Homebrew toolkit

[Seraphimsan]

I find that I'm much more comfortable with my code over some libraries some one else wrote. Which is why over the last 4 years I've been putting together my own, pretty universal, library. Do you guys find that you do the same? Opt out to write your own standards over the easier downloading of a file, for the sake of comfort? Or am I the only one foolish enough? :B

[Return 0]

I do the same thing. I just feel more confident with libraries written by me when I have the time. Of course some things it just doesn't make sense to rewrite.

[Seraphimsan]

oh god yea, no point reinventing the wheel :P

[helios]

Do you guys find that you do the same?

Depends on whether I'm interested in the subject and whether there are options available. The last wheel reinvention I did was a simple ZIP class (the ZIP specification is very well written and easy to understand, by the way) to handle archive splitting and a couple compression schemes.

for the sake of comfort?

It's often less about comfort and more about robustness and safety. The more a piece code is used, the faster its bugs are found. When you finish reinventing your wheel, it's been used exactly zero times, while the wheel you could have used may have been used millions of times over the course of many years.

[z05DSL3A]

For me it comes down to cost and/or licensing. If the commercial product cost to much or the open source license is incompatible then it often leaves little choice but to make your own.

[jsmith]

I find that I'm much more comfortable with my code over some libraries some one else wrote.

That's because the author of the code knows it best. You're more comfortable with your own libraries
because you understand yours better -- you know exactly how it works (hopefully...). Third party
libraries are black boxes -- to some extent even if you have the complete source code.

And then there's the "I can do it better" syndrome. In many cases, if you're just pulling down code off
someone's home page (or, sadly, even off the Sources section of this website), chances are, yes, you can
do it better, because it wouldn't be hard. But I wouldn't think of trying to one-up any of the boost libraries
or one of the more popular graphics libraries. While I'll understand my own implementation better, it
definitely won't be better.

[chrisname]

I have yet to master the art of reading other people's code.

Then again, when you see horrors like this:
http://warp.povusers.org/OpenLetters/decode_mb_cavlc.txt
it doesn't exactly make you want to try to learn...

Edit: this one is worse: http://warp.povusers.org/OpenLetters/fill_caches.txt

[firedraco]

cbp = 0; /* avoid warning. FIXME: find a solution without slowing
                down the code */

Use a constructor maybe?

[chrisname]

I think this is the most important comment in the whole of the second file: //wow what a mess, why didn't they simplify the interlacing&intra stuff, i can't imagine that these complex rules are worth it