, hash tables, etc. for it to work.

Re:Skeptical

ChrisDolan on 2007-11-30T04:09:35

I don't know, it seems to me that Boost has made a big impact for one. It's everywhere in the C++ world. Sure, legacy projects have lock in -- that's true in every language -- but not new projects.

Re:Skeptical

educated_foo on 2007-11-30T05:31:45

The problem with Boost (and the reason I don't use it) is that to use one Boost library, you almost have to use them all. It's like adding a whole other language: you pay an enormous up-front cost for using a single feature, so it makes sense to avoid Boost unless you have to use many features.

Re:Skeptical

ChrisDolan on 2007-11-30T07:43:11

Well, maybe that's because they don't have a CCAN to allow cherry-picking features. :-)

Re:Skeptical

chromatic on 2007-11-30T08:02:09

That's a pity, too. From my understanding (as I try to avoid C++ in general), Boost makes the language much more palatable at the expense of violating the principal design principle: you don't pay for what you don't use. (Then again, I always hate writing code to convince the compiler not to do the wrong thing by default, so maybe C++ isn't very good at that rule itself.)

Re:Skeptical

educated_foo on 2007-11-30T10:11:26

Yeah, C++ may not be for you. I really like the language for its "pay only for what you use" and compile-time evaluation (template metaprogramming) features, but there is significant pain required for each.