Keith Wesolowski wrote:
> ... a generic set of Makefiles that works on recent versions of common
> systems (Solaris, GNU/Linux, BSD, and HPUX) would actually be more
> useful...

This makes a lot of sense and seems, AFAICT to be a viewpoint with lots of
support. I also like your argument about gcc and Studio (e.g. it's
misleading to give a demotion only because an app compiles fine with gcc
but not Sun Studio).

So no doubt a revision to the experiment's rating system is needed.

On the topic of a machine-readable metadata repository itself and whether
converging on a single standard would be better than anything that could
occur in this arena... Of course, absolutely. But regarding your assertion
that the emergence of a widely admired and respected individual would
motivate the leaders of the existing camps to join forces...? Sorry, I
don't think that would have much impact on the motivation needle. (In fact
very little if any, IMO.)

> ... It's pretty easy to take a machine-readable
> repository and offer a human-readable front-end like this one...

By saying "_a_ machine-readable repository", you're missing a key concept
though. An _arbitrary_ number of machine-readable respositories (not just
one) can easily feed into a single human-readable front-end.

> Even in the cases in which autoconf gets it right, libtool will
> *ALWAYS* produce broken libraries.  So really I'd give a big fat -4 to
> any component that delivers a library and uses libtool to build it,

If you can't beat 'em, join 'em...? Maybe? In other words, I guess I don't
know (yet) what I think of this. Q: Does anybody have a sense (or a SWAG)
of what percentage of open-source apps and tools that deliver libraries
use libtool to build them?

Eric

Reply via email to