The first test version of the upcoming Gnome 2.10 is now available. In the meantime, discussion about proposed modules to join 2.10 is under way.
GNOME 2.9.1 Development Release Available
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
is it just me, or do people seem to be screaming for a lot of kde features?
http://lists.gnome.org/archives/desktop-devel-list/2004-September/m…
I would love to see this in 2.10
I want thumbnails to be rotatable to stop pictures lying on their sides, and also when you open them they also open the correct way too…
Why is this thing being called 2.10? That seems a bit, odd…
Probably because KDE has a lot of features 😉
Why is this thing being called 2.10? That seems a bit, odd…
Did we miss an even number between 8 and 10? I can’t think of any.
People, version numbers are not decimal numbers. The dot is just a seperator. 2.8 does not mean two and eight tenths. 2.9.1 doesn’t even make sense as a decimal number.
Ugh, not this one again. It’s GNOME “two-point-ten” not GNOME “two-point-one-oh”. The period is not necessarily a decimal point — the French use it as a seperator (100.000,00), for example.
It works like this; x.y.z
x = Major revision
y = Minor revision
z = Patch to the minor revision
So in the case of GNOME, it has had two major revisions, and 8 minor revisions so far, with the odd patch spread throughout the release cycle.
So the developers need to add a “feature” because your to lazy to rotate your pictures correctly in the first place?
How do you propose they do this? How would the software know which way the picture is supposed to go without you telling it ?
please engage your brain before whining about needed new features.
How would the software know which way the picture is supposed to go without you telling it?
Actually… In some cases, the image metadata will provide this information. Most photo management tools will automagically do the right thing. 🙂
In more detail – the standard for digital camera metadata, EXIF, contains a field to record the orientation of the camera when the picture was taken. As Jeff says, most digital cameras do this, so any application that reads the EXIF data can work out which way the picture should be oriented.
oh, and this is a Very Good Thing, because of recompression. Most pictures taken with digicams are already lossily compressed with jpeg. If you happen to take it with the camera rotated, and then because the application can’t discover that and show the picture correctly you perform a rotate operation and then re-save the file, you have re-compressed it, again lossily (unless you saved to a gigantic BMP or TIFF or something file this time), thus losing more quality. Which isn’t the best thing to do.
thanks for the info Jeff and Adam 🙂
Just seemed to me that something like this could lead to alot of issues with things like you wanted the picture rotated one way and the auto feature could end up forcing rotation when you didnt want it.
But Im not a developer just a user so its not going to be any extra work for me 🙂
Let’s hope one day drawing speed in Gnome, GTK, or whatever is responsible for sluggish window drawing is finally brought up to par.
Let’s hope one day drawing speed in Gnome, GTK, or whatever is responsible for sluggish window drawing is finally brought up to par.
That would probably be your video drivers. GTK+ makes heavy use of the Render extension to perform fast blitting and other features, so if you’re running an ATI card on the fglrx drivers (which do not perform hardware acceleration), or an nVidia card without explicitly enabling support (by setting Option “RenderAccel” “true” in xorg.conf), you’re going to get some very poor performance, especially with compositing. GTK+ is much more sensitive to this than Qt is, and I’m not really sure why.
I’ve noticed that on nVidia cards, if you don’t have both AGP Fast Writes and AGP Sideband Addressing enabled, Render doesn’t work quite right; it says it’s working, but it doesn’t appear to be hardware-accelerated. Maybe someone better-informed than I can shed some insight on this for me.
Where is the list of the new features and fixes?
The problem has been identified. It’s Pango, and they are supposedly accepting patches for a “fast-path” for western languages.
It’s still sluggish even with RenderAccel, FW and SBA enabled… and I have a Ti4200-8x. Anyway, the toolkit feels slow even in MS Windows… Try GAIM or GIMP and you will see by yourself.
It has nothing to do with GNOME but it’s huring its reputation. That’s quite sad.
I would like to see a search function and filter in Nautilus.
I would like to see a menu restructuring similary to what Ubuntu has done with their Applications/Computer menu. It really cleans everything up and seperates applications from system configurations. Seems like something the HIG might agree with…
I really hope Gnome doesn’t go ape-$&it with features like KDE did. Gnome has always felt lighter, simpler, and smoother than KDE, IMO…I hope it stays simple.
On another note…it looks like us Gentoo geeks may see a 2.10 release before we even get a safe look at 2.8!
Gnome must have some gnarly implementation issues for (some) distros to have such trouble w/ it?
Ubuntu’s Preview release came out on the *same day* as GNOME 2.8. Although GNOME is a large, complex software stack (as you’d imagine, for a complete graphical desktop), it is not outrageously difficult to integrate and ship.
GNOME 2.8 has a lot of issues. None of them however are show stoppers. You could try out the testing branch of Gentoo if you are interested in the latest release.
Depending on the kernel you use, the type of filesystem you use(i.e reiser4, reiserfs, etc) and your device management system (to use version 2.8 you have to move from devfs to udev), you may or may not encounter problems.
The upgrade from 2.6 to 2.8 was relatively painless, but it wasn’t without problems.
Nobody’s suggested Autopackage for inclusion yet? GNOME needs a distro-independent software installer. I mean, people DO like to be able to install new software on their Linux systems without having to jump through firey loops hop over hungry alligators. Dependency issues are a pain in the ass and for the newbies out there, they really don’t like command line tools to get what they want. Don’t get me wrong, I like apt/synaptic, but speaking for my mother, I don’t think she wants to scroll through a huge list of nonsensical crap so she can download a new solitaire game. Believe it or not, it actually makes more sense to go to a website, download a package, double-click on it, and the packages installs itself.
Howdy all
Does anyone know if this is still on track to actually happen?
My preference is KDE but Gnome always seemed more polished and if the two could one day integrate nicely this would be a god send to all of the open source desktop users.
Oops i should qualify integration aswell, i mean as in a plays-nicely-with kinda sharing gooey way, not that they`ll merge etc etc .. just getting in before the trolls
P.S. on a personal note and sorry ofr being off topic, Eugenia is there any plans to have some form of membership to OSNews?
Just to stop others posting with a member`s alias etc etc ?
I know you are already corrected, but I thought I should point out what David actually said:
“I want thumbnails to be rotatable to stop pictures lying on their sides, and also when you open them they also open the correct way too…”
If you read carefully, you will notice he wasn’t actually suggesting having it done automatically. It seems he wants to manually set the rotation in the file manager and then have it pass this information to the viewer/editor when the file is opened from the file manager.
BTW I am very surprised that the GNOME file manager doesn’t already process the EXIF data. If nothing else, it could speed up the thumbnail generation process, if a suitable thumbnail is included in the EXIF data.
Don’t bother. Your mother probably doesn’t know what GNOME 2.6 or 2.8 is. She probably doesn’t know she is running GNOME. And even if she wanted to install games, assuming the solitaire clones in gnome-games aren’t enough for her, she probably won’t know where to begin searching.
She probably doesn’t even care for or know about upgrading GNOME biannually. That’s the reason autopackage won’t go far. All mummy needs to do is click on that “upgrade the system” button over there on the panel, and her distribution’s package manager should handle the nastiness of package resolution, package hunting and packaging from her.
That’s the reason package managers exist afterall.
This has been happening, and it will continue to happen. KDE, Gnome, Rox, and various other desktops have been working on shared components and shared technology on freedesktop.org. With each release, the desktops are more and more interoperable.
There won’t be any release where we say “This is the KDE integration release.” It’s a gradual process.
I would like to see a search function and filter in Nautilus.
The filter is there. Use control+s on a nautilus window. Also available in the “Edit” menu.
Yes, I am aware of that feature. But its functionality is insufficient. I wouldn’t even call it a filter. It is a pattern matcher. You enter commands using regular expression characters, then Nautilus proceeds to select the best possible files that match your criteria.
Unfortunately, not many people understand how to use regular expressions or pattern matching commands. In addition, a search utility is more forgiving, more intelligent, more convenient and a lot more easier to use than a pattern matching function.
Finally, even when using Nautilus’ pattern matching function, the files that match your criteria are only selected, not filtered. It’s just not usable, especially when you have folders containing a large number of files.
Is it just my impression or is gnome churning out releases more often? If so are these releases as feature-packed as before?
Basically I’m trying to understand whether commercial investment is having a significant effect on gnome’s development.
On the matter of versioning, can ppl stop being such dorks concerning .10 releases. Why do we have to have the same discussion every time. Version numbers aren’t meant to be decimals that you can simplify at your wim. NetBSD 1.62 isn’t NetBSD 1.6. It’s two dot ten. It happens, live with it.
Since the release of Gnome2, the project has adopted a 6-month release cycle to ensure there isn’t the long period without a stable release that occured between 1.4 and 2. This means that the development version only takes on board working features ensuring a smoother transition from development to stable and hence a better end result for users.
“Is it just my impression or is gnome churning out releases more often? If so are these releases as feature-packed as before? ”
This is because they switched to time-based releases rather than feature-based releases after GNOME 2.0.
This means that they release every six months, and put in the features/enhancements that are ready, rather than set a target of features/enhancements and release when they are ready.
This is keeping in line with the “Release early, release often” philosophy of free software.
Some features might have to wait another cycle to get in, but it means that releases that contains many enhancements and bug-fixes won’t be put on hold just because some wanted feature is missing.
This is keeping in line with the “Release early, release often” philosophy of free software.
Actually, that’s more Open Source philosophy than Free Software philosophy. Think Linux vs. Hurd. The Cathedral and the Bazaar by ESR was not just targeted at proprietary software development, but also monolithic Free Software projects.
I want to be able to open/save files on ftp, sftp, smb, http etc. via the open/save dialog without having to mount the filesystem first.
At the pace integration is being pursued in the Linux desktop, maybe in 50 years we will have something.
It is a pity several opportunities were lost along the way regarding a common component model shared between the major applications/platforms. Without such a thing there will never be true integration, I think. In the GNOME field, CORBA/Bonobo seems to be dying. In KDE, DCOP/KParts seem to be doing very well, but they’re very KDE-specific. OpenOffice.org and Mozilla each have their own incompatible component models. Also, maybe fewer graphical toolkits would be better, maybe a modern replacemente for the Athena widgets, shipped along with the X server, could serve as a core for the other toolkits to build upon.
I think it will be very long before significant integration happens in the Linux desktop.
network transparency like KDE?
I don’t have words to describe the joy of using GNOME 2.8 in Ubuntu. The whole system rocks. It has come a long way and now I see what these guys (the GNOME developers) have been saying about GNOME being great. I could taste a little bit of it in Fedora Core 2 test 1, but since then they have improved it a lot. The system is stable and fast (for me and my machine). No crash yet for 3 to 4 weeks and 2 PCs. Of course Linux is comprised of many spread programs and libraries, including the kernel, so many of the improvements are spread among all its parts. Even the integrated Firefox contributes to this lovely desktop. Problem is: this system isn’t newbie friendly yet, and we can’t promise that all your hardware will work with it. If the manufactures don’t support Linux by default, it’s expected that some equipment won’t work at all. Come on guys, support Linux.
Mark McLoughlin from Red Hat is working on menu editing and hopes to get it working and enabled for 2.10, along with a simple menu editor.
The problem with the stuff in Fedora Core 2 and 3 (where the editing code is present but disabled) is that the code is very incomplete and doesn’t work very well. Its based on Gnome VFS (original idea was to edit the menus through Nautilus) but that type of file-based implementation doesn’t work very well and screws up the XML files sometimes, besides being slow.
yeah…but this is typical for open source projects.
how can i get them transparent?
like the button on the left:
http://az-lantech.de/ubuntu/applet.png
the right ones look ugly in grey.
…which is exactly why I won’t use the testing branch currently available on Portage.
I’m not so eager to use gnome 2.8 (in no hurry actually, 2.6 is great, stable, and very good looking) – that I would “switch to Ubuntu” either.
That would never happen, not this cat!
I guess you would have to write a GTK engine without a background, relying on the composite extension of X.org. But even if that was sufficient to get it working, all your applications would be transparent too.
ignorant,
Integration was never a goal. Interoperability is, now. You need to stop viewing desktop Linux through the Windows-OS X spectacle. Desktop Linux is not owned or run by one company. Sit down and imagine how thousands of software applications designed by thousands of people who may or may not have seen, spoken or communicated with each other, is expected to work seamlessly as if they are designed and developed by one unit.
It is somewhat of a miracle that Desktop Linux works at all. However, there is a movement gaining momentum among GUI and desktop technology developers. The goal is to ensure a user’s preferred desktop environment, and its native applications, can share base technologies with other desktop environment, and their applications. freedesktop.org is the body promoting this movement.
No, desktop Linux is not going to be a single desktop run, managed and developed by people who think alike, sit alike, eat alike and share to same vision for the one true, powerful and indispensable desktop environment. As a long human taste and preferences remain diverse and unique, there will be no “one desktop to rule them all” on Linux ever! We have no choice but to accept, embrace and deal with it.
http://freedesktop.org/
Depending on the kernel you use, the type of filesystem you use(i.e reiser4, reiserfs, etc) and your device management system (to use version 2.8 you have to move from devfs to udev), you may or may not encounter problems.
I’m currently using gnome 2.8 with devfs. While it doesn’t take full advantage of gnome 2.8, you don’t have to switch to udev to use it.
“At the pace integration is being pursued in the Linux desktop, maybe in 50 years we will have something. ”
—
I am as sceptical as you are. freedesktop.org just about finalised a trash spec. things move on in a faster pace now. dbus as the system IPC for kde too changes quite a lot of stuff. you are right about kparts. its the next thing to target for cross desktop functionality. btw, openoffice and mozilla can already use use gnome-vfs.
“Mark McLoughlin from Red Hat is working on menu editing and hopes to get it working and enabled for 2.10, along with a simple menu editor. ”
I hope you guys do this. its one of the major FAQ’s on gnome now. its been quite sometime since i have seen redhat guys posting around websites like this. keep that up. talk and blog more
”
I’m currently using gnome 2.8 with devfs. While it doesn’t take full advantage of gnome 2.8, you don’t have to switch to udev to use it.”
—-
right. thats true. however devfs is known to be racy and currently depreciated and its better to switch anyway
right. thats true. however devfs is known to be racy and currently depreciated and its better to switch anyway
I’ll get round to doing it eventually. I don’t have time to babysit linux right now. It’s currently working well enough and I can’t afford to break anything right now. But that’s another rant for another day…
am as sceptical as you are.
—
I meant NOT sceptical
No need to switch immediately but its good to keep this in mind. btw which distro are you using. did it come with devfs enabled. the only distro proactively using devfs was gentoo. since gregkh from ibm uses gentoo too, he has enabled gentoo to use udev now. if you are using gentoo you dont really get to complain about baby sitting 🙂
other distros like mandrake and suse has devs which was optional?
I’m using Arch Linux, and yes, it came with devfs (I installed v0.6). Sorry about the baby sitting comment
I usually like to try out the latest shiny toy, but some days I just want to veg out so to speak. Especially when I’m swamped with work and/or study. But heh, somehow there’s always time to post on the forums :/
“Believe it or not, it actually makes more sense to go to a website, download a package, double-click on it, and the packages installs itself.”
No. No, it doesn’t.
“No need to switch immediately but its good to keep this in mind. btw which distro are you using. did it come with devfs enabled. the only distro proactively using devfs was gentoo.”
Actually, Mandrake used devfs by default from sometime around 8.x (?) through to 10.0. 10.1 switches to udev by default.
I have been an ardent supporter of GNOME for the past 3 years-supporter in the sense that I use and encourage others to use it and have invested hundreds of hours in learning many aspects of how GNOME works and how to work with it. But GNOME-2.8 has proven to be a disappointment. The vast majority of things GNOME-2.8 work and work well-but certain things which are buggy have a severe negative impact of the total impression which it leaves. My personal frustration with it stems from the fact that there are so many things that are done so well and the GNOME-2.8 desktop comes so close to being a trully awesome desktop environment that it is trully painful…-so close and yet so far…
The buggiest part of the GNOME-2.8 is by far the gnome-vfs. It is as if there was zero testing before release. What so strage is that the GNOME-2.6 version of gnome-vfs simply did not have any of these problems-they are all new. The show stopper for me has been issues with Samba, gnome-vfs simply goes off into lala land when confronted with a samba timeout.
I was forced to switch to cifs as opposed to smbfs due to these problems in gnome-vfs. (no other file managers did more that twitch when confronted with timeout errors-which are par for the course when using smaba). gnome-vfs also has some curious notions about deteriming free space in automounted smbfs/cifs shares-without using “unix extensions = yes” in the smb.conf on the server nautilus just assumed that lack of information regarding freespace on the device, which is normal on autofs automounts, meant that there was 0 free space left-with unix extensions and cifs it now sees free space-but how it calculates that is a mystery to me as it bears no relation to what ‘du’ returns or what is actually present.
As is the case for how it determines file permissions of cifs shares. Or how gnome-vfs when confronted with a link to an automount directory of a cifs/smbfs share on the desktop simply chooses to probe the same link(and in so doing the automount) every 5 seconds religiously even though its not in use. As it stands the extremely buggy implementation of gnome-vfs cripples the entire GNOME-2.8 desktop if used in a production server-client setup. I have managed to find workarounds to get most things functional-but it has been a nightmare, to put it mildly….
Minor annoyances….
1) someone changed the MIME-type for dealing with msword documents. The new MIME type system used in GNOME 2.8 decided to use application/ms-word as the MIME type for msword documents when the entire would is using appliation/msword. This kind of oversight is so mind-numbing that it almost hurts. There are hundreds of thousands of word documents available in the internet and by default all GNOME applications which use the MIME type system fail to correctly associate the proper icons and applications to use when dealing with msword documents. The person who made this change did so believing that the ‘alias’ keyword available in the new MIME system functioned correctly-in the xml file ‘application.ms-word’ there is an ‘alias’ set to ‘application/msword’ -yet ‘alias’ is non-functional-ie. it does not work. Hacking this to work correctly cost me 2 hours of time trying to discern how the new MIME type system is implemented for there is zero currently available documentation regarding GNOME-2.8 or the new MIME type system in use.
2) Nautilus seems to be moving in rather curious direction-Nautilus now offers to excute any and/or all files which happen to have the executable bit set. Firstly this is foolish beyond belief-it may be ‘easy’ for those comming from the windows world-but it is contrary to default security provided within the Linux environment. There is no attempt made to ascertain whether the file itself is an executable-merely whether the executable bit is set or not. Which means: files loaded from file systems which do not support the attributes in use in file systems used under Linux-NTFS/FAT/VFAT are incorrectly seen by nautilus as ‘executable’ files-even though the there is no such thing as an executable msword document or pdf file.
3) Nautilus offers 3 options concerning caching and previews of files: one can either AlWAYS use them, only use them for LOCAL files or NEVER use them. It is a standard procedure to locally mount Samba shares when these shares are continuously in use. The LOCAL setting ignores the fact that these locally mounted directories *are* not local directories-thus resulting in an absurd waste of network bandwidth.
4) The application/smil MIME type incorrectly launches OpenOffice when one clicks on a smil link in epiphany. SMIL is *the* MIME type which is used for Real Audio and Video….Of course it is not the responsibility of GNOME developers to account for the missing “MimeType=application/smil;” line in the /usr/share/applications/realplayer.desktop file. But launching OpenOffice when clicking on a smil link is rather absurd. ‘glob’ types also only take into account lowercase extensions….yet routinely when one deals with windows files under Linux documents are saved with uppercase extensions…ie.DOC, etc.
5) By default the new HAL/DBUS system in use shows all entries present in the /etc/fstab file and those discovered by HAL/DBUS in the ‘My computer” browser window. It was bad enough that nautilus/metacity only now, with 2.8, provides the ability to *not* display these mounts on the desktop-but now that one can choose to disable the display of these desktop icons they have snuck into the ‘My Computer’ nautilus browser window. Please note-for single user systems this may be ideal- but for multiuser systems this is a usability nightmare. If one deletes these entries from /etc/fstab they are reinserted by HAL/DBUS. There is no option available to turn the display of these mount points off. There was a patch provided to the gnome-device-manager to turn this off and one of the developers complained that such does not belong in the UI. I had to recompile the GNOME packages to turn of HAL/DBUS support to prevent my users from automounting the second drive in my system to which they should *never* *ever* have access to.
6) the semantics in use for .desktop files have changed but no documentation of these changes is available. .desktop files of the TYPE Link no longer function-sending nautilus off into lala land for several minutes. Again no problem if this has been deprecated..but it would be nice to have *any*(some) documentation regarding such changes.
Don’t worry…I am in the process of writing bug reports …maybe these issues can be cleared up in a timely fashion…
Although I have criticized much here I still believe in GNOME-again it this ‘so close but so far’ which is so agonizing….
As far as the executing issue goes, I can say that’s certainly not true. There’s an option for whether you want to execute or edit shell scripts, yes. (defaults to asking). But for other files set executable, this simply doesn’t happen.
I’ve actually run into this as a problem– Nautilus will not even execute files associated through the kernel binfmt-misc interface.
As for the smil opening in OO.o, that’s most likely OpenOffice’s fault– OpenOffice may well claim that it will open them. The easiest way to change this is to save a smil file locally, choose properties on it, go to “Open With”, and remove OpenOffice from the list. (Or simply select/add a more suitable program)
Is there a place where we can make sugestion to gnome developpers?
Because I would love to see the edition of toolbar becoming like firefox. The actual way of ‘locked’ icon, where you have to unlock each one you want to use and move is quite frustrating. As well as the add applet that create unclear and overload menu list.
If it was like firefox, you would only have to rightclick and selection ‘edition’. than all icon would be unlock, and an box would appeart with all the goodies that are possible for addition.
Does everyone like the ‘centralize’ way of evolution? I would rather prefer a macosx indenpendant apps that work interractively.
I agree with the earlier commentor. If autopackage is all it’s cracked up to be then inclusion of the already developmed (look here: http://autopackage.org/gallery.html) installation tool would be very nice.
I don’t buy the argument against this. It would be useful to have a better installation system, packages are still a pain even with apt-get. I don’t care if my grandma can or can’t figure it out, I would like this feature personally (and I know Linux well enough).
I don’t see this replacing packages per say but it’s a nice option for vendors, especially commercial, to package their own products (unless there are some licensing issues).
I say all this completely ignorant of how well autopackage works, but the idea would be refreshing.
In regards to your response about going to a website to get a program.
Yes it would. Maybe not for a lot of programs but I have yet to see a great solution for commercial vendors. It’s not like they’re going to have you add their site as a repository for installation through synaptic. It’s a pain. Plus those repo’s don’t have a lot of stuff so you still have to go to the developer site and install the rpms, which usually require dependancy’s, and so you’re back in hell.
“OpenOffice.org and Mozilla each have their own incompatible component models…”
So the conclusion is to not use Firefox under Windows XP because it’s not “integrated” or what?
Just stop this debate, alright. I can’t hear it anymore. Please use Windows and be happy with it.
1) many gnome developers read osnews but its the best place to comment on various issues
suggestion ideas
* use bugzilla – file it as a detailed enhancement explaining your use cases
* choose the appropriate mailing list and present it
—
I already read somewhere about firefox like feature
Evolution has already started splitting up. the latest versions have evolution data server and eplugins which will enable it to modularise everything while being integrated on the whole
I hope to see a lot more Project Utopia cases in Gnome 2.10.
Also, in general, Gnome’s implementation of Project Utopia has been hit-or-miss for me. I first built the desktop on Slackware using Garnome, and HAL was poorly configured. None of the published features worked for me.
Then I used Dropline Gnome on slackware. This worked a bit better, CDs were detected as expected but USB Mass Storage devices didn’t seem to work properly.
Finally, i tried Ubuntu. Ubuntu is god. I’ve yet to be disappointed by this new and polished OS.
For commercial stuff, you may have a point. If you really, really must use it. Everything else is better handled through the distro’s packaging system, IMO.
Firefox in Windows *is* integrated, as I see it: I assume XPCOM is built around COM, which is the glue that binds everything in Windows (besides the evil registry). So I assume it is not difficult for Firefox to talk to OOo in Windows. My point is this: Linux misses badly something pervasive like COM. Bonobo and DCOP were attempts to create such a thing, with varying degrees of success. Maybe DBUS will evolve to something like COM, who knows, but even if it does, it seems it will take a long time, so long that it may not matter at all when it does.
I think autopackage inclusion in 2.10 would indeed be very good. Some people here are discussing the use of it by saying things like ‘apt is way better’, or ‘installing apps from a website is the way to go’. IMHO, apt is really good, autopackage is nice, but the combination of both would be excellent.
E.g. I’m using Ubuntu now (lovely btw) but I searched the whole www for a mcatalog deb. Can’t find it anywhere. If the author of this program provided an autopackage, my problem would be solved.
I find it really curious why nobody like ximian is supporting autopackage. Maybe that’s why it’s taking so long to deliver a stable version. If I’m right, a stable version is scheduled for end of this year, so it could be included in gnome 2.10…
DBUS is an IPC mechanism. it cannot act as COM subtitute. bonobo and kparts can. bonobo is currently depreciated (unofficially) and kparts is probably a better option. discussions related to this should go into xdg list in freedesktop.org
It’s my understanding that xdg is not a list for just a component technology, but other things too.
So where are the discussions for a Bonobo replacement?
So where are the discussions for a Bonobo replacement?
—-
it was done in irc. i am not sure its logged somewhere. the main driver behind this effort is Havoc HP AT redhat.com. its a long term plan and until people have a similar replacement the gnome project wouldnt depreciate it officially. it should be noted the gnome didnt drive the bonobo effort either. Miguel did but he later dropped the ball and wanted to move on to the mono stuff.
I wonder why, since the release of the first KDE (and later Gnome), developers are continuously trying to address the same problems?
Integration, duplication of functionalities and *absolute* lack of others, applications that fall short of Windows in terms of integration and speed; and no one addresses these problems, everybody adds “half completed” functionalities, no one “fills the holes”.
Commercial operating systems do the opposite: they offer, *when* they are sold, a set of core functionalities, and developers will later use them; think MacOSX, and all the stunning technology they offered in the desktop field.
Can you immagine Apple releasing MacOSX with a “partial implementation” of Aqua, a still non working version of the dockbar, and trying to mock everything up so it would seem it works, but then it just is pure crap under the surface?
I think we’ve all seen this undreds of times in Linux distros (and still we see in many… usb plug in integration with desktop on the fly anyone?)
In Linux nothing is “released”, but always “worked on”, and when a stable release is released, usually the lacking functionalities are so many people go for the “development” branch anyway.
After years I got fed up with that; I miss it thoroughly, but my laptop now does work wonderfully with XPPro and Mozilla and OO; and, mind you, OO and Mozilla in Linux aren’t but ugly clones of the ones in windows, which are faster and well integrated in the system… why Mozilla is not able, at version 1.8 and after almost 4 (5?) years of opensource, to integrate gracefully at least with the Open/Save dialogues of KDE?
I wonder how many people are like you and just switched back.
You make some interesting points and I think the answers are pretty simple actually.
First is the ad-hoc nature of open source development, with no strong vision. The KDE guys got the technology right, but have had interface issues as well as issues surrounding QT. I believe the Gnome people are on the right track with the interface, but Bonobo has been a failure and thus a nice component framework has been missing.
You’ve got a whole myth behind the supposed superiority of open source development. It’s just that – a myth. The only reason that linux desktops are in ok shape now is because of brute force momentum and not intelligent design. And also because there’s been corporate backing. Things would be a lot worse if it was all volunteer working part-time.
Bottom line is that meritocracy sounds all warm and fuzzy, but it doesn’t beat strong, intelligent leadership with clear vision. That is why the benevolant dictatorship of the kernel has been so successful.
“After years…”
Yeah, right….
“Commercial operating systems do the opposite: they offer, *when* they are sold, a set of core functionalities, and developers will later use them; think MacOSX, and all the stunning technology they offered in the desktop field.”
OS X – yup, that’s a good example of *good* closed-source design (though, of course, building on a free base).
Now think of all the bad ones! Are you honestly trying to tell me that no commercial developer *ever* shipped a release with claimed functionality not present or not working? That commercial software development is a bunny-filled utopia of joyous design with not a single bean-counter pushing for a release on March 16th because that’s what they told the press and, besides, they need a bump in their figures for the next quarter?
As Dexter Holland once so rightly said: “Man, I ain’t buying.”
libnotify/notification-daemon have been proposed:
http://galago.sourceforge.net/specs/notification/
It uses DBUS to communicate.