“Project GoneME is the first attempt to try moving the GNOME Desktop into a new direction. As a long years contributor and ex-GNOME Foundation member I got quite unhappy with the new direction that some core decision takers have chosen without further feedback with the community or their participants” he claims while in search for similar-thinking devs to join him. More here.
Ali Akcaagac Attempts to Fork Gnome
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Sounds good to me. I just hope the integration is there between the traditional spatial and the tree view.
Glad to hear I’m not the only person who wants this feature.
Pekwm, firefox, thunderbird, gkrellm, vim/gvim, streamtuner and Beep most frequently used apps. None of these are KDE/GNOME apps, so you mean this means a lot of redundancy? Where?
Well, let’s see. On your desktop, you’ve got 3 different toolkits (GTK 1.x, GTK 2.x, and XUL). You’ve got at least two networking libraries (Beep must roll it’s own abstraction on top of sockets, and so must Firefox and Thunderbird). You’ve got at least three four text-editor widgets (Mozilla one, GTK+ 1.x one, GTK+ 2.x one, and gvim’s). Neither streamtuner nor Beep are native gstreamer apps, so they each have their own sound layers as well.
Now, compare this to a comparable KDE setup, which would consist of Kwin, Konqueror, KMail, Kate, and Noatun. In total, you’re talking about one toolkit (Qt3), one networking library (KDE’s), one text-editor widget plug-in (choose from qt’s, kate’s, or vim’s), and one audio layer (aRts). The advantage increases when you run additional apps. Kopete uses the same network library as KMail or Konqueror. KHelp uses the same HTML widget as Konqueror and KMail. Everyone uses the same remote filesystem library (KIO). As you keep piling on the applications (what about office program, CD burner, etc?) the marginal cost for the desktop environments is much lower than for the stand-alone application environments.
good luck buddy, but you won’t have it.
just because you don’t like button order [aka doesn’t seem windows XP and you also don’t like this and this] you do a fork.
what’s wrong in switching to KDE?
afterall it’s GPL and it more Windows-like.
So just use KDE and not your fork. What is more interesting is that people talk about combining the two DEs and you just make another one.
Why don’t you join XPde: http://www.xpde.com/
enough with “oh they ate my meatball” forks
Oh yeah, now that you are angry at the Foundation, you’re going to fork Gnome. That will get them! You go girl!
OK, let’s try to think about the facts here:
Changing the Buttonorder back to normal again.
OK, good luck in this one, try to convince all developers which have GNOME projects to do this. Because you’re certainly insane if you intend to do it yourself. In the end you’ll have the biggest havoc someone will ever witness because only a small part of the GNOME based apps will have the button order _you_ like.
Spatial Nautilus.
OK, so you don’t like spatial. Why don’t you start your own GNOME distro (not a fork) which starts with navigational mode by default? Oh wait, that would just sound silly, right?
GConf did upset plenty of people in the past.
You surely don’t sound like a programmer, the GConf API saves alot of dumb work, manipulating configuration files for every project you do is plain dumb. Every advanced platform has and should have it. Look at java, one of the cool things about java 1.4 was their new properties model, which is an hierarquical one. Unlike INI files. The “windows registry” syndrome which ppl babble against GConf claim is purely FUD. Win registry has two problems: a) it’s binary based b) you store hardware details there. Other then that it is a perfectly good system for saving preferences. The win registry pollution is the same as you have in your home, where you have 1000 .dot files.
Control Center is also a thing to worry about. Control Center is also a thing to worry about. How often have you been upset that you need to launch one capplet after another only to change preferences.
Never, don’t even know what you’re talking about…
Removing a lot of bloat is also a point to discuss.
Could you be more vague please?
Worth of conversation is also getting a KHTML port so we can finally have a native Web browser rather than using Mozilla or Firefox.
Khtml is far behind gecko engine, nevertheless I wish you good luck.
Removing unnecessary things such as esound.
I vote for that!
The GNOME HIG is partially adoptable here but there are parts that need to be reworked.
Agreed.
In the end this has so much FUD and so little interesting features that it isn’t funny. Especially not worth creating a fork for.
Pehaps I was unclear, or maybe you got me wrong..
1.”On your desktop, you’ve got 3 different toolkits (GTK 1.x, GTK 2.x, and XUL).”
Actullay that’s two. Gtk2 and XUL. I’m sure you know better than me, but firefox is built with –enable-toolkit-gtk2 (not 100% sure what it does though). Gtk 1.x is not installed though, that’s for sure
2. “Neither streamtuner nor Beep are native gstreamer apps, so they each have their own sound layers as well.”
Not sure what you mean here. A) I’m not a Gnome-user, therefore I do not use gstreamer. B) Streamtuner does not play any audio, AFAIK, it just redirects the stream to beep or streamripper (Console app).
3. “Now, compare this to a comparable KDE setup,”
Woha, hold your horses…
I never argued that KDE apps were bigger than my equivalents in them self, and obviously the advantage of shared code will aggregate with the number of apps that are loaded. However, I doubt that the benefit will make up for the bulkiness kde brings to begin with, with such a small number of applications. I will have a check on this though, as I’m about run KDE for some testing anyway.
4. “choose from qt’s, kate’s, or vim’s”
That’s interesting. Do you mean that the text-editor widget plug-ins of qt and kate are optional?
5. “what about office program, CD burner, etc?”
I’m a mental
Vim/LaTeX and sc covers my needs for the most of the time. CD-burning is made from CLI.. or xcdroast.
6. “the marginal cost for the desktop environments is much lower than for the stand-alone application environments.”
As I said before, that may very well be true. But I think that is highly dependant on how many different apps you have loaded. But we’ll see. I’ll give KDE another test-run.
No it doesn’t, because if you don’t strictly follow one metaphor, the interface gets unpredictable if you don’t know it by heart.
That’s the problem, I know. That’s why I said I hope that Dave Camp, or someone else, could, perhaps, achieve a clean, working implementation of tree view in the spatial mode Nautilus. I don’t know how well it would be possible though? If not, well, then I choose the navigation model – if I have to choose between the two models and there is no possible combination.
I’ve been very open-minded and used the spatial Nautilus for some time now. I know by experience that the navigational model is easier and faster for me.
However, let me restate this once again: even now the so-called spatial Nautilus is maybe only 2/3 spatial. It does have some navigational features even now too. Why? Simply because people do seem to need those navigational features too.
Usability is not the same as spatial… Thinking like that would be blind dogmatism.
Maybe sometimes usability could even be a combination of the best features of different models? Impossible?
Would work by opening a tree in a seperate window than the normal spatial window, or by having two views you can switch between.
When you want to quickly find a directory, view the tree, when you want to work with things, view the normal spatial mode.
Also a spatial tree could be simpler than the normal system wide tree, because you can have a floating top level.
Actullay that’s two. Gtk2 and XUL. I’m sure you know better than me, but firefox is built with –enable-toolkit-gtk2 (not 100% sure what it does though). Gtk 1.x is not installed though, that’s for sure
Nevermind. I forgot they had a GTK 2.x version of gvim. Still, that’s two space-hogging toolkits compared to one. Firefox with –enable-toolkit-gtk2 causes XUL to use GDK for rendering, and Pango for fonts. I doesn’t actually cause it to use GTK+ widgets.
Not sure what you mean here. A) I’m not a Gnome-user, therefore I do not use gstreamer. B) Streamtuner does not play any audio, AFAIK, it just redirects the stream to beep or streamripper (Console app).
My point is that if you use more than one application for audio, you’ll get redundancy unless you have a common audio library. That means the minute you fire up a jukebox program, you’ve got two audio libraries loaded.
I never argued that KDE apps were bigger than my equivalents in them self, and obviously the advantage of shared code will aggregate with the number of apps that are loaded.
The whole point of an application platform is to mitigate the costs of having a wide-variety of apps. If you’ve got an extremely-limited selection of apps that you use, then you’ll probably end up ahead with the fluxbox approach. However, if you use lot’s of different apps, it’s better to have them share a common set of platform libraries, even if that causes the base system to get bigger. Thus, you can’t really say it’s bloated, because it includes the functionality applications would have to reinvent anyway. On my SuSE system, I’ve got a ton of KDE applications loaded (Konqueror, KDevelop, Quanta, KOffice, Kontact, Kopete, K3b lots of utilities, some graphics apps, etc), and those apps, along with the base KDE system and Qt, takes up about 250MB of binaries and libraries. I doubt you could get a comparable set of programs in Fluxbox in less than that amount of space.
I hope he manages to gain a large enough momentum to change some things and bring some respect back to the Gnome project from its _real_ users, and not some imaginary, theoretical Joe User who doesnt care in using Macs let alone Linux and Gnome, but whom the project still tries to please in the first place, blinded by world (desktop) domination delusions. and leaving its current and loyal users behind who helped to make it what its today.
Everybody saw what happened to Xfree86, and how quickly literally _everybody_ left it back in the dust as a usable alternative emerged. My guess is that this will happen in a similar way when the Gnome Project keeps to ignore its community and its wishes as it does now, because, when oGALAXYo’s idea catches on, there will be an alternative. XFree86 did in the end not succeed with its authoritative and imposing development model, and arrogance towards its community, and failed, and so should the current Gnome.
Go oGALAXYo go!
” But the “KDE and GNOME are bloated” claim is not something that makes a great deal of sense.”
IIRC, stock KDE and GNOME are ~100 MB, Dropline GNOME is ~200 MB
or not?
I guess everybody needs a patron somewhere along the way.
Project keeps to ignore its community and its wishes as it does now, because, when oGALAXYo’s idea catches on, there will be an alternative. XFree86 did in the end not succeed with its authoritative and imposing development model, and arrogance towards its community, and failed, and so should the current Gnome.
Oh come on. Gnome has _not_ been arrogant toward the community. The Foundation has been under direct assault by people like oGalaxyo for some time regarding these issues, and the Foundation handles these people with dignity. oGalaxy on the other hand never handles any dispute with dignity.
If you read carefully (which I doubt you did) you would figure out that there is nothing wrong with GNOME as project, idea and philosophy. I was talking about certain people who sit in upper positions who cause an harrassment on others by namecalling them, slandering them etc. What you have read from me for years (thanks for confirming that I am at least part of GNOME for years, even the way you did it isn’t the way I would have seen it) then you may understand that these problems are there for years as well. As long nothing changes nothing much will improve.
I am talking here about people, developers, contributors, users who got permanently taken to the shambles, where people are being talked in the back and stuff like this. This all has nothing to do with wanting to fork GNOME (even if I had the idea in doing so a couple of times).
People mistake features such as rounded windows, drop down shadows, transparency, and the most preetiest icon’s ever to be user friendly. In the end these are just cheap tricks/gizmo’s to convince Linux/KDE/Gnome converts from Windows into thinking we really have something here.
The Linux desktop is not half as mature as Windows XP graphical interfaces in design and implementation.
Throughout Gnome and KDE there are too many options that end users and converts from Windows XP in this generation could not hope to understand, nor should they.
“Throughout Gnome and KDE there are too many options that end users and converts from Windows XP in this generation could not hope to understand, nor should they.
”
stick with windows. saying that gnome has too many options makes your whole comment worthless. Linux people develop UI for them to use. if you want to stick with windows nobody is threatening you to convert. you dont have anything to say in here
Miguel de Icaza said:
“I do not think anyone in Gnome objects over someone porting
KHTML over to Gtk. Ali or anyone else is free to port it,
and if folks adopt that technology then great.”
You got the developers, why doesn’t Gnome do it? Because…’if it ain’t invented here, it’s no good’, that’s why.
“I do not think you will see anyone in Gnome asking for more
bloat, I think the opposite is the case. It is simpler to
complain than to do something about it.”
So why then, every new version is as slow or slower than the preceding one, why the persons in charge of development LISTEN for a change?
GNOME is not for DEVELOPERS, it’s for USERS.
GNOME itself wants to adopt all types of technology which somehow scares me e.g. a lot of talks about Storage, a lot of talks about MONO and Python and so on
Sounds pretty conservative to me. Think ahead to the time when Longhorn is released. Advanced search capabilities, extreme integration, all kinds of new and exciting features, and a developer’s utopia. From the sound of it, you’re proposing to strip down Gnome to the bare minimum…in fact, you’re proposing to go _backwards_. What kind of a position will it be in to compete with Longhorn when the time comes if Gnome follows your philosophy?
I’m sorry but I would agree with other posters that you come off sounding bitter. I disagree with virtually every one of your points (especially about spatial nautilus…have you used it extensively at all? I find myself pissed off when I have to use navigational now) and I really think that if Gnome were to follow your philosophy it would be 0 steps forward and 10 steps back. We’ve had GoneME already…it was Gnome 2.2. Now we’re in 2004 and we’re using Gnome 2.6, which I find _much_ more usable.
I’d like to wish you luck in your project, but I really think developer’s talents are better used elsewhere like helping Gnome in adapting to changing technologies and preparing for the future, not looking back on the past. Like it or not, progress will occur and whether you agree or not with the masses, this is the direction the project has taken. All you can do is voice your opinion. Forking, in my opinion, is just petty.
Metic wrote:
Sounds good!
I hope he finds a good way to implement it that keeps the interface clear and usable.
The idea is to have a listview, which acts like a tree. So you have those arrows to expand a folder right inside of the window. The root of the tree would be the location of the folder and if you doubleclicked some folder, it would open in a new folder window. So it does not break the spatial metaphor at all.
As for the rest of your statement, Julian already summed it up very well. Combining features of both metaphors will only make both less predictable, which is really bad for usability. So the GNOME idea isn’t to offer only spatial folders or only browsers or only a combination of both, but to offer both and make it easy to use whatever works best for a certain situation. Maybe it’s not easy enough yet, but of course there can always be improvements to this in the future. After all, GNOME 2.6 was the first version ever which went this path and I believe that they already did quite an impressive job on it.
Currently you can activate the “Browse Filesystem” item to get a new browser instance whenever you need one (move it to your panel for faster access) and if you want to open a specific folder in the browser, you can chose “Browse Folder” from the context menu. It’s not possible right now to open a spatial folder window from within a browser, so maybe that would be one possible improvement to discuss about.
I don’t like the idea of a “mode toggle” as in OSX, because that would make it even harder to understand that spatial folders and browsers are supposed to be differently acting objects, not just two modes of the same kind of object. However, what I could imagine would be a way to quickly open a folder in the browser while closing the current folder and the other way around. This would act almost like a mode toggle, but keep the separation clear.
Where did you read that? I have read the maillist thread that was posted a while ago and even Havoc was not against the idea. Miguel might be the leader of the priject but he is far from representing the mind of every GNOME developer. The obsession with Mono is one proof: many GNOME developers simply don’t want to touch it with a 10-foot pole.
–enable-toolkit-gtk2 causes XUL to use GDK for rendering
Thank you for the explanation.
That means the minute you fire up a jukebox program, you’ve got two audio libraries loaded.
Ok. I see, point taken.
really say it’s bloated, because it includes the functionality applications would have to reinvent anyway.
Not to be inflammatory, but isn’t that a bit of sophistry?
If I have to load a lot of libs containing code that I have no use for because I don’t use any programs that use it, isn’t that bloat to me? One man’s gain, another man’s loss, I guess.
Point is, if we leave KDE for a moment, that I can use a low-end box pretty much without penalty (ok, a little patience is needed) with the described setup. A couple of months ago I tried to use gnome on it (300Mhz 128MB) and it was absolutely useless, slooow is just the beginning of the description. I never tried KDE on that box, but I doubt it would have fared that much better.
As for the number of toolkits: I don’t know how much better Konqueror has become, but at least in the past konqi was less than perfect, so for some sites you HAD to keep mozilla/phoenix/firebird/firefox around anyway. How’s that for bloat? Gtk in some form + xul + qt. Yummy. So in my experience the theoretical advantage of sharing a single toolkit between the apps, still is quite — theoretical. This will certainly get better as time passes on, but for now things are as they are.
To return to you original question What apps do you WindowMaker/Fluxbox folks use anyway?
You got answears from one at least, and it turned out that the situation wasn’t quite as bad as you wanted to make it look like. As you noted before, in the end it all boils down to what apps you really use. So you got all those programs loaded, but I think you make the same mistake as a lot of people propagating for those fancy DE’s. You assume that everyone has a computer that is as fast as yours, and that they want the same apps as you do.
I can use vim for tasks that you use three different programs for. So when you say I doubt you could get a comparable set of programs in Fluxbox in less.. the question is “comparabel in what way?” In functionality or in functionality *and* apperance? I.e Vim and aspell is not as sexy as openoffice, and not as easy for the recent windows refugee, but they can pretty much do the same job. And that with significantly lower system requirements. Furthermore, as my needs quite probably are not the same as yours, why would I want the same setup as you?
One size does not fit all. It’s as simple as that.
I see lot of guys posting here running Ali down because he’s forking GNOME and not contributing to it (sticking to their stringent although lop-sided IMHO, rules).
And to think GNOME is the flag-bearer of everything open and free and the power to do things “differently”. Otherwise, I for one, would have been glad to see all the existing GNOME developers join KDE band rather than start over with a new desktop of their own.
If I have to load a lot of libs containing code that I have no use for because I don’t use any programs that use it, isn’t that bloat to me? One man’s gain, another man’s loss, I guess.
I suppose it depends on your definition of bloat. I think most people consider something bloated when it is a lot larger than it needs to be. When the extra size is taken up by features that are needed in for a product’s target market, and the code is not such that these features take up much more space than expected, then you can’t say that’s bloat. You’re welcome to criticize the software for not being optimal in situations where the user has a much smaller set of requirements than the general population, but that’s something very different than saying that the software is bloated. Look at it this way: is GCC bloated because you don’t need it’s full C++ compliance and powerful optimizer to generate good code for your PDA? Or is it just overkill for your requirements?
You got answers from one at least, and it turned out that the situation wasn’t quite as bad as you wanted to make it look like.
It’d be nice if you didn’t take statements out of context:
“What apps do you WindowMaker/Fluxbox folks use anyway? Unless they are KDE or GNOME apps, there is a very high probability that there is tons of redundant code in your system, because each app has to reinvent a lot of common functionality, instead of sharing it.”
I never set out to prove WindowMaker/Fluxbox was bloated. All I had to prove was that full environments using these WMs have a high probability of having redundant code. Even if the few applications you listed was all you used, there was still quite a bit of redundancy in there. In a more common desktop, where you might be running more things (address book, email client, etc), that probability goes up.
You assume that everyone has a computer that is as fast as yours, and that they want the same apps as you do.
No I don’t. I never said that WindowMaker/Fluxbox/etc had no place. I said that proponents of those programs couldn’t really levy the “bloated” claim against KDE and GNOME. KDE and GNOME have to serve a much wider variety of users than do WindowMaker/Fluxbox. As a result, they have to provide more general functionality. This doesn’t make them bloated, just targetted at a different market.
So when you say I doubt you could get a comparable set of programs in Fluxbox in less.. the question is “comparabel in what way?” In functionality or in functionality *and* apperance?
Just functionality, if you consider that features designed to serve less knowledgable users to also be functional elements.
why would I want the same setup as you?
I never said you’d want the same setup as me. I said that you cannot justifiably criticize GNOME and KDE of being bloated because serve users with a wider range of needs than yourself.
One size does not fit all. It’s as simple as that.
It was you who implied that one-size fit all, by claiming that KDE and GNOME were bloated. Systems that need to be general are not bloated. There may be solutions that are more optimal for a specific niche of uses, but bloated doesn’t mean that. It means that something is too large/too complex for what it does.
Very good points. I agree 100% with what you said.
If we look at what happened to Xfree86, and how fast it happened.
We could possibly see Gnome dead by the end of the year
Replaced by GoneME. Provided GoneME can deliver some results.
And I’d realy love to see that.
There is nothing worse than developers (Gnome devs) that completely ignore users and force their own ideas how users should use their desktops.
I couldn’t word it better than you.
“imaginary, theoretical Joe User”, very nicely put.
The fact is that Gnome project took dirrection as if Gnome was suddenly going to have 100% market share on every hardware and OS platform, and every grandma and grandpa are going to use it.
Meanwhile they (Gnome) left all the long time Linux users with a dumbified desktop, Windows like registry configuration, the most pathetic file manager ever seen in modern day software(Nautilus) and file-browse dialog that made me lough when I saw it in Gnome 2.6 (those buttons that show your location, what the hell is that?) and that’s sub Windows 3.1 level feature-wise.
The fact remains that no Joe User is using Gnome (or Linux for that matter) right now and will not be in forseeable future. Gnome needs to deliver features and configuration options for it’s current user base which is long time Linux power users.
So you’re forking Gnome and KDE together? I don’t get it totally. Can you clarify?
I also wanted to say good luck to Ali and I wish good luck to the GoneME project.
It’s about time someone brought some sense into Gnome and made it usable again.
I agree (after reading Ali’s GoneME website) with every point he raised.
Gnome needs those changes so it appeals to power users again.
It’s the only hope for its survival.
Grandmas can go and buy Macs if they want something easy and dumb
Don’t dumbify my Linux!
Long live GoneME!
I like the direction of Gnome as a user but like with the GUI the whole has to be kept simple as Rainer and oGalaxyo has pointed out with libraries and sound deamonds under Gnome. Keep the base simple and uniform so that programs developed for the Gnome desktop utilise the same resources and can interoperate easily with each other.
It also will be nice to have Epiphany without the need for all of Mozilla to be installed but for me it is fairly irrelevent cause I use Opera and can uninstall Mozilla/Epiphany.
Also I would like more centralised preference settings alla Windows style of control panel or other for hardware/software settings. At the moment Gconf is a little vague in it’s structure. Still it could be tuned to work nicely and it would be great if you could control X (and other) settings from within it and also have Gconf give you the name/location of the text/xml file it’s changing with the setting selections being manipulated (would help new comers understand the configuration structure of Linux if they so chose).
It’s getting there and speed of Gnome is getting better and for folder navigation a few folders deep Spatial navigation has great benfits over heirachial. If I want to dig deep into the systems guts which isn’t very often, then I use the Navigational Nautilus version with a really hard to use (NOT) right click of my mouse and selecting it from the menu.
Let’s keep in mind the vision of Gnome before slamming it. To make Linux desktop use easy for the average computer user. With that, I think Gnome is succeeding where no-one else has been able on Linux.
If your wondering about how to use Linux as a power user, try learning CLI. You still have the terminal interface to CLI within Gnome if you want and it would seem to be the easiest way to prove your worth as a Linux Power user. Geez, even Windows has a terminal CLI interface and I bet many so called Windows “Power Users” have no idea how to use it or CLI in a rescue situation to fix Windows. The tech support people don’t they always suggest re-installing the OS as the fix for all times. Bah Humbug to them all (-:
Can’t you just move to KDE and leave GNOME alone? I’m sure the KDE project would appreciate someone of your AMAZING SKILLS much more than those supposed ingrates at the GNOME project.
It’s obvious with about half a minute of reading the Galaxy has no idea whatsoever about half the issues, and has no conception whatsoever about future extensibility or maintainability. The only good thing I can say is that he’s obviously never going to be able to code this, so we don’t really have to worry about the fork.
Yeah, sure, let’s see him rewrite basically all the underlying systems. Give me a break. For damned sure anything he puts out is going to be a massive GUI mess, given how little understanding he has of the issues (his reasons come down to “looks like X”, “I don’t like Y”). His bloat rant is especially laughable.
The _constructive_ thing to do would be to work within the framework of the original GNOME project. You can optimize to your heart’s content, port KHTML, and write options for what system-level dirs you want to see. But, of course, when you have a huge ego, the only way is your way.
-Erwos
After reading the article and the various background I came to look at the comments hoping to figure out if ali had valid points or not. Unfortunatly, I found a stream of short posts doing little more than attacking the individual making the comments or blatantly asserting he was wrong. Mixed in seemed to be a few individuals like me curious about his points but not really knowing enough to comment definitively combined with some dogmatic assertions about UI. Perhaps the issue is that these arguments have already been thoroughly refuted in other forums and people are tired of making detailed rebutals. I can certainly appreciate this position but if you are too lazy to do anything besides assert the poster is wrong why post at all? At least you could include a link to where this issue has been addressed in detail.
I was particularly bothered by the host of posts who dismissed the issue by attacking ali. He is accused of being a spamer or otherwise annoying but not much evidence is cited besides posting to both osnews and slashdot (ohh the horror two online communities might discuss the same topic). My favorite was the post attacking him by saing “He’s known to cause seriously heated debates in the Gnome and KDE communities.” Of course we could never trust someone whose ideas were controversial. Even if his online behavior was completly disgusting this particular declaration is polite and dispasionate and deserves to be addressed on the merits. In the 21st century we should have realized that the truth of an idea should be evaluated independently of the person who originates it.
Many of the UI debates are just silly. Quite clearly the same UI is not equally good for every person and posts about how you can and can’t use spatial browsing well simply aren’t going to change this fact. Furthermore statements like users ‘shouldn’t’ be browsing to /etc are just silly. Are people who choose to do alot of browsing/reading of configurations (like I do) while in user mode violating the ten commandments of computer use? Perhaps there is some case to be made that computer users who prefer other ways of organizing their desktop should be coerced to the spatial way but some argument for this needs to be given. There is certainly the preception among many gnome users that they are being coerced into one UI style (just because they can change some config options isn’t a convincing objection…the very issue at hand here is ease of use) and it seems some justification of this policy or evidence that it is not the case is warranted.
Also while the statements about what most or newbie computer users prefer may be correct where is it written in stone that this is the goal of UI design. Consider the dewey decimal system’ it is certainly less approachable than the more simple system of arranging books alphabetically by title but because it locates similar subjects next to each other it is clearly a superior system. The very notion that it might be good to coerce users to a new metaphor (by perhaps making it the default) argues that it might sometime be in the interests of ultimate productivity to convince simple users to learn a more powerful interface. For instance mixing spatial and navigational aspects (maybe one spatial and one navigator browser…distingushed by window color or something….maybe this already exists) might give users more power even if it sometimes presented more options (and thus more confusion) when they first began using. It almost seems as if GNOME has the unstated goal, not of serving the interests of the current user group or at least not the developer ones in online communities, and instead creating a UI to let linux steal users from microsoft. It would be foolish to deny that a great deal of linux’s success is due to it’s nuts and bolts feal and attractiveness to power-users so we should probably give proposals, like Ali’s, which seek to turn a system into a more hacker friendly/bare bones system serious thought.
I simply don’t find the claims made that mixing metaphors is necesserily bad very convincing (although unlike many of the posts here these were at least good points). For one it seems that humans can get used to quite bizarre and ad-hoc schemes/metaphors given time so at most we are debating about how easy it is to learn. As I pointed out above it is hardly clear that this really is what we ought to maximize. Furthermore, people employ mixed metaphors all the time with great ease. Consider a filing cabinet at a buisness that groups employment records by year in differnt cabinets and employee name in the cabinet. Also (with some fear of being smitten for blasphemy) I am given to question the real utility of interaction metaphors. Should we really take the significant evidence that many people don’t understand the hierarchiecal metaphor of files and folders as proving that this simple concept is just beyond their grasp and they are in need of the even simpler spatial metaphor? Perhaps it just means alot people just don’t understand the machine via these metaphors. What evidence supports the metaphor as being the most significant facet in UI rather than other things like number of options, appropriate names/help (i.e. why couldn’t the anecdotes about people finding the spatial systems easier to use be simply because Music, Text, and Video are fewer options than are given to them in a navigational setting and the metaphor is irrelevant).
So in my prior post I complained that I hadn’t really seen the issues addressed. I was afraid it would go over length so here are some specific questions Ali’s proposal brought to mind that I would enjoy seeing discussed.
-Should we be coercing people to a new UI choice by making spatial nautilus default (if they weren’t coercing they would just have an option pop up (with a suggested answer) when each user starts gnome for the first time). Similarly with the capplet question.
-In the same vein what user base should we really be targeting. Do we want a system that is best for the average computer user, the average linux user, a power user or some sort of comprimise? Is the project specifically targeted to gain mainstream market share (in which case we should be copying windows metaphors right or wrong) or to be some sort of theoretically perfect UI (in which case the argument for simplicity vs. power becomes much more unclear)
-Some people didn’t seem to understand why someone might object to the idea of a centralized configuration system. I for one would be concerned that it would have similar problems to the registry (irregardless of the method of access) such as entires not being cleaned up. Also the very existance of a backend configuration database encourages the idea that it will be accessed primarily by programs and not by hand which ultimately encourages programs to create a ton of uninteligeable options even if nothing about the new system ‘theoretically’ makes it less accesible by hand. Finally the only reason to have a centralized config system in the first place is to allow programs to share information via the system and I worry that this might introduce more instability into the system. Using .config files makes a security slip up where you modify a key someone else relies on less likely. I’m not convinced (nor is he) that overall this is a bad thing but there are clearly issues to be discussed (policy…when should you use gconf as well as code).
-Finally his point about code bloat is once again a valid question of direction. Do we want to get the most functionality as quickly as possible or concentrait on building a stable elegant base. I’m not alledging there is a stability problem in gnome (I just don’t know haven’t used it lately enough) but one could concentrate on revising the basic code/standards until they were particularly well developed (i.e. rewriting the old base) or adding new features. Also should development concentrate on making packages dependent on the minimum installed base reasonable or instead push the entire gnome package together (potentially propagating instability) encoraging fastest development over slimness.
Ultimately, like any reasonable call for a code split (aside from incompetance of the development team) the issues he raises are primarily ones of direction and ultimate goals. Ali wants to focus on visual and code elegance targed at the power user while the gnome team seems more interested in UI consistancy (i.e. consistancy of the metaphors not the chrome) values fast development over systemic elegance (multiple interfaces to do the same thing….this is not necessarily a bad thing…sometimes the inelegant solution is the most practical) and appeal to some notion of the ‘average computer user’ (although I can’t really explain why. If they were really interested in gaining market share in the US/developed world wouldn’t it better serve them to copy microsoft UI deciscions? Perhaps it is better to target the developing world but it is a valid question).
“Or is it just overkill for your requirements?”
Perhaps the term “bloat” gets misused, and the correct term is overkill. However, when you look at such an application as gnome-terminal an see it eat someting like 20MB ram, when you are used to aterm using < 1, bloat is the automatical associasion.
“It’d be nice if you didn’t take statements out of context:”
If I did, I’m sorry — things tend to get messy when you try to think, when you really should sleep..
“I never set out to prove WindowMaker/Fluxbox was bloated.”
Well, you did claim that there would be large redundancies, wich could be translated as bloat I guess.
“…quite a bit of redundancy in there.”
The rest of the apps are console apps, and thus not very interesting. But what is interesting is “is it possible to stick with one toolkit?” As I pointed out, at least in the past it was not doable to stick with qt only, so you ended up with all three toolkits anyway. Gnome was better in that way, but it has other unpleasant drawbacks.
“In a more common desktop,”
This we have already agreed on.
“I said that proponents of those programs couldn’t really levy the “bloated” claim against KDE and GNOME.”
This we have already been through. See the first paragraph.
“Just functionality, if you consider that features designed to serve less knowledgable users”
Different opinion again. Those features are like supporting wheels on a five-year-olds bike. They are not part of the bike and they are not necessary, but they make the process a bit less painfull. But it’s quite obvious that the bigger apps you use the more they will gain by shareing code with other apps. KOffice compared to OpenOffice is a nice example.
“you cannot justifiably criticize GNOME and KDE of being bloated”
And with this we are back to the beginning again. It all depends on how you define “bloat”. Unneeded stuff is bloat in my mind. Maybe I got the definiton wrong, but a terminal emulator that needs 20M ram does seem to fit the description. Sure it does more stuff than a simple aterm, but 20x ..? And before somone jumps me.. I know that gnome-term doesn’t == gnome. But it is a very visible part of it.
“It was you who implied that one-size fit all”
I certainly did not. Where did I write that you’d have to surrender you SUSE box within 2 weeks or face the consequences?
“that KDE and GNOME were bloated.”
I’d say what bloat is, is as much a matter of opinion as it is a technical one. I never said that the ideas behind KDE and Gnome were bad or unhealty, but the implementation isn’t always all that great. Yeah, I know, “work in progress”, but this is about what we have now, and not in two years.
“Systems that need to be general are not bloated.”
Maybe. But to reverse a bit. What does Gnome offer that windows 2000 does not? My old “Senilix” has been running w2k quite happily, and yet gnome made it *crawl*.
Does that not indicate bloat?
Unneeded stuff is bloat in my mind.
The key question is what do you mean by unneeded, or more appropriately, unneeded according to whom? If all you need is MS Paint, then Photoshop is bloated. But for a graphics pro, all that bloat is absolutely necessary features. So it’d be stupid to say that Photoshop was bloated.
There are not a whole lot of superfluous features in KDE/GNOME. Component parts are critical in allowing different applications to share functionality (eg: HTML widget). Network-transparent filesystems (KIO, GNOME-VFS), are needed features. Networking libraries are needed features. KDE, in particular, has no marketing department to satisfy. The features in KDE’s core libraries are there because developers want them, which means that their applications use them. Just because *you* don’t need them, doesn’t mean that everyone else doesn’t need them. Ergo, you can’t say that these are unneeded features.
gnome project need
1) stop stupid binding (sorry guys ) perl ruby php python java
2) implement mono ASP.NET ADO.NET yo yo
fuck off all patent problems with debian i install
all features on the fly without buy any distros.
3) more power of choice to user
4) gconf out or at least rewrite most of all parts.
at this time it’s suck.
5) winfs copy,we need a system like winfs for file organization.
6) features features features
7) there are many good gnome developers but they lack of
leadership,too many zealots on the mailing lists.
1) stop stupid binding (sorry guys ) perl ruby php python java …
Why binding with different programming languages is a bad thing?