28
Jan
In the beginning it was a neat idea. They wanted a common set of standards for interop on the desktop... Then somehow it all went wrong, somehow implementations started to pollute these standards. This has caused quite a quandary, now somehow we have these standards that are bound to technology, some of which we could argue is pretty poor technology. IMHO this is reason to abandon efforts to follow these standards until they become standards once again, instead of implementations.
In the last 3 months KDE has integrated more with GTK and Gnome without these standards, this furthers my annoyance with freedesktop.org. While things like dbus and their xserver are interesting, they are implementations and imho inappropriate for a standard. It would have been far more effective if there was a protocol established that projects could implement natively. If we choose to use common implementation cool, otherwise don't trap people into them.
I don't see a reason to follow standards that cause use to be forced to use a platform. It seriously sounds like a Microsoft approach to things. "You can work with our platform, if you are running on it"...
oh well... the nice thing about standards is there are so many to choose from.
- geiseri's blog
- Login or register to post comments
- 1868 reads
Syndicate Blogs
Comments
I'm Afraid I Agree....
Then somehow it all went wrong, somehow implementations started to pollute these standards. This has caused quite a quandary, now somehow we have these standards that are bound to technology, some of which we could argue is pretty poor technology.
Well, I agree here. I know some people say that you cannot prove specs and standards without implementations, and that is certainly true, the fact is that the specs and standards do not seem to have been developed along with the implementations. The implementations seem to have completely taken over. Could D-BUS be re-written to C++/Qt through a set of verifiable standards? I don't think so.
It would have been far more effective if there was a protocol established that projects could implement natively. If we choose to use common implementation cool, otherwise don't trap people into them.
Yep.
I don't see a reason to follow standards that cause use to be forced to use a platform. It seriously sounds like a Microsoft approach to things. "You can work with our platform, if you are running on it"…
Well, quite. I wonder why all the Gnome people seem to support it....
The emphasis seems to have shifted from standards to "Ooh, look at our swish new X server" type thinking. Freedesktop seems to be a project to shift all the major components of a desktop Linux distribution into Havoc and Red Hat's lap. Nothing has hardened that belief more than Havoc's announcement about XFree and X.org "basically merging". Well even if they are what the hell has that got to do with him? We also got some rubbish after that about Freedesktop only being a container project and not forcing anything on anyone. Yer, right...
fd.o focus
Err, dude, it's not about shifting all the major components of the desktop into Havoc and Red Hat.
This seems to be a major misconception, so let be be pretty clear.
JUST BECAUSE SOMETHING ISN'T KDE, THAT DOESN'T MAKE IT A GNOME CONSPIRACY. NO, REALLY. I'M SERIOUS.
Keith's new X server is nice, sure, but that was basically unannounced. Slashdot picked it up and said "dudes, check out this shiny transparency!". That's it. Most of the fd.o focus has been on xlibs, the platform, and other unsexy things that Gentoo users don't really fawn over too much.
Havoc didn't make those comments. He said that most of the former XFree86 core team was now a part of X.Org (actually old news if you followed public, but obscure, pages); he was the only one of the fd.o informal leadership group who was at LWE NY, so it was him that made the announcement. X.Org lives on fd.o now (I have my own opinions about it, but that's irrelevant).
I never said fd.o was only a container project. We're a multi-faceted project. That means we do a few different things.
We're not forcing the platform on you (plural). I just think you'd (plural) be stupid not to adopt it, from a personal point of view, but there you go. :)
Re: fd.o focus
Sorry if you're seemingly getting a pasting here Daniel, but it is not personal at all, I assure you. A lot of people have some genuine concerns here, as I'm sure you can appreciate.
JUST BECAUSE SOMETHING ISN'T KDE, THAT DOESN'T MAKE IT A GNOME CONSPIRACY. NO, REALLY. I'M SERIOUS.
I don't see a Gnome conspiracy. However, I see a lot of potential power-plays going on at Gnome now between Ximian, Red Hat and Sun. Sun are a lesser issue because they don't seem to know what to do, but hey, that's just Sun. Red Hat have Bluecurve and desperately need to find some way of making the darn thing work. Ximian have Mono and want that to be used at the core of Gnome. Yes, I saw Miguel de Icaza's posting about Mono not taking over Gnome and I saw it as crap. I was right. I don't want to see KDE getting involved in any of this, even indirectly in technologies that are used.
Keith's new X server is nice, sure, but that was basically unannounced. Slashdot picked it up and said "dudes, check out this shiny transparency!". That's it. Most of the fd.o focus has been on xlibs, the platform, and other unsexy things that Gentoo users don't really fawn over too much.
It was an example more than anything else. I suppose when you post screenshots of lots of transparent windows and incredible eye-candy people are going to react :).
Err, dude, it's not about shifting all the major components of the desktop into Havoc and Red Hat.
I really, really, really, really want to believe it. But, Bluecurve still exists, and it is still as big a mess as ever. Guess who benefits directly if Freedesktop.org technologies get widely adopted? It doesn't matter how much anyone believes in Freedesktop, no one can be naive to that.
Havoc didn't make those comments. He said that most of the former XFree86 core team was now a part of X.Org...
Granted, but that was not what was announced (original newsforge announcement, and on OSNews):
Havoc Pennington of freedesktop.org announced that X.org and XFree86 have essentially merged...
X.org didn't say that, and XFree certainly didn't and it says here that Havoc Pennington announced it. XFree vehemently denied it right afterwards, so where did it come from then? I'm willing to believe it was just a cock-up because I cannot verify directly how it was announced, but that was the way it was announced.
I never said fd.o was only a container project. We're a multi-faceted project. That means we do a few different things.
Yep, OK I apologise for that. I heard the phrase 'container project' somewhere, but can't verify it or remember who wrote it, so it is dismissed until I can.
We're not forcing the platform on you (plural).
Well, there are many definitions of forcing, such as Microsoft not forcing Windows on you, so I think we're just going to have to move slow on this, if at all.
I just think you'd (plural) be stupid not to adopt it, from a personal point of view, but there you go.
Well, the key word here is adopt, and it is a rather big clue. I certainly think KDE should use Freedesktop technologies where they are just plain good enough, and preferably standards and specs that can be implemented natively in KDE (of which there do not seem to be many of any real use). I don't see anything wrong with that. However, widespread adoption of D-BUS etc? No. A large cross-pollination of KDE, Gnome and other projects only benefits one project and organisation directly - Bluecurve and Red Hat. However impartial you are, you can't get away from that.
More clarifications
I don’t see a Gnome conspiracy. However, I see a lot of potential power-plays going on at Gnome now between Ximian, Red Hat and Sun. Sun are a lesser issue because they don’t seem to know what to do, but hey, that’s just Sun. Red Hat have Bluecurve and desperately need to find some way of making the darn thing work. Ximian have Mono and want that to be used at the core of Gnome. Yes, I saw Miguel de Icaza’s posting about Mono not taking over Gnome and I saw it as crap. I was right. I don’t want to see KDE getting involved in any of this, even indirectly in technologies that are used.
... okay. But how is that relevant to fd.o? We shouldn't use D-BUS, because Red Hat made it and they want to take over the world?
I really, really, really, really want to believe it. But, Bluecurve still exists, and it is still as big a mess as ever. Guess who benefits directly if Freedesktop.org technologies get widely adopted? It doesn’t matter how much anyone believes in Freedesktop, no one can be naive to that.
... okay. But how is that relevant to fd.o? We shouldn't use fd.o, because Red Hat made Bluecurve and they want to take over the world?
Granted, but that was not what was announced (original newsforge announcement, and on OSNews):
Havoc Pennington of freedesktop.org announced that X.org and XFree86 have essentially merged…
X.org didn’t say that, and XFree certainly didn’t and it says here that Havoc Pennington announced it. XFree vehemently denied it right afterwards, so where did it come from then? I’m willing to believe it was just a cock-up because I cannot verify directly how it was announced, but that was the way it was announced.
Don't believe everything you read, dude. Havoc isn't that stupid, let me assure you.
Yep, OK I apologise for that. I heard the phrase ‘container project’ somewhere, but can’t verify it or remember who wrote it, so it is dismissed until I can.
Let me assure, as one of the fd.o admins, that it's not just another SourceForge.
Well, there are many definitions of forcing, such as Microsoft not forcing Windows on you, so I think we’re just going to have to move slow on this, if at all.
We have to move slowly, because there are many definitions of the word 'forcing'?
Well, the key word here is adopt, and it is a rather big clue. I certainly think KDE should use Freedesktop technologies where they are just plain good enough, and preferably standards and specs that can be implemented natively in KDE (of which there do not seem to be many of any real use). I don’t see anything wrong with that. However, widespread adoption of D-BUS etc? No. A large cross-pollination of KDE, Gnome and other projects only benefits one project and organisation directly - Bluecurve and Red Hat. However impartial you are, you can’t get away from that.
No, it doesn't. It benefits the desktop. It benefits users who don't f**king care what desktop they're running - they just want a shiny, purty desktop that works. Think about it: if there was no need for Bluecurve and other integrative technologies, why would everyone be focussing on it so much now?
I tried to explain the concept of multiple toolkits to my mother a while ago, as to why my desktop looked so all over the place and crap (I was using GNOME on my laptop - more lightweight as my laptop sucked). That's when I realised it was ludicrous.
Standard and implementation
I think the original idea with standard and implementations was that producing only specs is worthless, while producing a spec with a working implementation is a lot more intelligent. At least, you can expose your specification to real life and see the real problem arise. That's how RFC work. In this spirit, I find it a good thing that fd.org has both standards and implementations.
Now, those implementations are a lot on the Gnome side. That's because they have somebody there who is paid to work on that stuff. The glib vs QTL is a problem with no solution. When you code, you have to use lists, maps and string at some point. Then you have to make your choice for one of the available template library. The pro blem with the Qt one is that it is not available without the rest of Qt. Else, we could develop a C binding for QTL.
Re: Standard and implementation
The glib vs QTL is a problem with no solution. When you code, you have to use lists, maps and string at some point. Then you have to make your choice for one of the available template library. The pro blem with the Qt one is that it is not available without the rest of Qt. Else, we could develop a C binding for QTL.
As Windows developer who has increasingly come to a Linux world over the past two years, I'll solve this problem for you. No one wants to know C, and no one will want to know glib. Clear enough? Unless you really need to write some real low-level stuff, which the vast majority of developers will not, this is a none issue. The world is C++, despite the hype of Java and .NET (which just hasn't happened), and this is a big promotional point for KDE.
Templates/Containers
Then you have to make your choice for one of the available template library. The pro blem with the Qt one is that it is not available without the rest of Qt.
Well there is TinyQ http://www.gnu.org/directory/libs/cpp/TinyQ.html and if someone does not like GPL, I heard that the C++ standard library contains such things as well
As someone who runs KDE on Mac OS X...
...I find myself in some ways worried about the freedesktop.org thing as well. I think they're doing very interesting things, but you've summed it up in it not being suitable as a standard (at least, in my mind, not yet). It seems like a big step backward to start preparing to standardize on stuff that pretty much only supports Linux in any useful way (dbus, keith packard's new X server, etc...)
I'm sure someday perhaps some of these start-from-scratch projects may eventually replace the things they were designed to replace, but isn't it a bit early to be jumping to another track? It seems like a lot of this is still at the "proof of concept" stage.
Granted, I've only had a cursory look at some of these new technologies, so perhaps they're more cross-platform than I thought, but one of KDE's (and even GNOME's) strengths is their portability.
Could you please help out?
I know you're way busy, but could you please join the platform list? Having your (or Ben's) input would be way handy, as you're in quite the niche here. I'm going to start posting a few things and getting plans together ASAP (one of my friends has been in ICU for a while on >60% oxygen, I've had my 'net access cut off at home, and other long stories, so I've been too busy to do it), but yeah. I'd love to have your input, seriously.
Please help the platform out; I think you could be really useful. :)
don't confuse hosting and endorsing
fd.o has evolved into a hosting platform for many interesting (and not so interesting) pieces of software, but I don't think that defeats its purpose as a place for discussing interoperability standards.
KDE should make use of the stuff where it makes sense and ignore what's not appropriate.
I don't think it's time to abandon fd.o, I think it just starts to make sense.