Hi Holger, My proposal is just to see if there can be a way to glue up all the bits n' pieces together. There is a lot of information scattered at multiple places, which if mined properly, can serve as one of the best resource. Debian Bug Reports contain a lot of valuable data. But is it connected ? Unless one goes to the BTS and scans for the logs, they can't find it. In the same way, all cool hacks and best uses of particual software is described at these places; mailing lists, blogs, irc et cetera. Can't there be a way to glue all these data to each other ? IMO, debian-devel and debian-user must be having great amount of data, which can be invaluable. But given their high volume, not much is mined. I was thinking if there could be ways to inter-relate data to each other. I was looking at Launchpad. I haven't used it much but am starting for one of the projects. Mark's ideas is the same. They are drastically different than the way we work today (or have been working for years), but his vision is the same, IMO. And that is to inter-relate all these data sources and create invaluable resource out of it. For Debian's current infrastructure: hit the nail on the head earlier in the forum * Forums: They should be linked to mailing lists also. Like if someone starts a thread in a forum, the same should get posted in the mailing lists. And the same would go for replies. * Wiki: I'm not sure if there is the relational model with other packages, in place for wiki. Is there a way to find all wiki pages related to kdelibs5, from the PTS page ? That would help. I think, data should go to just a single location. Mailing lists, forums, wiki, BTS - all should have a single database. All should be made inter- related. And then all should have tags, so that for focused interest, one could focus on tags. A maintainer could take Tags with BTS and the highest priority item. The benefit is that all the relevant data is inter-related and hopefully mined. To start with, it should begin in Debian and who knows, maybe some day all Free Software projects could follow the same model and be inter-linked. But these are just ideas. I have no implementation details apart from a half confident example of Launchpad. But I think they are in the right direction. Ritesh On Friday 08 May 2009 16:45:18 Holger Levsen wrote: > Hi Ritesh, > > thank you for your suggestion on how to improve Debian! Even though I'm > closing this bug on the assumption that it ain't useful to report arbitary > wishlist bugs about things which could be implemented to improve Debian, no > matter how sensible they are. > > This assumption is based on three main arguments: > > First, there are several places to collect such ideas which IMO are better > suited, like the Debian wiki or personal idea collections. > > Second and more importantly, if you want such a system, I think *you* > should implement a working prototype. For example, wiki.debian.net was just > done and then, as the Debian community liked+used it, it was moved to > wiki.debian.org. > > Third, yes, discussion in advance of doing such a prototype is useful. But > you don't need to file bugs to discuss. > > > And even though I agree with Neils points in his reply, please don't be > discouraged by all of this. If you think it's a good idea, by all means go > for it and show its a cool thing! forums.debian.net is probably an example > of something which many Debian developers don't (or didnt, when it was > started) consider particulary useful, but today it has many happy users :-) > > > regards, > Holger -- Ritesh Raj Sarraf RESEARCHUT - http://www.researchut.com "Necessity is the mother of invention."
Description: This is a digitally signed message part.