From: mad@math.keio.ac.jp (MAEDA Atusi) Subject: Re: Why so many binary distributions (was Re: Nethack 3.1 is out! Who will port it?) Date: 4 Feb 1993 11:49:00 GMT
In article <1993Feb3.231836.6109@midway.uchicago.edu> goer@ellis.uchicago.edu (Richard L. Goerwitz) writes:
>jwinstea@fenris.claremont.edu (Jim Winstead Jr.) writes:
>>I can't imagine people who have all that free time to pull zillions of
>>sources down from the net and compile them, fixing any porting
>>problems as they go, choosing options, and applying patches (once you
>>find them).
>
>This ain't the way a good OS archive works. Once and for all: If there
>is a standard source distribution, then diffs are all that need to be
>supplied for the port. You download the source tree once, then apply
>diffs as they come in. The first distribution is large, but then subse-
>quent downloads are trivial. And you retain control over the configur-
>ation. Doesn't waste time. Saves it.
Patches are not so small in some cases. They occasionally become larger
than the original source (e.g. Linux kernel).
How about distributing sources *and* binaries, and let people choose
they prefer? For popular softwares ported to Linux, just binaries
(plus diffs to compile them on Linux, if needed) would suffice, because
you can get sources from many other archives. I think most people who
write or maintain software for Linux try to distribute this way.
This way, those who want sources can compile them by themselves,
taking full control over software they're using, while people that
don't have enough time or disk space can be happy, too.
;;; Keio University
;;; Faculty of Science and Technology
;;; Department of Math
;;; MAEDA Atusi (In Japan we write our family names first.)
;;; mad@math.keio.ac.jp