KDE's Nepomuk Doesn't Seem To Have A Future

[omer666]

You spent 17 million Euros on a Semantics Web set of frameworks, for KDE? Who had the authority to waste that sort of funding? You've got an agning Window Serving environment, the likes of The GIMP, Inkscape, Digikam, Blender all of whom with that infusion would have made a much greater financial boon to Linux, never mind Scribus [another obvious duh] and you folks pissed roughly $25 million+ on fucking Nepomuk?

Too goddamn hilarious.

Oh yes they did and you know, it's not even close to the dumbest decision they are able to make. I do think that our EU may still become something great, but this kind of things happens so often that it's depressing, and as a matter of fact, it could be possible that most of this money actually went to the commission, and not to developpers themselves. But this is symptomatic: people don't care about EU, so EU doesn't care...

[CrvenaZvezda]

Seriously? Do you guys really believe that the EU gave the KDE project ?17 million just to play with? Really? Like there is no one else involved and benefitting from this project? KDE was merely piggybacking on something way bigger than themselves.

[liam]

Just forgot something quite important:

Baloo, now, is architecturally similar to Tracker, but it does more than Tracker, since it is designed to reach feature parity with Nepomuk. All this is done while eating significantly less CPU than Tracker (a great effort was spent optimizing the indexers, since Nepomuk was so slow, and Baloo simply reaps from that).

The question is: Will the Tracker devs join Baloo? Will they refuse to cooperate with Baloo? The reverse question has a simple answer, Baloo does more than Tracker, so Tracker cannot replace Baloo, but Baloo can, theoretically, replace Tracker.

Tracker's development has really slowed down since the fall of Nokia and the ending of a few contracts so I'm not sure how much help they could provide. The main problem with tracker is there substantial number of joins it performs, but they are necessary if you want to support the kind of rich query tracker allows you to do.
BTW, what can baloo do that tracker cannot? You would've thought the Jolla folks would've gone with the qt solution rather than tracker if it was at all possible.

[GreatEmerald]

Yea, Nepomuk in the latest versions of KDE has been good enough that I don't bother disabling it now. And I do enjoy fast searches. Anything improving it even further is good.

Well, in web browsers is very easy and convenient to avoid just that: Use a "incognito" session or delete everything by pressing ctrl+shift+del.

It would be faster to just create a new user.

Gives you deep searches, Boolean and Regular Expression searches, date and proximity searches - stuff you really need. Along with the industry's best preview pane. And it's incredibly fast, although it will hog your CPU if you are searching 10's of thousands of files at a time.

In other words, it's a GUI frontend to `find`?

[Temar]

So, no, guys, Nepomuk isn't gone. It is going to be replaced by something that does 80% of what the original Nepomuk did, using 20% of the resources (a good twist in the 80/20 rule)

Why does the KDEPIM community always replace components when they are not yet finished? Nepomuk and Akonadi were forced onto the user's desktop in an unfinished state and now you are doing it again. Didn't you learn anything?

Many people have Akonadi disabled, so they won't mind. There are people however, who actually use Akonadi and now they get a replacement which is only 80% feature complete? I really hope those 80% features are at least stable or KDEPIM will be running into the next disaster. You guys really know how to annoy your users the most.

[curfew]

The KDE devs should be aware of this (for example: what if I share my computer with other people, and I don't want to create another account for them? KDE applications store a LOT of personal data in the system...). I gotta admit, though, I'm not very knowledgeable about how these apps work.

You should be aware of that you're doing it wrong and that it's got nothing to do with what the devs should do. If you need to share your computer with others and cannot be watching over them, then you could easily create a visitor account or so, and just keep it hidden for the most of the time, if seeing it in the login manager annoys you.

[curfew]

You spent 17 million Euros on a Semantics Web set of frameworks, for KDE? Who had the authority to waste that sort of funding? You've got an agning Window Serving environment, the likes of The GIMP, Inkscape, Digikam, Blender all of whom with that infusion would have made a much greater financial boon to Linux, never mind Scribus [another obvious duh] and you folks pissed roughly $25 million+ on fucking Nepomuk?

Too goddamn hilarious.

Dumbass. The money was not given for developing an application per se, it was dedicated to developing and researching of semantic desktop paradigm. In that sense it's irrelevant what the actual code looks like, more important is what "intellectual property" was gathered around it.

[[Knuckles]]

The project probably wasn't all KDE. Most of it wasn't KDE. But the goal of research, is *research*. Sometimes you just learn that something is infeasible, and you identify issues, and you propose other options. That's why it's research, it's supposed to fail most of the times.

[yoshi314]

Neposuks and Akoncrappi. Goodbye and riddance. Damn bloatware that they tried to shove down our throats.

can't really say for everyone, but you could build kde with those features disabled on gentoo. So it's not like it was obligatory.

[caligula]

I think the Nepomuk works on wrong abstraction level. Modern filesystems support metadata for files. In fact there could be all sorts of things like per-file compression based on the metadata. How to achieve this? Instead of MySQL use SQL VFS built in to kernel and connect to desktop via kdbus and let systemd start it. This way it's desktop independent.