Jerome Yuzyk posted on Fri, 16 Nov 2012 19:23:50 -0700 as excerpted:
> With all the hassles added by Akonadi and Nepomuk and Strigi for some
> higher "social/semantic desktop" purpose, does anyone actually _use_ the
> stuff? Or just the devs that thought it up? Googling "KDE Social
> Desktop" returns mostly 3-year-old links to articles about what it will
> do someday, but I'm wondering, "What does it do today?" There have been
> 9 major releases to date. Is everyone (but me) social-desktopping away
> happily and not writing anything (useful) about it? Even the
> Social-Desktop mailing list archive is pushing a year without any
> updates.
>
> Googling "KDE Semantic Desktop" doesn't provide much more current
> information about why I should care now, years after the project was
> started.
>
> All I know is that is causes a lot of problems, for me and others,
> without any obvious benefits. All I've seen for me is every once in a
> while, for no apparent reason, an app as simple as my email program
> suddenly consumes up to 2G of memory and I have to kill it and restart
> the Akonadi server. For email, something I've used in one form or
> another since the 1990s - _15 year ago_.
>
> So what's happening with this grand vision?
My honest opinion? In general, it's a tool for those that really aren't
comfortable with "traditional" computers and the way they work. The kind
that has so many icons on their (traditional icon-based) desktop they
don't all fit, because that's the only place they can find things.
For this type of person, the resource usage is arguably worth it, because
it lets them find things in more "human" ways, "The picture I remember
that I'm looking for was taken on my vacation in Italy in 2003. It was
of a street performer, and I remember recording his music too, so there
should be both the picture and the music. They go together. Where did I
put it? It's on the computer. What subdir? It's on the computer."
Vs. a more "computer literate" person, who would handle it this way:
"The mp3 and jpeg should be in ~/vacations/2003/italy/rome. I saved each
day's stuff in a separate subdir and don't remember the day and thus the
specific subdir, but while I have lots of jpegs from the trip, I only
recorded a couple mp3s so I'll do a file search for *.mp3, and even if I
don't recognize it immediately from the name, I can get the subdir and
timestamp from those, and the jpeg I'm looking for should be with one of
them and have a timestamp within about half an hour."
Those are two very different ways of organizing and storing stuff, one
that "just works" with the way computers normally work, one that doesn't,
at least not very well, without the help of strong tools like "semantic
desktop".
To the one kind of person (the one who identifies the files as mp3s and
jpegs, not just pictures and music), apparently including you, and
*DEFINITELY* including me, all the extra resources required to do that
semantic-desktop tracking are a waste, because even if we don't know
/specifically/ where something is, we naturally work and think enough
like a normal personal computer does that it's just normal for us to be
able to pin down the location well enough that an ls (very possibly using
tab-completion at the command line =:^) or grep can find the specific
file we need in "close enough" to real time, that the help of the
"semantic desktop" isn't worth the resources in time, memory and index
storage space to first build the index and then keep it updated, let
alone having to deal with the host of additional bugs that by definition
come with all that extra complexity.
The other type of person has never felt really comfortable working with
computers in any case, and it has always been "like pulling teeth" to
translate their desires and the way they normally think, into something
the (traditional) computer can actually understand. As a result, they're
constantly fighting with it, never really understanding it. Normally,
all they do is finding some "magic" ritual that seems to work, they
couldn't tell you how or why, and they repeat that every time they want
the same type of results.
This type of person has /always/ taken quite a while to get anything
(outside their specific magic rituals anyway) done on the computer, if
they can do it at all, because it's constant trial and error, hitting
this button, typing that bit in, because it "might work", or because
they've seen someone else do it that way, not because they have any real
understanding.
So for this type of person, the resources that the "semantic desktop"
requires aren't wasted at all. Yes, it might make the computer run a bit
slower, and it might crash at times, but the success rate is at worst, no
worse than the way they did it before, simply guessing, fighting the
computer all the way, because they really did NOT have any idea what was
going on or how it worked. Before, it was trial and error and fighting
the computer all the way, to maybe finally get what they were trying to
get. Now, it might take as long, and perhaps it crashes instead of
working some of the time, but the results are no worse than before, and
at least now, the process is a bit more automated, leaving them free to
do anything /but/ computers with the time freed up.
That's painting in huge overbroad strokes, but IMO, it's a reasonably
accurate "3000 meter view" (aka 10,000 foot view) of the situation. Even
if it doesn't well explain the /other/ side (something I can't be sure of
since I'm /not/ the /other/ side), it certainly seems to explain why most
"computer literate" folk seem to have little use for this "semantic
desktop" stuff, particularly given the resources it uses and the
additional bugs and crashes it triggers, since for us, it's not /that/
hard to get at the data the traditional way, because we just naturally
think and organize that way, second nature.
FWIW, that's from my viewpoint as a gentooer, which I guess could be
classed as a geek's geek, certainly so if one accepts that people that
actually enjoy working with Linux tend to be geeks in the first place,
then someone who chooses gentoo for the additional control over one's
computer it allows, must be a geek even to his Linux geek peers.
I chose gentoo in part /because/ it exposes build-time options such as
the semantic-desktop to the individual sysadmin (in an automated way, of
course anyone could do it if they want to build it themselves, manually,
doing the standard configure/make/make-install, gentoo automates that
while exposing the choices available at that configure step =:^).
And, at some point, after the newly akonadified kmail2 (then kdepim
4.6.1) ate yet another email, and I had to figure out if/how I could get
it back, I decided it simply was NOT worth it. Top priority was then
finding an alternative to kmail, that the devs hopefully wouldn't break
the working version of, just because they could, YEARS before the new
version was mature enough to properly replace it, as the kdepim folks did
with kdepim, repeating the same pattern of kde4 in general, before it.
I ended up on claws-mail, here. It's unlikely they'll up and break it as
the kdepim folks did kmail, because a major touted feature of claws-mail
is its scriptable extensibility, and doing what the kdepim folks did to
kmail would break all the scripts and customized hotkeys, etc, that are
one of the major reasons people use claws-mail in the first place.
I also had to replace akregator, the only other kdepim-based kde app I
really used, since while it didn't directly depend on akonadi yet, it did
depend on kdepim-common-libs, which depended on akonadi. As it happens,
I replaced it with claws-mail, with its feed plugin, as well, altho
unlike most it would seem, I prefer keeping mail and feeds separate, so I
setup two separate claws-mail instances, one for feeds, one for mail.
With both kmail and akregator gone, I could purge all of kdepim and
akonadi from my system. In turn, with that gone, I could set (gentoo/kde)
USE=-semantic-desktop, rebuild the kde I still had installed, and
exterminate semantic-desktop from my system as well. After a bit more
cleanup of the bits and pieces, I had removed virtuoso, mysql, rasqual,
redland, and soprano, too.
And even tho I already had as much of the semantic-desktop turned off at
runtime as was possible to turn off, no nepomuk running, no strigi, etc,
getting rid of all that junk SERIOUSLY sped up what remained of kde! So
much so that while I had argued kde 4.5 was really the first kde4 that
matched kde3 in usability and quality, finally ridding kde4 of all all
the semantic-desktop dependencies (save for strigi, still pulled in for
the headers but it can't actually run as there's no backend for it to
use) and running kde4 without them, was the first time I could honestly
say that the kde4 experience had not just matched but actually EXCEEDED
kde3 for me. Ironic, that, given that the whole semantic-desktop thing
was a huge bullet-point feature of kde4 in the first place, but building
without it really WAS what made it finally worth the switch from kde3,
for me! =:^)
So yes indeed, I perhaps more than most, tend to hate semantic-desktop
and the bugs and dead-weight bloat it has saddled kde4 with, and I'm
definitely glad to be able to be rid of it here!
But even with that, I recognize that for people of a type far different
than me, people who really don't understand computers and who tend to
spend more time fighting them or at best, simply following rote "magic"
that they've found works, than actually working with and understanding
how and why the computer works as it does, for THESE people, semantic-
desktop may be a very useful tool indeed!
So there's a time and a place for those semantic-desktop features. But
that time and place is about as far removed from me and my systems and
it's possible to get!
--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman
___________________________________________________
This message is from the kde mailing list.
Account management: https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.
