Hi, in this bits: 1. Scientific citations 2. Prospective packages in VCS 3. Proposal to move Blends code from SVN to Git 4. General situation of Blends 1. Scientific citations ----------------------- Due to the effort of Charles Plessy to gather machine readable information about upstream[1a] there is a set of references about scientific publications available. This is an ideal situation for scientific Blends like Debian Science, Debian Med and DebiChem because it opens the chance to strengthen the connection to upstream authors of software by advertising their scientific work. While the web sentinel provides a way to inject such citations directly inside the tasks files of a Blend this had several drawbacks which are now solved: 1. Citations do not need to be specified in every single task file in case a package is useful in more than one task. 2. The reference information is directly attached to a package and can be used there for other purposes as well. 3. The infromation of debian/upstream files is parsed in Vcs and injected in UDD[1b] were it can be used quite flexible. Technically the Upstream Metadata Gatherer[1a] parses the packaging VCSes once a day and creates a tar archive which in turn is parsed by the UDD[1b] bibref gather. Once the data are inside UDD there are several chances to use these also independently from the usage inside Blends sentinal pages. For instance you can create a BibTeX file of all citations inside Debian which was explained at Debian Science mailing list[1c]. [1a] https://wiki.debian.org/UpstreamMetadata [1b] https://wiki.debian.org/UltimateDebianDatabase [1c] https://lists.debian.org/debian-science/2012/04/msg00077.html (and other mails in this mailing list) 2. Prospective Packages in VCS ------------------------------ The Blends tasks files may contain so called prospective packages[2a]. On one hand these entries are informing users what software the packaging team is working on and on the other hand these entries ensure that metapackages are built nicely without changes once the package in question has reached the Debian mirror. It has turned out that some tasks files became quite long because of these prospective packages entries and on the other hand information of packages that are just prepared in VCS of a packaging team is duplicated. To enhance this situation a parser was written that gathers information of all machine readable files in VCS of the Blends packaging teams and injects this information into an UDD[1b] table. From there it is read when the tasks pages of the web sentinel are rendered and thus the extra information about prospective packages in the tasks file becomes redundant. The raw archive of this information can be found here[2b]. [2a] http://blends.alioth.debian.org/blends/ch-sentinel.en.html#s-edittasksfiles [2b] http://blends.debian.net/packages-metadata/packages-metadata.tar.bz2 3. Proposal to move Blends code from SVN to Git ----------------------------------------------- A proposal was posted on the Blends listi[3a] to switch the repository maintaining Blends code from SVN to Git. If you are involved in maintaining parts of the code and are not happy with this move this is your chance to raise your voice now because there are chances to leave some parts which you might be interested in untouched if your personal workflow would be spoiled by this move. [3a] https://lists.debian.org/debian-blends/2012/10/msg00002.html 4. General situation of Blends ------------------------------ Debian Accessibility Team works constantly and successful to enhance accessibility features of Debian BrDesktop Became a bit silent in the last time - some bits of the drivers might be interesting. DebiChem Team works constantly on maintenance of existing and new packages which are useful for Chemists. Debian Edu Has released Debian Edu based on Squeeze and made good progress to gather new team members. Debian EzGo Became a bit silent in the last time - some bits of the drivers might be interesting. Debian GIS Team works constantly on maintenance of existing and new packages used in Geographic Information Systems as well as OpenStreetMap. It might not harm if the team would be more visible in and outside Debian because there should be potentially a lot of more users and developers. Debian Junior Recently some new light might be at horizon to revitalise the project - at least the mailing list[4a] recived some promising mails. Debian Lex Did not really started ever, recently there was some activity on the mailing list to vitalise the project a bit. On the other hand it was suggested by Debian web team to state on Debian Lex website that the project is discontinued. Debian Med Constant progress to package even more software for medical care and biological research as well as a constant increase of developers in the team. The progress was boosted by the yearly sprints (see last Debian Med bits in March[4b]) Debian Multimedia Team works constantly on maintenance of existing and new multimedia packages. It might not harm if the team would be more visible in and outside Debian because there should be potentially a lot of more users and developers. Debian Science Constant progress to package even more scientific software which is not yet covered by Debian Med or DebiChem team. Also the number of tasks is increased. It might make sense to branch some Debian Mathematics or Debian Physics project from Debian Science to add more stucture to those topics and create more dedicated teams. In summary I would like to say that I do see good chances for improving some Blends as well as using the technique in other fields as well to make Debian more popular in additional workfields. [4a] http://lists.debian.org/debian-jr/ [4b] https://lists.debian.org/debian-devel-announce/2012/03/msg00009.html Thanks for your interest in Debian Pure Blends and feel free to read more about it[0] Andreas. [0] http://blends.alioth.debian.org/blends/ -- http://fam-tille.de
Attachment:
signature.asc
Description: Digital signature