libclc: thanks

  • Thread starter =?ISO-8859-1?Q?Bj=F8rn_Augestad?=
  • Start date
?

=?ISO-8859-1?Q?Bj=F8rn_Augestad?=

I'm writing to thank everyone who has been working on and contributing
to the libclc project. It's been a very educating experience and quite
entertaining at times as well.

I've come to the conclusion that the project isn't working very well.
The number of contributors is not very high, the number of downloads is
close to zero and supporting the DS9K is just too limiting (for me).

Unfortunately I see no way of solving these problems, and I've therefore
decided to leave the project and instead probably limit myself to
writing POSIX C.

But again, thanks to everyone who was a part of the project. Best of
luck if you decide to continue the development of libclc. If so, let me
know who'll be the new maintainer(I recommend Hallvard) so that I can
transfer permissions to him over at sourceforge.


Best regards,
Bjoern Augestad
 
M

Malcolm

Bjørn Augestad said:
I've come to the conclusion that the project isn't working very well.
The number of contributors is not very high, the number of downloads is
close to zero and supporting the DS9K is just too limiting (for me).
It was a nice idea. I think the problem was that you ended up trying to
compete with the standard library.
Another nit-pick was that the prefix clc_ was a little bit too long, making
code ugly to read.

However I think that the ng can have an influence, and it is a good idea to
use that influence to try to develop a library. The more people who use a
library, the more incentive there is to make it better, and the more
incentive for other people to learn it.
 
R

Randy Howard

malcolm@ said:
It was a nice idea. I think the problem was that you ended up trying to
compete with the standard library.

Even that is/was not possible. The standard library isn't restricted to
being implemented in pure ISO C. In sense, libclc is doomed if it can
never provide a general API, with platform specific details hidden
underneath. Doomed is probably too strong. EXTREMELY LIMITED is
probably more appropriate. It is sort of absurd to try to extend the
standard library, while doing so with a more restrictive set of
coding restraints than it itself requires.

There is clearly an argument that it is valuable to have a library that
runs on a wide variety of operating systems and exposes traditionally
platform dependent functionality through a single API. Lot's of people
have tried. I think this really points back to the standard library
being somewhat incomplete. Why is there so much wheel reinvention going
on amongst C programmers? It's not the language syntax, it's the lack of
a standard library with functionality less ancient than the PDP-11.

Unfortunately, it is difficult to construct real-world applications in
pure ISO C. In practice, most such applications could rationally be
constructed with a very large % of the code in conforming C, and a much
smaller portion of "outside the box" C. Doing such by abstracting the
functionality on top of platform specific details in a library could have
a lot of value, but could never meet the constraint of "100% pure ISO C"
required here. So, in a real sense, everybody loses because of all the
duplicate work. Perhaps the solution would be a lib99percent that
working on everything in general use except for the DS9K. That would
in fact be far better than nothing at all, but it would be OT here.
However I think that the ng can have an influence, and it is a good idea to
use that influence to try to develop a library. The more people who use a
library, the more incentive there is to make it better, and the more
incentive for other people to learn it.

Unfortunately, not even the standard library's implementation could be
discussed in this newsgroup, because it is forced to do platform dependent
things, and that is OT here.

Thinking on it again, perhaps "doomed" was the correct term after all.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,744
Messages
2,569,484
Members
44,903
Latest member
orderPeak8CBDGummies

Latest Threads

Top