レナート   Wunschkonzert, Ponyhof und Abenteuerspielplatz   ﻟﻴﻨﺎﺭﺕ

Posted by Benoit Jacob at Thu Jun 12 19:35:53 2008
Since Gmane doesn't seem to allow navigating through the discussion thread, let me also add here my summary:

- KDE already has knotify, but it does not operate at toolkit level. Thus, two possibilityes:
-- Either KNotify is enough: then this is the end of the story.
-- Or it is really useful to have a solution at toolkit level. In which case you should talk directly to Trolltech. There are many trolls on the kde-devel list, but not the right trolls for this topic. Even Thiago tells you that he doesn't feel competent on this topic.

- I think that one explanation for the lukewarm reception of your proposal is
that you came up with an already implemented proposal; it would have been
better to let people join much earlier, during the design stage. I just
talked with an Akonadi developer who told me they had been discussing with
Evolution people since the early design stage before any code was written.

Posted by Albert at Thu Jun 12 20:44:50 2008
Well, let's see your first blog, you announced something like 3 STANDARDS AND FREEDESKTOP SPECIFICATIONS that noone uses right now and noone outside gnome had probably ever heard about.

And you expected a nice reception?

That's not the way one gets nice words from the people in the other side of the fence, sorry.

Posted by Lennart at Thu Jun 12 20:52:33 2008
Albert: Uh? You cannot count, apparently. (Welcome to the club!)

Also, as I mentioned, the relevant KDE people had seen the two specs before, not in their current state, but they saw them, months before.

Also, I didn't expect anything from them. And I am not sure how you got the idea I was complaining about their reception.

Posted by Lennart at Thu Jun 12 20:55:21 2008
Benoit: of course Gmane allows you that. It's very good at that actually.

And, personally these days I believe that discussing things with nothing concrete in your hands is pointless. To know what you want to discuss you need to have implemented it. It's that simple.

Posted by Jonas at Thu Jun 12 21:09:38 2008
Well, a disclaimer first: I'm not a developer. Not in KDE and not in  Gnome. I'm just a user of the first that wants better cross-DE compatibility across the board.

That being said, apart from the reasons outlined by Benoit and Albert I think it may also be a case of bad timing.

This kind of integration, at least to my non-developer ears, sound like it would have to become a part of the core modules (regardless of which DE), right? And KDE is in a transition phase from 3.x to 4.x right now. A transition that has, for the time being at least, brought along quite a few regressions that the KDE devs are quite busy fixing. It doesn't sound like a very good idea to introduce yet another new concept/lib that may (or may not) cause regressions until the KDE 4.x series has stabilized.

Posted by Benoit Jacob at Thu Jun 12 21:35:33 2008
Lennart: now I can see that clicking on the thread title shows a nice treeview. I was expecting "Next in thread" links like on other archives.

Part of the discussion has to be done before implementing anything, part has to be done while implementing, part can be deferred until a first implementation is done.

Another issue is that whatsoever, this whole discussion plus the needed discussion with Trolltech should have happened before even submitting anything as freedesktop standards. This is just my uninformed point of view, but it is looking like freedesktop is much closer to GNOME than it is to KDE. That might be just fine, but many people including me were somehow believing that freedesktop was neutral between GNOME and KDE.

Posted by joe at Thu Jun 12 22:56:25 2008
"This is just my uninformed point of view, but it is looking like freedesktop is much closer to GNOME than it is to KDE."

How about you keep you uninformed views to yourself till you get informed on facts? Deal?

Posted by Benoit Jacob at Thu Jun 12 23:04:41 2008
joe: sure, as soon as you explain me why that is not the case? I pointed out what looked to me like evidence that Freedesktop.org is much closer to GNOME than it is to KDE; could you elaborate on why you disagree?

Posted by Oliver at Thu Jun 12 23:09:28 2008
So libcanberra looks like a really good idea, same with sound theming. Nice work!

But what I don't quite understand is why you did that posting to kde-devel. I'm not a KDE or Gnome developer (but use both DEs); but after reading the initial few kde-devel posts and your summary, I got the impression that you didn't quite intend to get KDE in the boat at all...

Your posts were from the beginning quite full of derision and some sense of "here's the ultimate Linux sound event solution, and if you KDE people were good developers you'd of course use it, but secretly I know you won't use it, proving how bad devs you are"... This is at least how it looked to me.

I'm really interested: if you think honestly about it, what kind of reaction did you expect from the KDE devs, apart from "who are you to tell us how to make sound events, without even having looked at our current solution"? If you were really keen on making a sound event lib for Gnome and KDE, would you really have approached the KDE devs like this? Or, in other words: if on the other hand you wanted to "sell" some Gnome devs on a sound event lib, would you have written anything similar?

Posted by Oliver at Thu Jun 12 23:10:49 2008
[Btw. maybe you could extend the "Comment rejected" message in your blog with " - please enable cookies"]

Posted by Lennart at Thu Jun 12 23:42:29 2008
Benoit: You apparently have no idea what fdo actually is. It is no standards body or anything. It's just a server where you can host your stuff if you ask nicely. Just because something is hosted at freedesktop it doesn't need to be any good, in fact among the stuff fdo hosts is a lot of total crap. That said libcanberra is not even hosted on fdo. Might be even worse crap... ;-)

So, stop phantasizing about "submitting as a freedesktop standard". Makes no sense.

And I can repeat this once again, the specs are not that new, and I know that a few relevant KDE people were aware of them months ago.

So, please stop criticizing how I handled these things. The whole point of this blog story was to make clear that while the KDE people might not have been actively involved with writing those specs they were very well kept in the loop.

So much about the specs.

And about libcanberra: yes, it is completely new stuff, and I wrote an email to get the KDE people into the loop, to check if they are interested or not. Apparently their interest is rather minimal, but that's fine. I never complained about that. But for fuck's sake, stop whining we wouldn't try to work with them and increase cooperation.

Please, stop commenting any further on this. Or do your homework and read my blog entry for once and get your facts straight. Thanks.

Posted by Lennart at Thu Jun 12 23:45:02 2008
Benoit: and those rumors about the closeness of fdo with GNOME vs. KDE. What nonsense is that? Everyone can host a project or spec there. Regardles if he writes himself with a g or with k.

Posted by Cyrille Berger at Thu Jun 12 23:45:41 2008
Comming with a solution to a problem that was allready solved in KDE (and Gnome too ?) and not bringing any advantage and expecting people to jump and be over enthousiastic about it is extremely weird, to say the least.

The feeling that freedesktop.org is Gnome-oriented comes mostly from the fact that once someone from Gnome has implemented something, it comes to fd.o and says "hey guys this is a standard for doing this", it would be the same if KDE developers goes to fd.o with plasma and says "hey guys this is the standard for making applets on your desktop". This kind of collaboration can't work because it doesn't takes into account the problems of the other side of the freedesktop.

If you really want the collaboration to be successfull, you will try to involve everybody from the start.

First you define your problem, then ask other people from other desktop environnement about their problem, then try to find how you can solve the problem in common, then define a spec. As for the implementation, who cares if they are two of them, or a single library ? (if you want your spec to be validated you will need two anyway...)

Posted by Lennart at Thu Jun 12 23:54:48 2008
Oliver: I am not sure what you are expecting me to expect from them?

So, a few guys wrote a spec or two, another guy implemented it. Then he went and asked some even other guys if they might be interested to join in. They said, yes, but only partially. And that's the end of the story.

The response from them is absolutely what I expected. Some stuff we'd be able to agree on, some stuff not. Perfect! Paradise! Rock'n'roll! I mean, if we'd always agree on everything we wouldn't have both KDE and GNOME around these days anymore, we'd only have a single project. And if we'd never agree on anything, then we would have no fdo around. So the response I got is perfectly in line whith what was to be expected.

The intention of this blog story was to explain that I actually do communicate with KDE people. Apparently I totally failed with that intention. Now I get even more flak from those dudes with no clue about what they talking.

Posted by Lennart at Fri Jun 13 00:00:59 2008
What is this, an invasion of people who comment before they read?

Cyrille: if you'd had read a single line of what I wrote in that kde-devel thread you'd know what nonsense it is that you claim "all those problems have been solved before".

AND FOR FUCK'S SAKE, I HAVE NEVER EXPECTED ANYTHING BUT A BIT OF AGREEMENT AND A BIT OF DISAGREEMENT FROM THEM.

Posted by Stoffe at Fri Jun 13 00:25:16 2008
Meh. Ignore the trolls, Lennart. You do lots of awesome stuff all the time and you've handled this well. Those who can, do, those who can't, troll. It will always be alike this and these people will never read up on anything, not that they'd understand it anyways.

Posted by E at Fri Jun 13 03:04:31 2008
Why not come up with a better name?  The naming of so many gnome related technologies is so ... odd - it wouldn't surprise me if the name alone caused less buy-in - especially from the kde folks.  Call it libsoundevents or something normal.  Canberra is a city ... and a library for handling sound events?  Wha?  At least is isn't called libgsoundevents.

Posted by Benoit Jacob at Fri Jun 13 07:56:30 2008
quote: "So, stop phantasizing about "submitting as a freedesktop standard". Makes no sense."

Look at your own previous blog entry. You announce two XDG specs in #1 and #2, and then in #3 you call them "standards":
quote: "Of course, what would the mentioned two standards be worth if [...]"

So it seems that nobody know whether FDO is a cross-desktop standards org, or just a repository for any desktop-related stuff. You seem to say it is the second, but still called "standards" stuff you put on it.

I admit that I didn't realize at first that only announcements 1 and 2 were as "FDO standards" and in particular that libcanberra is just one implementation of them, not itself a "FDO standard". I was confused by the fact that you announced all that together.

Still I believe that a "FDO standard" is something that must be discussed publicly with KDE (not just a few individual devs) before putting on FDO.

I mentioned FDO being close to GNOME because there are many other examples of specs being prepared by GNOME folks, then put on FDO, and only then discussed with KDE folks! I don't see any example of the converse having happened. Not that the specs are bad but it would still feel good to be able to interact earlier instead of being presented with a finished product.

Cyrille and I definitely aren't trolls, we are developers who happen to contribute in KDE and who want things to be clear. We're not here to make you lose your time. If I say something wrong it's because I don't understand something which might suggest that it wasn't clear in the first place. I don't like too much being treated like an anonymous troll on OSNews. See how Cyrille and I use our full names.

Posted by shamaz at Fri Jun 13 10:15:34 2008
ahem. This discussion is turning into a KDE vs Gnome one...
The point that really strikes me is  that everybody is talking about libcanbera.
Well, the first point of FDO projects is to have a common SPECIFICATION isn't it ?
If it's not optimal or too hard (I don't know) to use libcanbera in kde, is it really a problem ? If they propose another way to attain compatibility with the specs it seems ok for me.
just my 2 cents...