eWEEK is currently featuring an article on the UserLinux and KDE debacle. Indeed, exciting times are ahead as we forge on with our plans for both KDE::Enterprise in general and the KDE/Debian project. You've heard of cool hacks to make KIO slaves accessible on a system-wide basis, you've heard of GTK/KDE integration (don't be surprised if KDE Gimp, a.k.a. Kimp, makes an effortless return), you've heard of the new Debian tools in development, you've heard of a KDE/Debian live CD. All of this and more is coming to fruitation. See you next year!
Dot Categories:
Comments
That's one *hell* of a teaser! 2004 is going to be one kick-ass year for KDE!
Hi,
so what's the deal about UserLinux at all. What can it be that is not already out there. Being a Debian fan, may I say that if not Debian GNU/Linux then SUSE will offer exactly that?
Sure, it's all buzzword, lots of promises. More or less what they actually can release has to be "Debian testing" with non-free addons like closed source drivers, but which exactly who will validate?
Who will validate? How will UserLinux exactly be financed? What exactly will it allow a corporate "user" that Debian not already does. Debian has a open structure, will even give all choice and ... everything but quick series of releases.
If it's just for the desktops to use and a development platform to target, SUSE will do for now. One day, Debian will also be good enough for most companies. So?
Yours, Kay
From lurking on the list, this is what I gather ...
The central non-profit (UL) will brand the distribution and certify support providers. It will not have any technical staff and will be financed through fees that the support providers pay to become certified. I don't think it will be a large scale operation.
The technical decisions will be made based on a "meritocracy", with Bruce presiding as a benevolent dictator (to step in when consensus cannot be reached, or any other reason he deems important enough--like the GPL vs LGPL decision). The existing Debian infrastructure will be used to maintain the distribution source code, and UL will contribute back to Debian.
For example, Bruce is primarily working on the installer right now, which (AFAIK) is one of the main things holding up the next release. (Well that, and 500+ release critical bugs ...) From what I understand, he wrote the first installer so this seems like a good fit.
The central non-profit will provide a central 800 # where support calls go--these leads get farmed out to the support companies that have been certified.
I guess this is the main difference w.r.t. a corporate user--a central point of contact for support that remains constant and controls what is in the distribution. It's different from existing models in that the support is provided in a distributed model--for example, he has made it clear a couple of times that the certification process should not rule out sole-proprieterships.
That's my understanding ...
"For example, Bruce is primarily working on the installer right now"
Not another bloody installer.
The new Debian installer.
http://www.debian.org/devel/debian-installer/
I have really no faith that UL will survive. I mean, really. When it comes down to it, it's just another distro based on debian. Does it mean anyone will use it?
Take this as a factor. Novell, now owning Suse/Ximian, will be pushing linux hard. The places UL is trying to cater to, small IT shops, alot of those IT shops are running Novell services. So now that Suse is backed by Novell alot of those IT shops will just go with the same company that they have been trusting for years. Why would they use a fresh distro that has not "proven" itself yet and is in all technicality, crippled.
UL is just a big mess of political nonsense that no one wants to deal with. If they want to deprive their users of quality free software, then it's their own necks they are hanging themselves by.
quality-free software? LOL!
Read again. And don't forge.
Don't worry, I was just kidding :-)
See the URL below for some news about upcoming Novell services:
http://www.linuxplanet.com/linuxplanet/reports/5175/1/
Wow! The article says the GUI will support either KDE or GNOME! Good to see Novell keeping their KDE support. Maybe the Ximian division will start doing KDE development?
Does the article say it'll support *both*?
I didn't understand it that way.
Yes, it does say that in my opinion.
No, it doesn't.
It just says they will choose one of the two.
To avoid a big public tohoowabohoo now, they will announce later, *after* the fact, that they only support a GUI of Gtk/Gnome.....
;-(
.
.
I see 'a GUI that will support "either KDE or GNOME," ' in the article.
IIRC "either ... or" means a mutually exclusive or relation.
No. It's just a way of talking that might not be familiar to non-native English speakers. I'm pretty sure he means BOTH.
Could you ask *him*? Maybe write an e-Mail? And then report back here with a dot story?
Actually as a "native english speaker," the statement is ambiguous. It could mean both, or only one or the other. When I read it it seemed to me to be one of them would be supported but not both, and they'd tell us after the fact. I think you are being overly optimistic.
The following quote from the depth of my archives came to my mind:
"The warning message we sent the Russians was
a calculated ambiguity that would be clearly understood."
-- Alexander Haig
(Though I don't know the original context. Cold war, AFAIK)
Waldo? Your take on that? ;-)
Not necessarily. You can say "the M-16 can fire either a 40mm grenade or a 5.56mm round." That means it supports both. So the statement is really ambiguous, it can mean it will support both GNOME and KDE, or GNOME or KDE.
"You can get it for people as a stocking stuffer,"
I think that tells you how seriously you can take this.
i like it when i do a poo and let it out slowly-it feels guuuuuuuuuud ;D
riggghhhhhhhhhht.................wierdo lol
I read a linked article from the posted eweek article. And Bruce says something interesting.
But, in the end, Perens said he "made a decision for this particular project. I don't pretend that it was a democratic decision, and we're moving on to other levels. I can't be everything to everyone."
So basicly he was lying when he even asked for debate. He had his mind made up. Nothing was/is going to change it. It's not UserLinux, its BruceLinux. I thought this was supposed to be a *COMMUNITY* project. I though *USER* inplied user involvement.
Also something interesting from the article. UL has no known developers and the KDE/Debian initiative has 35 developers. Interesting, to say the least.
So despite Bruce's efforts, he isn't going to kill KDE.
It's the developers, stupid.
To paraphrase an election slogan. KDE and Gnome both have very active development communities. Neither will go away. In fact, despite what some think, it is in no-ones interest that either go away. Monocultures either proprietary or free, don't work well.
There are some that wish KDE would go away, but they only speak for their companies, not any community. This fuss is not about free software, it is commercial competition pure and simple.
Derek
> There are some that wish KDE would go away, but they only speak for their companies, not any community. This fuss is not about free software, it is commercial competition pure and simple.
Exactly. These people only wants to use what is free without giving back to the free community, or even outrun their free competition in the future.
Is this really just and fair?
> I thought this was supposed to be a *COMMUNITY* project.
A true community project doesn't need one million dollar funding to kick off.
wonder why they chose userlinux.com and not userlinux.org...
Nor does it need backers who according to Bruce do not want the publicity at this time. So the obvious question is, who are these backers? SCO?
See
http://cgi.userlinux.com/cgi-bin/wiki.pl?Proposals/GraphicalUnicity - it will be interesting to see how they want to make a Qt branding theme without including Qt (forcing ISVs to ship statically Qt-linked applications which don't support style plugins).
I think this initiative is exactly the right thing for KDE to
be focusing on right now. In the past, distros just didn't
step up to the plate and bridge the gap between what
KDE offered and a complete desktop OS system. SuSE
probably came closest, but even they fell short. (In that
sense maybe the Novell merger will turn out to be a good
thing for KDE, sinced it probably helped spur on this work.)
My only question about all this is can the KDE/Debian
initiatives be given their own project name? (KDE OS?
Kebian? hehe) Seriously I think the association with
UserLinux will end up as an historical footnote. Why not
move now to brand the new project so it has a more
recognizable identity, independent of UserLinux?
UsereLinux is bollocks. Enterprises are never going to use it or even get to trusting something like that. Since proprietary software will be developed for it anyway, if by some miracle it is successful, the logic behind it is totally flawed. I think many in the free software/open source communities, with no experience outside, are very, very deluded.
a sub porject of Debian ... nothing more.
Unfortunately it seems like UL is trying to be YET ANOTHER distro ...
Eugenia from OSnews said it best
here and
here.
So, a distribution chooses to not ship KDE, and it's a "debacle"? Wow. There sure is a heck of a lot of noise coming out of KDE on issues like this at the moment, and it's not looking professional or justified at all. It's KDE people who are always pointing out that more distros ship KDE as the default desktop environment... Do you hear the GNOME people whinging about it? Nuh-uh, they just get on with the job.
I hope you are aware that the wording is mostly used by the "press" and other rather unrelated and univolved audiences. If you care to dig into the KDE developer community you'll find out that the vast majority doesn't really care and does business as usual.
Ah, but it was core KDE developers who were involved in, and largely instigating, the whole mailing list mess. :-)
I counted two developers who happen to be involved with Debian as well, so they were affected both ways (no business as usual possible in that case). You don't expect them to abandon one side just due to Bruce's decision, do you?
Certainly didn't expect the behaviour.
What "the" behavior?
...about UserLinux standardizing on GNOME/Gtk because it's "easier for enterprise developers" is that there are no VB-like tools based on Gtk, and at least two nearly mature ones for Qt. I have yet to find a business with more than 100 or so employees with no VB apps, and those VB coders aren't suddenly going to pick up C overnight. And then of course there's Access [shudder], which isn't really duplicated at all under Linux but at least KDE has Rekall if people can stomach a little python.
And no, Glade does *not* count. (Neither does wxBasic, at least till they come up with an IDE like VB.)
I wonder how the UL and Ximian guys will react when HBasic or Gambas hits 1.0 (which will, in all likelihood, occur sometime this month.)
http://www.go-mono.com/
http://www.gtk-sharp.sf.net/
http://www.eclipse.org/
While I agree with you on the VB like apps, I don't agree on the point "easier for enterprise developers".
When you start a project with eg. Gambas, then the code must be GPL if you distribute it, as it is linked against the GPL-QT version.
You now might state, that you can keep it in your company, so it's not a "distribution". But how many companies have subisdiaries in the same contry or even overseas. I would need a lawyer here, but from my point of view, this is then distribution if the same tool is used in these subsidiaries. Therefore it then must be GPL.
Hbasic - qt/.Net based
http://hbasic.sf.net
Gambas - gtk based
http://gambas.sf.net
fruitation ==> fruition. Please see dictionary.com for more information.
I never heard of them ?
What can we do with them ?
Any link ?
First, let me note that I'm historically a (classic) Machead (and developer), my views are not well aligned with any UNIX party line. I'd take Copland/Gershwin over Jaguar/Panther anytime. Recently, I was "thrown into cold water" and had to port the 2.4 kernel to some embedded hardware, so I gained a few insights into Linux development. Here's my opinion on the QT/UserLinux/Desktop stuff (and a bit more):
1.) Desktop choice
Funnily, I run non-default desktops: KDE 3.1 on Redhat 9 at work, and Gnome 2.4 on Mandrake Cooker at home. Reason for KDE was that Gnome 2.2 was sluggish and so feature-poor that it was no fun. Gnome 2.4 has caught up well enough to be usable, and now I prefer the "mousefeeling" of Gnome (to bend the McD-Testers "mouthfeeling" impression a bit). I have no idea how quantify this, and I may be biased because Nautilus has been done by Ex-Apple-Finder people. But it certainly has something to do with a cleaned up and simple appearance. And, not trolling, I feel KDE is closer to the spirit of Windows, while Gnome leans ever closer towards the Mac (spatial Nautilus!). One example is the OK-Button order where Windows diverged from the Mac way for no good reason (well, possibly because of ancient SAA-Key-Controls, or just to look different). Thought food, maybe.
2.) The QT-not-LGPL issue
Forget your QT-centric arguments. Period. No matter how much QT is worth and businesses can fork up. It's not how it works out there. I code for some large companies with enough petty cash to buy the whole Troll(.no) shop, and the QT licensing is a TOTAL showstopper in any engineering decision process. Problem is that these large companies have a Dell|HP|whatever Windows-only default policy as a decree from the "decisionmakers". Anything else will have to go through a purchase process with these people. That will put engineers on the radar of the PHBs, stupid questions ensue, and you end up with a "do it with our Windows tools, and use our new [whatever-XML-tool], btw." order. No logic, no economics, politics.
Unless QT salespeople win a whole shop over from the top, Linux development slips in from the grassroots, and engineers will chose a freely available and DEPLOYABLE solution because of said politics. By the time, the slip-in is a success to stay, the GTK/Gnome chain is entrenched. And if it fails, there is no evidence or commitment in the books. While with QT, the proponents would be scapegoats, a la "You made us buy these tools and now you don't deliver." (Actually that can happen even if they do deliver...). For these very reasons, some hardware production support software I work on will be done on PyGTK (plus, it can be rather easily put on Windoze laptops for field support).
All similar to the VHS story, where the superior formats (2000, Beta) were license encumbered and lost out. Qt was the first _decent_ toolkit, but by now GTK is good enough (and so is Glade, though it sucks goats).
3.) What the KDE team should do (just from _my_ point of view)
Scratch the remaining itches. Five different demo disks, Knoppix, MandrakeMove, some Slackware-based stuff, whatever, don't make a difference, if the stuff doesn't work great beyond the CD its on. That includes system administration and software deployment. KDE went over the well-established Motif standard and came up with something new. It may be time to think bold again. Here's my totally uneducated masterplan:
- split up KControl into KControl and KAdmin. KControl does everything the non-privileged user wants to do in his session.
- KAdmin controls the system (and requires root): Clock, Firewall, services, mounts, global shares, and whatever drakconf or yast does. In a pleasing manner, not mirroring *.conf cryptics.
- Unlike XST/GST, don't use backends for KAdmin. Stick to the LSB (or so), accomodate slight differences where easy, but don't cater for an exotic variety. So there's one way, and that can be brought to work perfectly in this terrible mess of system configuration.
- Provide a KDE reference platform for KAdmin. This will also be used to create the "Official KDE Reference CD". At this point, welcome to the distro business, btw.
- Now rethink packaging. While yum or apt-get might be your friends here, if I want to install an application, I want to install "Application 1.23" in German, and not appl-1.23-9kdx, appl-i18n-de-1.23-9kdx, libappl-1.23-9kdx, and libwslinfro-0.91-1kdx (that was "lib-whatever support lib is needed for real operation", btw). Mark "main" packages, give them clear names. For initial distribution, metapackages will be needed if the fine granularity below must stay. Pack everything not in the base system into the meta-packages provided for download.
- I've given away above, that I'd call the whole lot "KDX". A nice, industrially compatible, buzzword-ready TLA. Canonical software might be named like "Kate 2.2 of KDX", and third-party software could be like "Scribus 1.2 for KDX". The end result is that I download ONE file with that name (or maybe application-1.23.kdx), the installer takes care of not overwriting a newer version of libwslinfro, sees that I'm German, and stuffs in appl-i18n-de-1.23-9kdx. I have a tea while the progress bar wiggles along, and when I come back, the app is ready to run.
- All this DOES need the commitment to throw a certain amount of (script file etc.) compatibility overboard, but it CAN be done in a way, where glue to legacy systems can be maintained. It might actually provide convenience to "common" distros, where "kdx-to-dpkg|rpm|tgz" tools would totally automate packaging beyond the pain of distro-specific specfile maintenance. I don't think this break with legacy is much bigger than what KDE originally did with the accepted Motif.
Note: With C++ and the fragile base class problem (and mentioned autoconversions), this might mean "source packages for deployment". With the amount of iBooks and upcoming AMD64 systems, this might actually help a lot.
- When all bits are in place, make bootable ISOs for the popular platforms that work nicely as CD-boot systems with an USB dongle, which can "infect" clueless PCs through a friendly graphical installer (we're talking Desktop, NOT headless MIPS-based 1" servers). Et voila. World Domination.
Well, almost. Friggin lobby the trolls to LGPL at least their Linux QT. Even with all the shit above, the other side is "good enough". If people were willing to pay a bit extra for much better computers, we'd be all using Macs now.
Thanks for reading and hope you enjoyed the rant!
Rich
haha,
qt actually has a non-Linux market. And there are no gtk applications out there.
qt creates an incentive to license your programs as gpl whereas in the gnome world we see an incompatible license mix.
Qt is not expensive and used by many companys that even don't serve the Linux or KDE market. It's nothing special for a company to buy a toolkit such as they buy VisualC, Delphi and so on.
try directing your plentiful energy towards raising awareness and education. You will need these skills once you are sent down to fight in the corporate trenches. (Reading Dilbert helps understanding, too).
> qt actually has a non-Linux market
So has Borland. I like Delphi. It's just not mainstream. Heck, people still keep the Amiga alive. But it is entirely irrelevant, as this discussion is about a Linux distribution with an appeal as broad as possible.
> And there are no gtk applications out there
I assume that in your eagerness to reply, you forgot "on Windows". Well, except for our little in-house stuff. And maybe the Gimp. Or Pan. I like Pan.
> qt creates an incentive to license your programs as gpl
Young friend, whether I have an incentive or not is totally irrelevant. I am not in the position to license out anything at all, for I just carry my skin to market in exchange for a little money. The people who give me money 0wN everything I do for them in return. This decision is made by corporate lawyers in a country far, far away, behind a large ocean. Lawyers may look like people in suits, but really they're totally different from us people inside. I just know I'd be royally screwed if I did anything these lawyers don't agree with.
> It's nothing special for a company to buy a toolkit
Oh yes. Just yesterday I saw a big industrial company walk into the software shop and pick up a copy of Mandrake 9.2 Powerpack. I thought, *wow*, what a cool company not to get Windows XP. Well. Fun aside. You must not only be young, but in a very remote location, too, to not have heard about the evil forces summoned by trying to purchase ANYTHING in a big industrial company. Two vertical layers of approval plus providing a vendor certification is for entry level skills. This gets you about a nice ballpen and a sketchpad. With software, other obstacles get involved, from license management to the IT department.
Now get this in your tiny head: In big companies, Programmers don't decide about toolkits. "Decisionmakers" do. They do so by looking at a glossy 3-page-foldout advertising the toolkit and then deciding. It helps if the foldout has a lot of blue that matches their suit and says "Microsoft".
You CAN NOT run a QT purchase past these "Decisionmakers" as long as it's not corporate policy to use it. The only way to get Linux in at all is for the people doing the work to take what they can get for free and just make it work. In most cases, the reality is not really acknowledged, but accepted once it works.