[tex-live] TL13 status
zdenek.wagner at gmail.com
Wed Apr 10 13:15:57 CEST 2013
2013/4/10 Markus Kohm <komascript at gmx.info>:
>> but another factor is the fact that the design of CTAN
> I don't think CTAN is the limiting factor. Manpower is the main limiting
> factor. The "team" of most packages consist of exactly one person, who is
> software designer, manager, programmer, author of the manual, editor of the
> manual, major beta tester, test facility manager, release manager, support
> manager and support executive assistant. And sometimes this person has to
> become even the legal division of the "team".
> Several of those one person teams provide more than one package and therefore
> have all these jobs more than once. Several of us find it already hard to have
> a live beside our OpenSource engagement.
Moreover, upload to CTAN and addition to TeX Live, MiKTeX (and maybe
other distributions) are asynchronous. If a distribution maintainer
asks me to incorporate just fxes to an arbitrary release of my
package, not the new features, I could start a branch and using the
log file port just the fixes. However, such branches will have less
tested and will be less stable than the latest version.
Requirement of adding just bug fixes may be counterproductive. many
years ago CentOS 4.0 decided to include ghostscript 7.x and during the
live of 4.x versions do not update to 8.x. Using ghostscript 7.x in
OS/2 I found and reported a bug. At that time 7.x was obsolete, so the
bug was fixed in 8.x only but years after that CentOS 4.x still
distributed that 7.x version with a known bug. I did most of my work
on CentOS but I knew that I must not use its ghostscript but use the
8.x version from OS/2.
> Providing and supporting a feature and an additional bugfix branch is simply
> unrealistic in this situation.
> BTW: CTAN is almost such a tiny team too and TeX Live team isn't larger a lot.
> Both teams are to small to spend time with "nice ideas".
More information about the tex-live