On 2019-10-17, 08:39 GMT, Jan Engelhardt wrote:
PackageHub and rpm-catalog are counterexample of this opinion.
Just to explain my position: 1. In my opinion (and I am afraid I am not alone) Groups are not used anywhere it matters. Yes, I have to admit I have no idea what PackageHub and rpm-catalog are (first just from the name looks like some kind of website, right? second will be probably some software). If one looks for some piece of software, who won’t just run search "openSUSE typeofsoftware" in the Internet Search of your choice? Is there anybody who really cares for PackageHub and rpm-catalog (and Gnome Software, for that matter?) 2. Because of #1, for most package maintainers Group label is nothing anybody cares about. It used to have to be filled with something which gets through rpmlint and that’s how far most package maintainers IMHO cares. 3. Given #2, groups are such mess, no user uses them for searching for their software, because nobody knows whether Network Manager is in Productivity/Networking/General, System/Networking, or something else. See below on problems with tree hierarchies. 3. Given #3, no sane software developer uses them in their software, so -> #1 Yes, of course, possible solution would be to make an Herculean effort to fix groups to be perfect even against the will of all package maintainers. The problem is that you still haven’t persuaded software developers to use it. I think it is a prevalent notion among people who care about classification, catalogues and similar stuff, that the tree hierarchy is inherently flawed and insufficient to classify which at least approximately corresponds with the real life. Even software is just too complicated to be squeezed into simple hierarchies and too complicated hierarchies are too complicated to be useful for the end-user. See for example https://www.w3.org/Provider/Style/URI The answer to this problem with strict hierarchies was the movement of folksonomy (https://en.wikipedia.org/wiki/Folksonomy), which basically means tags and search through them. It is better than strict hierarchies, because suddenly multiple tags can deal with the ambiguity of the classification, but the problem is (as anybody who tried to find a story on https://archiveofourown.org/ can atest), that unbridled generation of tags leads to the endless duplicates. Limited tags are again very hard to manage and require constant maintenance. However, I don’t think that we are limited just to strict hierarchies or tags. There is already one system, which is well maintained and easily utilized. Every package has Description field, and with zypper se -d (or https://archiveofourown.org/) this can be easily searched. And that is the reason I believe whole effort to create working Group: or Tags: is waste of time and effort leading to nowhere. Best, Matěj -- https://matej.ceplovi.cz/blog/, Jabber: mcepl@ceplovi.cz GPG Finger: 3C76 A027 CA45 AD70 98B5 BC1D 7920 5802 880B C9D8 Basically, the only “intuitive” interface is the nipple. After that, it's all learned. -- Bruce Ediger when discussing intuivity of Mac OS http://groups.google.com/group/comp.sys.next.advocacy\ /msg/7fa8c580900353d0 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org