16 Apr
2011
16 Apr
'11
3:13 p.m.
It was thus said that the Great Alexander Schreiber once stated:
On Wed, Apr 13, 2011 at 04:33:17PM -0600, Richard
wrote:
Agree. At work [1] we had two projects that needed a DNS resolving
library and settled on C-Ares (one of the half dozen DNS resolving libraries
out there). Total crap. We needed to hack in support for the DNS record
type we were dealing with [2] and even then, it's networking model did not
match the rest of the programs (there were two projects that needed it). I
spent a weekend writing my own DNS resolving library and released it as open
source [3], thus improving the lot (and by using my DNS library, the runtime
memory usage in one program dropped from 15M to 400K).
Also at work, we get to use a commercial SS7 stack that was first
developed in the late 80s. It too, is total crap (the documentation says
it's thread safe, but it's not; what little source we do have was so badly
written that a simple (but tedious) change improved performance) but since
we don't have the source code, we can't fix it. And our company paid an
insane amount of money for this "commercial quality" code base.
And our lead developer, who is NOT a fan of open source for the most part,
constantly bitches about the "kwalitee" of the C/C++ compiler we're using
(and wishes we could use GCC).
-spc (Fun times ... )
[1] implementing the call processing features (SS7) for cell phones.
[2] NAPTR (RFC-1348)---I did not know DNS was used in call processing
for caller ID information.
[3] http://www.conman.org/software/spcdns/
It's in C99 [4]
[4] Which is more than 10 years old, so probably on topic for this list
8-P
In article <9AADCE11E0DD44EA9EAB11392540FA36 at ANTONIOPC>,
<arcarlini at iee.org> writes:
Ok, _that_ claim is, admittedly, silly. There _is_ a lot of very high
quality open source out there, definitely, It such isn't the majority
of open source code in existence. Hey, anything written in PHP already
has a 99.9% chance of being utter crap to begin with ;-)
But the upside is, again: with open source, even if the code is crap, you
can actually take a look at the source, see that it is hopeless and move
on to something that works (or build it yourself) instead of buying lots
of duck tape, baling wire and chewing gum to keep the expensive "enterprise
solution" somehow from _visibly_ shitting over itself on a regular basis. If you mean "a significant amount of stuff
out there is essentially
unmaintainable"
you are almost certainly correct. But why is that such a big deal?
Its not a big deal to me personally, but its important to keep in mind
when people make the claim that *most* open source code is high
quality.