Anders Carlsson of Imendio outlines and explains (copy of the text) the API changes that are coming with the Gnome 3.0 release in the future. On other toolkit news, GTK# 1.0.4 and beta 1.9.0 (wrapping around Gnome 2.6) was released recently. You can rate GTK# here. Update: On yet another toolkit news, the latest release of the Ruby language bindings for Gnome, adds support for GnomePrint, GnomePrintUI, librsvg and Atk 1.7. Rate Ruby-Gnome here.
The Upcoming Changes of the Gnome Subsystem
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
49 Comments
2004-11-15 7:11 pm
2004-11-15 7:17 pm
2004-11-15 7:32 pm
2004-11-15 7:48 pm
2004-11-15 8:19 pm
2004-11-15 8:21 pm
2004-11-15 8:42 pm
2004-11-15 11:03 pm
2004-11-16 5:31 am
And now all that is left is a broken linked wasteland… really tho…. links werky maybe please?
Integration and simple APIS are the main goals.
Is amazing how good GNOME is despite the fact the APIS are 5 years old.
For me, the first two links in this news clip are dead. What gives? Is this on my end?
The APIs aren’t generally 5 years old. *Some* of the APIs are 5 years old. libgnome and libgnomeui are gradually being phased out in favor of moving UI functionality into GTK+, so a lot of parts of libgnome aren’t well-maintained.
http://people.imendio.com/
The links went down, yes. Probably an automatic measure on their server if too many hits get there at the same time. Try this instead: http://people.imendio.com/
It’s not GNOME 3.0, it’s GNOME 2.10.
I don’t think they’ll bump the version to 3.0 until GTK 3, which would break backwards compatibility. Note, there’s no plans for this at the moment, so things will continue in the 2.x series for the forseeable future (2.10, 2.12, 2.14, etc…).
We know the next version it’s 2.10. The API changes is for 3.0, whenever this will come out.
Go to http://people.imendio.com and read the (at the current moment) first 2 articles dated November 14, 2004 11:36 AM (outlines) and November 14, 2004 11:13 PM (explains).
It’s not GNOME 3.0, it’s GNOME 2.10.
I don’t think they’ll bump the version to 3.0 until GTK 3, which would break backwards compatibility.
Moving stuff into GTK and breaking backwards compatibility is exactly what Anders is talking about. And that’s why he refers to it as something for 3.0.
I’m not an expert on these API’s, but I wonder if it’s a good thing to do: breaking backwards compatibility. Backwards compatibility is one thing that made Windows so populair, both for users and for programmers.
Quote:
“I’m not an expert on these API’s, but I wonder if it’s a good thing to do: breaking backwards compatibility. Backwards compatibility is one thing that made Windows so populair, both for users and for programmers.”
Well, sometimes, there’s no other choice.
In order to make something better, you can’t always use the current “state” anymore.
In such cases, it’s better to change it.
Of course, you don’t do this every day.
Only with major releases.
And… it doesn’t necessarely need to break backwards compatability though, but in a lot of cases it does.
gtk only breaks binary compatibility better major releases and it does so only when its obsolutely required to move forward. its not a big thing for app developers because previous versions are designed to be parallely installable so your gtk 1.x and 2.x will continue to work.
http://ometer.com/parallel.html
remember that windows *has* broken app compatibility several times with the recent xp sp2 pack being a prominent example and .net rewrite being another major one
if apps were as much careful about breaking compatibility as gtk the situation would be much better
the worse offenders today seems to be glibc and automake
Even if you break compatibility, it’s no big deal, as the new and the old versions can co-exist peacefully.
It’s not like you can’t use GTK1 apps on a Gnome2 deskop.
Seems reasonable enough to me. Yes, you’re gonna see some breakage, but the change he seemed to be proposing didn’t strike me as all that difficult to implement.
The general crux of moving off libgnome is greatly appreciated. I do quite a bit of work with Mono and GTK#, and I often find myself wondering whether it makes sense to make use of the gnome libs (due to portability concerns).
In any event, GNOME 3.0 is still a bit off, I think.
-Erwos
i have bashed Gnome desktop in the recent past, dont get me wrong i dont hate Gnome, i think it is great just that it is difficult to compile from source and so HUGE even comared to even KDE, i would love to see it get slimmed down, features like drag & drop menu editing and Desktop icons created by drag & drop too. Gnome should not be so paranoid about user space, since if a user trashes his desktop all ya gotta do is delete the GNome related files in /home/*user and the next time the user logs on they get the default desktop again rebuilt by the root system (not like a user is going to bork the whole system)…
some of the BEST apps are GTK
If breaking compatibility means a substantial improvement in the overall API in terms of performance and flexibility, then I am all for it!!
It’s not like you can’t use GTK1 apps on a Gnome2 deskop.
And at least we won’t have ugly-ass, non-AA fonts to deal with when running legacy Gnome 2 apps on a Gnome 3 system.
What is the difference between osnews.com and planet.gnome.org? Honestly, if such two postings are worth a “news story”.
Well, planet.gnome.org is exactly that, a planet. All it does is syndicates gnome related blogs, of which the imendio postings are.
So planet.gnome.org is only posting the blog entries, osnews is featuring the very interesting topic as a news item.
Putting all the gnome stuff into gtk is a bad thing, as it makes it far more closely tied to gtk. They should try and abstract libgnome from gtk more, and work towards a common toolkit abstraction with kde. Then it might one day be possible to have them both use the same widget toolkit, which would be a good thing. Diversity is good, but only when it’s compatiable diversity – a future where you could choose gnome or kde and *independently* choose qt or gtk would be great.
>Putting all the gnome stuff into gtk is a bad thing, as it makes it far more closely tied to gtk.
I disagree. By putting things into GTK makes the platform more COHERENT and EASIER to develop for, because this way you don’t have to have two different functions that do effectively the same thing. You have ONE API that takes care the *platform*. And yes, this is a very good thing for Gnome devs, and if something is good for devs, it means that more apps will be developed, and this means that users will be happier.
This also means of course that GTk-only apps like Bluefish, gFTP, AbiWord, Sylpheed etc, won’t be able to use the new GTK version anymore, as it would require the Gnome subsystem to be installed, instead of just GTK. IMHO this is a good thing for Unix and Gnome users because if the devs of these apps decide to move to Gtk 3.x (that’s a version number assumption btw), Gnome users will get desktop integration for free with these previously gtk-only apps. The people who will actually “pay” for the transition will be the Windows users instead. But these apps don’t really shine on Windows (except Abiword which uses native Win32 code anyway for its Windows version), so the problem is NOT big at all.
I look forward to see ONE API for the Gnome platform that actually stays source-compatible from that point on. I don’t want to see more breakage for gnome 4.x for example. Do the integration RIGHT for Gnome 3.x and don’t break it again.
So it looks like, GNOME 3 and KDE 4 will happen in a predictable future. I think number one priority should be integration.
e.g use print/file dialogs from the actual DE, or use current UI renderer (Something like Gtk-Qt did).
Both, KDE and GNOME will break backwards compatibility in their next major release, so it would be the right time.
> But these apps don’t really shine on Windows (except Abiword
> which uses native Win32 code anyway for its Windows
> version), so the problem is NOT big at all.
Why do you think, they will brake Windows compatibility? And there are apps with a lot of windows users, like Gimp, Inkscape, Gaim, Workrave or XChat.
Am I the only one who finds this strange? I have always been told one of the great things with Gnome are the fact it’s written in C and this is the reason you find so good bindings to other languages. But here we have the bindings for GTK#, and Mono being hailed as the future by several key Gnome developers, but the bindings are 7 months out of sync and not even considered stable. Why are the bindings for one of the favored languages so outdated?
How is this surprising? Advocates for a particular language tend to cheer their bindings and say that it’s the only way to go.
The Python and C++ bindings for GNOME also have a large followings and large advocacy groups.
>> But these apps don’t really shine on Windows (except Abiword
>> which uses native Win32 code anyway for its Windows
>> version), so the problem is NOT big at all.
>Why do you think, they will brake Windows compatibility? And >there are apps with a lot of windows users, like Gimp, >Inkscape, Gaim, Workrave or XChat.
It seems strange to me, too. I think that GTK+ is cross-platform and will always be. MS could support GTK+ if they wanted, and help us develop beautiful cross-platform apps.
Things could get better with Cairo and Longhorn, as both will support something that could be very compatible…
This also means of course that GTk-only apps like Bluefish, gFTP, AbiWord, Sylpheed etc, won’t be able to use the new GTK version anymore, as it would require the Gnome subsystem to be installed, instead of just GTK. IMHO this is a good thing for Unix and Gnome users because if the devs of these apps decide to move to Gtk 3.x (that’s a version number assumption btw), Gnome users will get desktop integration for free with these previously gtk-only apps. The people who will actually “pay” for the transition will be the Windows users instead.
The people who will pay are the current Sylpheed, etc. users and the developers. The reason that Sylpheed, gFTP, and Bluefish — and arguably Abiword — gained so much popularity is because they were so much lighter than the alternatives. People who want a large, pretty e-mail client are using Evolution. People who want a barebones [graphical] client are using Sylpheed. Sylpheed hasn’t been ported to GTK2 because GTK1 is lighter (and because the author doesn’t want to rewrite everything). Bluefish stayed GTK1 for a long time. A lot of people don’t want to have to install five thousand libraries and a desktop environment just to check their e-mail or type up an HTML document.
Morty, there are gtk+2.4 and Gnome 2.6 bindings in gtksharp cvs right now.
From now on gtksharp won’t be playing catchup as much because the basic bindings infrastructure has already been established.
a working smb browser in nautilus. They still don’t support browsing MS networks properly. You can authenticate against shares but how useful is that if you can’t browse the machine to see them in the first place. booo hisss.
@ Lumbergh
“He’s german and KDE is some sort of national pride for them”
There are a lot of German Gnome users out there… but SuSE is the best Distribution for German speaking Linux beginners.
@ Anonymous 14:55
I can brows my MS network.
what do you people that disagree with using GTK with Gnome think GTK means???!?!!?!?!?!
GTK = Gnome Tool Kit…
Gnome needs GTK, if you keep changing too much of the underbelly of Gnome it will always be in a state of development and never be ready for the avareage user, i think Gnome developers just need to focus on makeing Gnome more streamlined/lighter, easier to compile & install from source or binary (.bin maybe) how about a distro independant installer similar to ximian’s only better because a Linux user can install or upgrade their gnome no matter which Linux distro they are running now that would be nice, and the menu situation – maybe use the applications directory as a template to build a user directory so a user can edit their own menu without changeing other users menus and if something else gets borked, all they would have to do is empty the config files in their home directory and the next time they logon it is all rebuilt to a default desktop (much safer than making system wide changes that effect other users)
GTK = Gimp Toolkit
http://www.gtk.org/
>there are gtk+2.4 and Gnome 2.6 bindings in gtksharp cvs >right now.
Yes, and the announcement told of a beta release of the bindings for 2.6. Since 2.6 was released more than 7 months ago and current Gnome now are 2.8 are the facts I commented on.
>From now on gtksharp won’t be playing catchup as much >because the basic bindings infrastructure has already >been established.
Sounds great, when are the release for Gnome 2.8 planned?
Gnome 2.8 is still bleeding edge. Even freaking Gentoo doesn’t have it in the main repository. The important thing is that the gtk+2.4 bindings are out there. Remember, Mono is also for business and not just for bleeding-edge hobbyists. They need to make sure there are stable bindings before moving on.
As I stated before, there should be less lag time now that the bindings infrastructure is there.
On a related note, Mono’s managed winforms implementation seems to be moving full steam ahead.
I’ve been using Gnome 2.8 on gentoo for ages now. Maybe you should try “emerge sync” and up date your portage tree.
Like showing how it is now, which I’m still not sure about, and how it will be (proposed obviously). I’ve heard a lot abot merging stuff into gtk but I would like to know the details.
With pretty pictures too
I use xchat and gaim on Windows all the time – what about them?
Given that the largest Linux distributor, Red Hat, supports GNOME by default, I suggest that your assertion may be a little out of line with reality.
GTK = Gimp Toolkit