KDE's Project Silk [LWN.net]

KDE's Project Silk

Posted Sep 22, 2009 10:47 UTC (Tue) by ballombe (subscriber, #9523) [Link]

The problem is not only that the server code is not FOSS, but also that such community site should be managed by its own community. It should belong to a non-profit whose members are the community members, with voting rights.

KDE's Project Silk

Posted Sep 22, 2009 11:52 UTC (Tue) by sebas (subscriber, #51660) [Link]

The idea is indeed to create interfaces for service-types (think of social-networking, web-shopping) as examples. Those APIs are independent from the service they interact with. Applications use these APIs to offer integration with webservices.

These APIs are backed by plugins. So a social-networking API would have plugins for facebook, opendesktop, linkedin, for example. The application developer doesn't care about which particular service is being used. The plugins that are backing the task-specific API are written using script languages, or are generated from webservice descriptions (if possible). The API also doesn't discriminate between free and non-free services. Using scripted plugins makes it easy to update these plugins whenever the web service interface changes.

Which services are offered depends on what plugins are provided. Of course it's easier to write such a plugin for a well-documented interface (think of a REST API like openDesktop) than for a closed service -- it'll become easier to use the service from silk-enabled applications if the service makes its content and interaction available in a documented, machine-readable way. Those plugins could also do screen-scraping as a last resort for less friendly web services.

The question how we, as a Free software project protect our Freedom is only loosely connected to this. The idea is to have three categories: open, free and ugly.

- "free": Free web service implementation (such as identi.ca, openstreetmap)
- "open": Documented interface, service is not necessarily free but has a stable, machine-readable API
- "ugly": closed and no openly documented interface

Obivously, we prefer and actively encourage "free" and "open" services, but there's nothing that stops people from implementing backends for "ugly" services.

KDE's Project Silk

Posted Sep 22, 2009 13:21 UTC (Tue) by anselm (subscriber, #2796) [Link]

I don't think you will ever be forced to use it.

Who knows? I'm already forced to run a completely extraneous MySQL server for no discernible benefit. It's as if the KDE-PIM developers have never heard of SQLite.

KDE's Project Silk

Posted Sep 22, 2009 13:44 UTC (Tue) by sebas (subscriber, #51660) [Link]

Akonadi (which is what you're probably referring to) *currently* requires
MySQL, its storage is abstracted so you could plug in a different storage
engine, however. The requirement is as simple as "right now mysql is the
only backend".

During the beginning of Akonadi development, the idea was to actually use
sqlite, but it was very problematic. Sqlite could not, at that point
handle the number of concurrent queries such a high-performance cache
would need, it also wasn't thread-safe.

It seems that sqlite has improved in those areas, and work on an sqlite
backend is well underway. There's also work going on on a postgresql
backend, by the way.

All that has nothing to do with forcing people to use certain webservices,
of course.

KDE's Project Silk

Posted Sep 22, 2009 20:56 UTC (Tue) by anselm (subscriber, #2796) [Link]

The problem is really when people like the KDE-PIM developers have their own notion of what is good for everybody and then proceed to cram whatever that is down everyone's throat. I'm a big fan of »Heuer's Law«: »Any feature that cannot be turned off is a bug.«

I don't *need* an SQL server to store a few hundred addresses. My machine is running an SQL server but I need that for other things (such as software development), and I certainly don't want my desktop environment to mess with it. Neither do I want the overhead of another MySQL-size daemon just for a bunch of PIM data (my machine isn't *that* big). Akonadi might be a great idea if it was optional, or if there was a backend that catered to people who need only very basic functionality, cheaply. I'm generally happy with KDE but this tendency towards a fixation on infrastructure for infrastructure's sake makes me look for alternatives.

KDE's Project Silk

Posted Sep 22, 2009 22:15 UTC (Tue) by sebas (subscriber, #51660) [Link]

Akonadi is not about MySQL, or a daemon of its size. Akonadi is about
providing a common synchronization and access layer for PIM data.

Even if you only store three emails in it, it still makes sense to make
them accessible for your applications other than the primary email client.
You also don't want to rewrite all the email protocols over and over, for
every new email application. That's what Akonadi takes out of your hand.

As I said, an sqlite backend is under way. That would be the one catering
to those very simple use cases you're talking about. It has only become an
option very recently, due to technical issues with it (see other comment).

KDE's Project Silk

Posted Sep 23, 2009 5:33 UTC (Wed) by dkite (guest, #4577) [Link]

I think you have it backwards. Akonadi has had very limited usage within
kde, gradually increasing as apps are ported to use it. In other words, the
ideas were implemented using tools that worked at the time. Proof of concept
type thing. As it becomes more useful, more integrated, the availability of
suitable backend libraries hopefully improves at the same time. But the
necessity of getting something out there for people to bang on forces bad
choices.

Really, all this episode shows is that for serious desktop work, developers
run into serious limits and sometimes walls that gives them the choice of
writing from scratch or using something inappropriate. In this instance due
to the lack of a lightweight full featured data/storage/retrieval engine.

Derek

KDE's Project Silk

Posted Sep 23, 2009 10:38 UTC (Wed) by segedunum (guest, #60948) [Link]

Ahhhhh. The whining refrain from someone complaining about 'bloat' on his machine. Quite frankly, most users want their desktops and applications to actually do things and they don't care about you. Nobody is interested in whether *you* don't think you need a SQL server. If you're that bothered then go and write something yourself. Some light reading:

http://www.joelonsoftware.com/articles/fog0000000020.html

KDE's Project Silk

Posted Sep 23, 2009 11:02 UTC (Wed) by nye (guest, #51576) [Link]

And people wonder why MS has near total desktop domination.

KDE's Project Silk

Posted Sep 23, 2009 11:21 UTC (Wed) by anselm (subscriber, #2796) [Link]

Well, maybe it's just me, but if I have to have bloat then I want that bloat to actually *work*. This is something that Akonadi so far has failed to deliver for me -- it insists on launching its MySQL server and trying to import stuff from our LDAP server, but then it basically quits and says »I'm broken, I can't do anything«. Of course once it has noticed that it is, in fact, useless it would be much too much to expect it to clean up after itself (say, by killing the MySQL server again). Incidentally, the help link on the dialog in question leads to a 404 error. I don't mind debugging stuff but if people actually seem to go actively out of their way to make it difficult then I have better things to do with my time.

I'm cool with bloat that actually delivers something worthwhile on top of what is already there. What I am less enthusiastic about is developers who appear to spend all their time pushing the envelope and don't seem to care about the day-to-day stuff. They're busy working on a car that will fly (and it actually does, in the wind channel in the lab) but the fact that the steering wheel falls off the moment the car is being driven out of the driveway into the road is below their attention threshold. (After all, flying is much more important, and the rest is just trivial stuff that the users can fix if they are »bothered«.)

I'm looking forward to the day when Kontact will actually be able to maintain a stable connection to an IMAP server that is running on the same machine, with no gratuitous hangs. This has never worked quite right as far as I am concerned -- it's OK most of the time but once or twice a day Kmail just hangs until either the IMAP server is restarted or a timeout happens. (In point of fact the IMAP server is only necessary in the first place because Kmail doesn't handle Maildirs the way it ought to, according to the author of the offlineimap utility, which doesn't appear to be a problem with other MUAs.) *Maybe* Akonadi will help there but I'm not holding my breath. I'm also looking forward to the day that the message editor in Kmail will once more handle Ctrl-K the way it is supposed to, which is apparently broken in KDE 4.x. There is more stuff in the KDE 4.x Kmail that used to work in 3.5 but no longer does -- but hey, the developers are working on an INTEGRATED FRAMEWORK so who cares about little annoyances like these? If the users are »that bothered« with them they can go and fix them themselves, because the developers are too important to worry about them. If a user dares voice an opinion that something the exalted developers have decided may not, in fact, be the greatest idea on Earth then they're »whining«. Right?

KDE's Project Silk

Posted Sep 23, 2009 14:15 UTC (Wed) by boudewijn (subscriber, #14185) [Link]

Come off your high horse and get real.

Do you really think the developers think that way? They aren't in it for
gratuitous coolness, as much as it might please your ego to believe that.
Don't be ridiculous. There are people working really hard to solve real,
hard problems -- problems that didn't exist in the days when KMail was
first designed, which only surface with today's email load. It's actually
one of the things that is developed commercially, instead of by misled
volunteer enthusiasts -- there are actual customers paying for akonadi,
kontact and kolab.

It's no doubt fun to rant like this, but you're simply wrong about the
developers motivations and priorities.

KDE's Project Silk

Posted Sep 24, 2009 8:16 UTC (Thu) by anselm (subscriber, #2796) [Link]

Whatever. But I suppose that even those »paying customers« would like their Ctrl-Ks to work correctly ...

KDE's Project Silk

Posted Sep 23, 2009 0:12 UTC (Wed) by alankila (subscriber, #47141) [Link]

It'd be mostly horrible idea to have, say, facebook.com and facebook2.com and then have these sites not integrate seamlessly. The value of many of those sites is in the size of the network. Have two sites and in the worst case both are equally popular and you have just neatly halved the size of the network and thus reduced its value by even more than that.

Anyway, google wave is on the horizon and might be just what everyone wants: social network where all servers interoperate by design, data is not held as hostage on anyone's server but copies of data are passed around to every participant who needs it, anyone can start a server, etc. It could well provide the platform that everyone could live with.

Oh yeah, it's supposed to have open-source frontends and backends as well.

KDE's Project Silk

Posted Sep 24, 2009 21:29 UTC (Thu) by nix (subscriber, #2304) [Link]

Anyway, google wave is on the horizon and might be just what everyone wants: social network where all servers interoperate by design, data is not held as hostage on anyone's server but copies of data are passed around to every participant who needs it, anyone can start a server, etc. It could well provide the platform that everyone could live with.

Sounds just like USENET to me.

KDE's Project Silk

Posted Sep 28, 2009 16:23 UTC (Mon) by nye (guest, #51576) [Link]

Then you're not listening :P.

You should take a look at Google Wave - the presentation at wave.google.com is worth watching if you have the time (it's about an hour and 20 minutes).

Of course you might still not like the sound of it, but personally it's the first 'next big thing' idea I can recall seeing which doesn't seem completely pointless.