On 07/02/2016 20:35, Mick wrote: > On Monday 08 Feb 2016 01:57:52 Andrew Lowe wrote: >> On 02/08/16 01:38, Mick wrote: > >>> But you should be able to stop it from indexing your files. If you go to >>> systemsettings/Desktop Search you can disable it. >> >> That's what I've come to the conclusion I'm going to need to do. This >> "semantic desktop", for me, is a pile of rubbish. I don't need it and >> don't want it, but the people at KDE apparently know better my needs >> than I do so they have embedded this stuff so deep you can't untangle >> it. It's just more gumpf that gets installed, for no other reason than a >> developers ego trip. >> >> Thanks for the comments, >> Andrew > > I think it was started by sponsorship via an EU project, back in the 00s, > when > Linux was seen as an alternative contender to the US MSWindows desktop > monopoly. To break into the corporate world the concept of a semantic > desktop > with a database back end was seen as key requirement. I don't know if they > were thinking of a corporate/federated database at the time, to share address > books and client info, but the concept was not managed carefully through a > staged release program. They dumped their half-baked semantic desktop and > KDEPIM nightmare onto unsuspecting users, who started losing their emails and > address books. This I recall was limited to one version of KDEPIM only, but > the credibility of the whole project was lost at that point. Thankfully KDE > has improved since. :-) >
and I was the dumb sucker on gentoo-user who did lose emails and address books But I think you are conflating strigi, akonadi and nepomuk a little bit. strigi always worked pretty good but it was a touch slow in early versions. That was sorted real quick with the usual code optimization and check-what-it-really-does developement cycle nepomuk looked good on paper but was horribly complicated and other dev had trouble with it. All of that is now mostly gone with Baloo (it's really a version 3 situation) akonadi was just a mess from the start. Another good idea on paper but I think the devs completely underestimated what this was going to take to work, in terms of man hours and cpu cycles. Right now it feels like a half-cocked solution without a problem to solve -- Alan McKinnon alan.mckin...@gmail.com
