On Fri, Apr 29, 2011 at 11:12:55AM +0200, dexen deVries wrote:
> 
> qmake (Qt's makefile generator) is mostly reasonable IMHO. consists of one 
> program (the qmake) which reads a rather simple project description 
> (<<myapp>>.pro) plus a bunch of platform description files 
> (/usr/lib{,64}/qt/mkspec/<<platform>>/qmake.conf + whatever it includes) and 
> outputs reasonable makefiles.
> 
> at any rate, the supposed replacements for autoconf/automake aren't shining 
> examples of engineering either -- usually big & complex. i guess it's more 
> about mindset (``let's check every itty-gritty detail and let's abstract away 
> differences between platforms'') than the problem space, thou.

The problem is not in the tool per se---R.I.S.K., used for KerGIS and
kerTeX (and others with no public version), is an example---but with the
programmers.

If programmers knew what they are using (C89 or C99 and that's all; or
POSIX etc.), the problem would be easily solved---these are the
cases "solved" by R.I.S.K.: programmer must know.

If the tool must "guess" what the program is using; furthermore if
for viral purpose and by "educational" repeating the wannabee
programmers are told to not care about standards, because GNU's
Not Unix and POSIX is bad, but use every chunk blessed by the GPL...

I don't know if there are black holes in the nature. But for sure mob
programming has managed to create computer ones; projects so bloated
that they are absorbing all the resources around with an emitted service
dimming more and more.

I'm finishing the integration of MetaPost in kerTeX (one auxiliary
program to fix and I can start testing), and I will have spent less time
from a very scarce free time redoing everything (distribution side) than
people trying to make TeX Live compiling for their plateform. (The
source code is the Medusa: you look at it and you are awed. That was
the aim.)

I claim this is a kind of lesson. (Same goes for GRASS -> KerGIS even if
nobody cared when I did it.)
-- 
        Thierry Laronde <tlaronde +AT+ polynum +dot+ com>
                      http://www.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89  250D 52B1 AE95 6006 F40C


Reply via email to