> What would be cool though is to know what kind of hardware are people using > I know that currently we have a closed database which doesn't help anybody > as the info. is requested repeatedly whenever a new bug is being registered.
It's a set of nasty problems: 1) Device and/or hardware identification (not always easy) 2) Filtering ambiguities and inaccuracies (take a look at smolt, how many "O.E.M Fill in here" motherboards are there?) 3) Supporting and structuring meta information, correct names, correct manufacturers. Even so far as the list features and metrics, icons and pictures. 4) Centralising and analysing, you can't do this until you have 1 and 2. Then you need everyone to be running on the same standard. Unfortunately, smolt and many other ideas have jumped strait to 4 without considering that perhaps 1 and 2 are important to sort out first. My own project, which I've shut down since, taught me a great deal about some of these weird and wonderful issues. In conclusion, I believe that the way forward is to fix these issues: 1) Gather dmi, bios, i2c and other system information sources with kernel level drivers able to deliver this information via sysfs or some other mechanism. 2) Populate and structure Hal (and it's replacement) in such a way to accommodate this "none functional" information sources. This includes being able to distinguish between a hardware object and a device object. 3) Be able to generate a unique hardware class ids that allow destinations, even when normal class ids are the same (examples of scanners, phones, motherboards with the same usb or pci ids but different metrics) Only then can you start fixing this problem properly. Regards, Martin Owens -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
