I am unable to find a clear distinction between an experimental and an optional
package.
The comments at the top of the list of 'experimental' packages
http://www.sagemath.org/packages/experimental/
make it very clear you are taking your life in your hands by using these.
The optional packages
http://www.sagemath.org/packages/optional/
gives no such warning, so one might infer that the quality requirements for
'optional' are higher than 'experimental'
But I do not see any list of the different requirements.
My particular issue is whether an optional package needs to build on all
supported operating systems, where it is appropriate.
Since sage 4.3.4.alpha1 builds without issues on 't2' and Sage 4.5 out in 3
months expected to support Solaris 10 on SPARC, should new optional packages
work on Solaris 10 (SPARC), unless there is good reason for them not to, such as
Valgrind, which does not support Solaris. (The same goes for all supported
platforms of course.)
I can appreciate given the warnings on 'experimental' packages that no such
requirement would exist, but I'm not sure about optional.
I think there needs to be clearly defined criteria for what makes a package
'optional' as opposed to 'experimental', as the wording at
http://www.sagemath.org/packages/optional/
http://www.sagemath.org/packages/experimental/
implies different levels of quality.
Other issues to address are
* Whether a vote is needed before a package can become optional?
* Whether a ticket needs to be opened before a package even become experimental?
(I think I know the answer is "yes", but there was a report recently of an
experimental package not having a trac ticket associated with it).
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
