thoughts on Cecil/Vortex packages
[Background: Cecil is a pure object-oriented language with
multimethods and other nice features; Vortex is a compiled
implementation thereof.]
I am strongly considering dropping this ITP. I realize that I could
also change it to an RFP, but am uncertain that it is of sufficient
user interest. Why?
* Stagnation: although upstream still exists, they haven't pushed out
a release since early 1999; this does not bode particularly well.
* Few users: Although Cecil is a very nice language, it seems to be
primarily of academic interest; AFAIK, the only major Cecil
application is the compiler. As such, I suspect it will mostly be
dead weight on the servers.
* Resources: on i386, the compiler alone (without any supporting files) is
a 27M executable. A full build is also rather time-consuming; my
estimate is ~10 hours on my 256M Celeron-500. Granted, I am
arranging things to spare the autobuilders from most of this work,
but I'm still not convinced it's justified.
* Fit: The upstream source doesn't have anything resembling "make
install"; I'd basically have to stick everything into a big
directory under /usr/lib, add some wrappers in /usr/bin, and hope
that nothing still tried to write into the tree. (Upstream seems to
assume a shared playground, which is fair enough for research
environments but not really appropriate for Debian.)
--
Aaron M. Ucko, KB1CJC <amu@mit.edu> (finger amu@monk.mit.edu)
Reply to: