Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
QT/Embedded to go GPL, provide credible alternative to X?

By 11223 in Technology
Tue Oct 31, 2000 at 02:19:11 AM EST
Tags: Software (all tags)
Software

Trolltech has announced that it will be putting it's QT/Embedded product under the GPL. Their product provides such holy grails as alpha blending. With all the work going on to replace or extend X in one way or another, could Trolltech have just provided us with the next direction for the Linux desktop?


(Please note: I am not an advertiser for Trolltech. I'm simply intruiged by the idea of a non-X desktop for Linux.)

The KDE developers have expressed some interest in porting their desktop over to QT/Embedded. If that happens, an already existing desktop/office suite will suddenly appear on a new non-X platform for Linux, perhaps giving us the tools to finally dump the X-Windows Disaster. Would this be dumping too much legacy software? Or is this the key to massive Linux adoption on the desktop?

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o announced
o the X-Windows Disaster
o Also by 11223


Display: Sort:
QT/Embedded to go GPL, provide credible alternative to X? | 12 comments (12 topical, editorial, 0 hidden)
Just so you don't take away my X (4.00 / 11) (#1)
by vsync on Mon Oct 30, 2000 at 10:06:50 PM EST

The KDE developers have expressed some interest in porting their desktop over to QT/Embedded. If that happens, an already existing desktop/office suite will suddenly appear on a new non-X platform for Linux, perhaps giving us the tools to finally dump the X-Windows Disaster.

That's fine with me, as long as they leave KDE compatible with the normal Qt and stuff. I actually like X, other than its little quirky shortcomings like fonts. I routinely display apps between different machines, and that's a beautiful thing. If you take that functionality away from me, I will come after you.

I actually can't see why people get so confused or hateful of X, other than not being able to wrap their minds around the whole "client"/"server" thing. It's not that hard, though, and should be easier now that people are used to Napster and such. The server is the thing that provides resources (in this case, a display and input devices). The client is the thing that uses those resources (the program that writes to that display).

Would this be dumping too much legacy software? Or is this the key to massive Linux adoption on the desktop?

Well, you wouldn't necessarily have to dump legacy software. As mentioned in their FAQ, the Berlin folks do have an X11 compatibility layer in mind.

Anyway, I don't think this would make much difference one way or the other as far as popularity goes. The people who are intimidated by X aren't the people who would tend to run Linux anyway, and the more "user-friendly" Linux setups usually go to some trouble to hide the complexity from the user. Basically, the ones who know are used to it, and the ones who don't know don't care.

--
"The problem I had with the story, before I even finished reading, was the copious attribution of thoughts and ideas to vsync. What made it worse was the ones attributed to him were the only ones that made any sense whatsoever."

berlin needs to exist, asap (3.40 / 5) (#2)
by sayke on Mon Oct 30, 2000 at 11:48:39 PM EST

i contributed a bugfix ages ago, but i havn't messed with it since... and its come a long way, to be sure, but its potential is incredible. if there ever was a worthy open source project, it would be this... i beg everybody who has spare time to hack on it.


sayke, v2.3.1 /* i am the middle finger of the invisible hand */
[ Parent ]

hmm (3.50 / 4) (#3)
by phexro on Tue Oct 31, 2000 at 03:17:17 PM EST

they don't say that Qt/E works without X. of course, they don't say that it requires X either.

X11 is a pretty big overhead for an embedded device; 6-8mb of ram usage, & 2-8 mb of disk space, depending on what fonts you install.

it would be interesting if Qt/E didn't need X, but then you have a compatibility problem, since you won't be able to run standard X apps under Qt/E.

guess i'll have to download the source and see for myself. :^)

Losing X (4.00 / 5) (#5)
by chuq_r on Tue Oct 31, 2000 at 03:20:53 PM EST

I'm all for getting rid of X, provided that the alternative is a better design. I'm really, really getting tired of XFree86 4.01 taking up, on average, about 40 megabytes of core on my system and being slow as a tortoise. (And yes, that is the amount it takes up after you subtract the amount of video RAM on the video card that ps and top report in their memory usage statistics. And no, I can't/won't go back to 3.3.x for various subtle reasons, not that earlier releases are all that much better.) That just seems excessive for what amounts to being a 2d rasterizer. I don't have much of a use for X's network transparency, and in this day and age where everyone can have their own Unix workstation for cheap, I can't imagine that most other people need that either. And with things like DGA and such going into XFree these days, I'm beginning to see less of a point to X as DGA routines don't work over networks. (That's the point of them; direct low-level hardware access.)

If QT/Embedded is low on RAM and high on performance, I don't think I'd have much of a problem using it or developing for it. But it's my feeling that something should be done. Throwing more and more RAM and more and more cycles at a problem doesn't seem like it is the whole answer to woes like these. I believe that people should start paying attention more to the efficiency of the code they're writing as well as how elegant of a solution it is. Even the best optimizing compiler in the world can't always help (much) with inefficient code. Elegance be damned if it's wasteful. ;)

Chuq

Gtk will have the same facility soon... (4.00 / 4) (#6)
by Paul Crowley on Wed Nov 01, 2000 at 09:15:37 AM EST

I've seen a demo of framebuffer Gtk, though I don't believe it's complete yet. So that could get rid of X for us too. Actually the thing I most want to see built is framebuffer Mozilla. Also check out Berlin
--
Paul Crowley aka ciphergoth. Crypto and sex politics. Diary.
Wish list for X (3.75 / 4) (#7)
by ObeseWhale on Wed Nov 01, 2000 at 10:20:44 AM EST

This is my wish list for a new, better, X for the desktop.

1) Clean, efficient, fast. Whether you like the OS or not, one thing you have to admit is that the Be Operating System did absolutely amazing things with their display code. Linux NEEDS a rasterizer/display manager with the kind of lean efficiency we saw in Be.

2)Proper copy and paste management. Currently you can't just "copy" an image out of Netscape and go ahead and paste it into your AbiWord document. I believe the GNOME people are doing something along these lines with Bonobo, but such a system might only work with GNOME apps (I'm a little unclear here, if someone could clarify this issue). We need direct, native support for Windows-style cut and paste regardless of whether someone is coding Tcl, GNOME, GTK+ or qt.

3)Interaction between applications. I should be able to drag an object in one window/app to another and see results. This would make a lot of sense with images and such.

4)Good driver support for 3d cards. DRI is definately getting there.

5)Native TTF font support. People might disagree with me on this, but I know when I moved to Linux and opened up Netscape for the first time, browsing was NOT fun.

---

"The hunger for liberty may he suppressed for a time; yet never exterminated. Man's natural instinct is for freedom, and no power on earth can succeed in crushing it for very long."
-Alexander Berkman
Re: Wish list for X (3.50 / 2) (#9)
by Ronin SpoilSpot on Wed Nov 01, 2000 at 05:08:57 PM EST

Ad 1) Nobody is going to argue against Clean, Efficient, and Fast, except that Clean might not be as unambiguous as it sounds.
Ad 2) Hell no! Copy and paste does NOT have to be a part of the graphical interface. The reason that windows can cut and paste graphics is that the applications know how to do it. If they didn't, the graphical interface wouldn't be able to do it. It would be restricted to screen-grabs. Keep the cut&paste buffer OUT of the graphical interface, and allow the GUI to access it just as any other application.
Ad 3) Again, that is at application level, not at GUI level. I though you said you wanted a Clean display manager :). All that is required for that is a common way of representing the data, like... OLE? You want windows-style cut&paste... you got it, when using Gnome programs. What you really mean is that you want a SINGLE windows-style cut&paste model... i.e., windows is good because it doesn't have competetion.
Ad 4) Driver support? Building something to match the hardware is rarely a far-sighted move. Design the protocols and designs first, then make implementations. The implementations care about drivers, not the design!
Ad 5) Native scalable font support, TTF is just one possiblity. But again, it
should be abstracted into a seperate program, just as the current font servers. There are font-servers for X that support TTF through the freetype implementation.

All in all, you sound like what you really want is another Windows Desktop, only faster, or perhaps I'm reading too much into it. Maybe you just want ONE standard so your programs could communicate. Sure, great idea, until someone thinks they can do it better. That is what is happening now.

X is not a desktop, and nor do I think a replacement should be.
I might want a display manager, perhaps without the network-possibilities of X, though that is one of the strengths of X. I sure want a better clipboard, and a scalable font-server (gawd, do I want that!!!), but you could have those already on X if your applications supported them, as the Gnome stuff or the KDE stuff or whatever other implementation already out there.

/SS
"This space intentionally left blank"
[ Parent ]
Some wishes fulfilled... (3.50 / 2) (#10)
by meldroc on Wed Nov 01, 2000 at 06:19:20 PM EST

1. Clean, efficient, fast. I agree with that - I think the X11 architecture is a bit antiquated. What needs to happen if X is to continue to keep up in the GUI marketplace is to revise the protocol, call it X12 - put newer features into the core - 3D and other acceleration support, alpha blending, anti-aliasing for fonts, clean out some of the legacy junk (does anyone use monochrome X11 displays anymore?)

2. Proper copy and paste. As was stated, this is an application problem, not an X11 problem. I'd recommend you pick one of the desktop environments (KDE or GNOME) and use it exclusively - KDE apps know how to cut & paste with each other, so do GNOME apps.

3. Interaction between apps. Again, pick KDE or GNOME and use it consistently.

4. Good driver support for 3-D cards - Getting there, but again I think the X Window system needs to be revised - time for X12. That or replaced.

5. Native TTF font support: XFree86 4.x has this, Red Hat Linux 6.x and up have TTF font support built into X. The tricky part is finding fonts (stealing fonts from Windows is easiest) and installing them so they're usable. Read the XFree86 Font Deuglification HOWTO for help with this.

[ Parent ]

Copy and Paste? UPGRADE! (3.00 / 2) (#12)
by dgwatson on Wed Nov 01, 2000 at 10:18:20 PM EST

I'm using KDE2 now (actually have been doing so for months), and I'm happy to report that copy/paste works perfectly between all the apps I've used - KDE, GNOME, and crappy^WMotif (like Netscape and RealPlayer).

[ Parent ]
Thanks for the font advice... (3.00 / 2) (#13)
by the coose on Wed Nov 01, 2000 at 11:10:32 PM EST

5. Native TTF font support: XFree86 4.x has this, Red Hat Linux 6.x and up have TTF font support built into X. The tricky part is finding fonts (stealing fonts from Windows is easiest) and installing them so they're usable. Read the XFree86 Font Deuglification HOWTO for help with this.

Wow, thanks for that link to the HOWTO. I would suggest that anybody having gripes with X and fonts go read that. Here's the link: http://www.linuxdoc.org/HOWTO/mini/FDU/. It provides some tweaks to existing X fonts as well as installing true type fonts. It was easy for me since I'm running xfree86 4.01 but 3.3.x users are covered as well.

[ Parent ]
What I'd really like to see ... (4.33 / 3) (#8)
by XScott on Wed Nov 01, 2000 at 03:45:46 PM EST

... is a pure OpenGL desktop environment (hopefully on top of a BSD or Linux, but OpenGL is pretty widespread, and I could imagine it working on other platforms). The latest hardware accelerated video cards are reasonably cheap, but dropping a VooDoo 3 in a slow box would bring most drawing operations to life. The operations are well defined, and the free software world could use Mesa as a starting point.

I imagine it working along the lines of having a predetermined view volume for the applications to draw in. The user Alt-Tabs between which application gets to draw into the view volume. Some well behaved applications (clocks etc...) get to make additional contributions after the application has filled in its details. Add a simple event model for the keyboard and mouse and you're good to go.

It could be client server (like X), but it wouldn't have to be. Marshalling large texture maps across a network might need compression.

A first and obvious application would be an XTerm type of thing. I imagine an MDI application where minimizing a terminal just pushes it off into the distance. This would be a lot like the Berlin stuff, except using 3D instead of 2D linear transformations.

A lot of legacy apps could be ported by having them draw into a 2D texture that maps onto a rectangular polygon. This is especially true of applications like Gnumeric which already use things like the Gnome Canvas for their drawing operations.

As an alternate way of switching between applications, maybe the applications that don't have the focus (occupy the main view volume) could draw into little miniature view volumes across the bottom (or side) of the screen. Click on one of the mini-volumes and it becomes the active one.

Someone is going to come up with a good 3D desktop someday. I'd greatly prefer it came out of the free software world using an open standard.

Ok, quick! Get to work everybody, and let me know when you're done. :-)


-- Of course I think I'm right. If I thought I was wrong, I'd change my mind.
X should be fixed, not dumped (3.66 / 3) (#11)
by mihalis on Wed Nov 01, 2000 at 07:36:33 PM EST

First off, full disclosure, I happen to quite like X. I like having a little LAN in my apartment and only needing one monitor for my workstation and my little webserver box.

Still, X does have its flaws, even though many of the problems mentioned in the the X-Windows Disaster are gone. The things that make me happy enough with it though in spite of its flaws are a) all the software that already works with it, and b) people are working on trying to fix X - see Keith Packard's work on a new rendering model for X here and here and a minimal complete X-server here (I think Jim Gettys is doing some work there too).
-- Chris Morgan <see em at mihalis dot net>

QT/Embedded to go GPL, provide credible alternative to X? | 12 comments (12 topical, 0 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!