Communicating Nepomuk's Benefits
Right now there are a lot of questions surrounding Nepomuk: What is it good for? Why does it take so much CPU / disk? Why do I need it if I don't plan on doing any socializing or semanticizing (?!) on my computer? This translates to a lack of enthusiasm, appreciation and even frustration amongst our users and even at times developers.
The thing is: all of those questions have good answers. Whether it's the performance improvements of Virtuoso, educating our users about how Nepomuk is being used by various "every day tasks" (such as indexing and relating email, contacts and other PIM data via Akonadi, all behind the scenes and with good perormance) or making it easier to figure out that it's quite easy to turn off the disk indexing parts if you want, these things aren't getting communicated very well yet.
There is a page about this on our Community Wiki. That needs to be fleshed out as needed, read by everyone who is doing communication on behalf of KDE and then shared as often and widely as possible.
As developers or technophiles with a deep understanding of these esoterica, we know the benefits of Nepomuk and get rather excited about it. Everyone deserves to feel that excitement, too. :)
Delivering Nepomuk's Benefits
This one is easy to describe, but harder to do because it takes effort. A lot of it over a period of time. Delivering on Nepomuk's benefits means using Nepomuk features in as many of our applications as possible in as many creative ways as possible so that people who use KDE can just start to take for granted that those features will be there in KDE applications. Think of it as the new copy and paste, maybe. :)
In 2010, if each of our application projects implemented even just one feature that used Nepomuk in some relevant way, even if it was completely behind the scenes, we'd end up with an amazing network effect as the user's data swirls around in the crunchy ontologies of Nepomuk.
I'll admit that we haven't done nearly enough of this yet in the Plasma Workspaces (we wanted to have some Nepomuk features in 4.4, but they got punted till 4.5 as they aren't quite finished to our satisfaction), so we'll be on this key quest like a dirty shirt in 2010. :)
(This article is part of the "Key Quests for KDE in 2010" series)
9 comments:
I think part of the issue at the moment is that KDE 4 uses a reasonable amount of RAM and probably many of the people that turn nepomuk off are people which hav 1G or less RAM. My new X200 laptop shipped with 1G of RAM and if it's between running nepomuk and a few more konq tabs probably would prefer the konq tabs.
I should also add, that on my desktop with 2G of RAM nepomuk is enabled and it is indeed very useful.
I think its important for KDE or someone at least to help Distros with integrating this properly. I'm still yet to come across a distro that has implemented Nepomuk in a half usable manner.
I have 4Gb of RAM on my desktop PC and I have allocated for nepomuk 1Gb (KDE SC 4.4 RC1) even that the index is only a 900Mb. On my Netbook what has *only* 1Gb I have gaved the 250Mb. The default is 50Mb what I feel is enought but I use so much the Nepomuk that I want it to be blazing fast.
I believe that the Nepomuk is a key technology for what KDE should invest more time. We need KDE apps to support it more and KDE SC need to be tweaked to use it as much as possible and really do the polishing for users. So using it feels natural and does not disturb the working.
I have one old PC with 512Mb RAM and even on that I keep Nepomuk enabled because it helps. And on that computer it does not slow the whole software system down at all, just improves the working.
If I would get Nepomuk to the Mac, I think I would love to see to replace the Mac OSX own Spotlight for a few weeks to see how well it compares it.
Ah i didn't realise that increasing the nepomuk storage capacity from the paltry 50MB to something like 1GB would help? And I am also assuming that this is on-disk space? and not memory usage?
Any way the storage capacity slider needs to be improved. to something like
minimum(50mb) ------ reasonable(512mb) ------- Ideal(1gb)
and the size there (50mb, 512mb, 1gb) could depend on the size of the directory that is being indexed?
The slider could be something along the lines of the thumbnail size slider (with predefined intervals (min, reasonable, ideal or even more)
At present there is no indication to user about the importance of the default 50mb setting
added a bug report for the above comment
https://bugs.kde.org/show_bug.cgi?id=222697
@Socceroos, have you tried Mandriva? They actually have paid developers to work on nepomuk.
http://doc4.mandriva.org/bin/view/labs/Mandriva_Smart_Desktop_2010-0
Perhaps one problem of nepomuk gaining acceptance is, that there are wrong priorities. There is fancy stuff in the works and talked about like new open and save dialogs, sharing of metadata with other people, the integration with plasma you talked about and probably more. This is all wow! However the basics are not even working yet.
The fulltext search, even though perhaps boring and only a small part of what nepomuk can do, would be something very useful to a lot of users and gently guiding them into the world of nepomuk. Provided it works unobtrusive, reliable and is well integrated. None of these is currently the case.
Also a very important question is how to deal reliably with generated metadata (which cannot be simply rescanned from disk) with regard to copying data around e.g. for backups. I won't invest energy in tagging files, and thus actually use this feature, if the data is lost sooner or later. For example I started cautiously tagging some of my files, sometime later all the tags were lost somehow (in this case I didn't even copy around anything).
This is my humble opinion. And of course everyone is free to work on what he/she likes, still may be a problem for the project as a whole.
I personally do not use tags. All my files are perfectly organized within folders, each with a custom icon describing the contents (e.g. video, music, pictures, etc). I never feel the need to "search" for files in my own desktop.
I also do not use local mail clients like KMail or local messaging applications (GMail works OK for me).
And frankly, neither of my friends does.
I'd be very interested to see some real world statistics about the usage of "desktop search" and "local mail clients" by the majority of users (not developers).
I that's the only usefulness of Nepomuk then I suppose I will disable it after installation, as will other users like me.
I there are more features planned/supported then it would be a good idea to start advertising them (i.e. features that improve production in an offline environment, with perfectly organized files in folders).
Post a Comment