On Mon, 2004-09-06 at 11:06, Richard Jones wrote:
> On Sun, Sep 05, 2004 at 03:20:56PM +1000, skaller wrote:
> > My personal take is: all 'make' based build systems
> > are fundamentally flawed because they're based on
> > the wrong concept. 
> 
> Can I stand up for 'make' (well, GNU make anyway) and point out
> that whatever its philosophical problems, it:
> 
> (a) works

No it doesn't. That's the problem. If it really did work,
why do imake, omake, ocamake, cake, and many other
such derivates exist?

It fails to be even marginally scalable -- for any significant
project it collapses almost immediately, and these days
it fails -- totally and abysmally -- to cope with 99% of
all C projects as well -- which is why those projects also
use M4, Auto* tools, shell scripts, make templates, Perl
or Python, and many other such tools.

Extend that to a major system requiring multi-platform
support, multiple target kinds (debugging, s
tatic/dynamic link ...) multiple developers, versioning,
documentation and testing, multiple languages,
etc etc and the collapse is total -- the tool becomes
an *impediment* to productivity. Mandating that the
whole of such a control system be written in Perl
plus a small set of external tools is a much better,
more expressive and scalable solution.

I'm not saying make isn't useful, I'm saying it isn't
a general purpose project build tool. It doesn't even
come close.

-- 
John Skaller, mailto:skaller@users.sf.net
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net



-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners