Weekly Edition
Return to the Press page
| |||||||||||||
|
| |||||||||||||
Ximian XD2: The Way Things Ought to Be (OfB.biz)
Open for Business
reviews Ximian XD2. "Ximian Desktop is fairly
non-invasive. While it offers to replace the distribution's login manager
with its own variant of GNOME's display manager, it does not force it on
you, should you wish to continue using another one. Once I logged into my
account on the test box, Ximian offered the choice to preserve my old GNOME
settings or replace them with Ximian defaults. My KDE desktop files were
also found and placed in a folder on the new Ximian desktop - a nice added
touch."
(Log in to post comments)
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 7, 2003 17:28 UTC (Sun) by danielpf (subscriber, #4723) [Link] This article touches on something very fundamental about the way people see "Things Ought to Be". This author praises uniformity à la Windows as an ultimate goal, a credo close to a religion. The most surprise free the system, the better, each change or irregularity is viewed as a disturbance, a potential menace. This reminds conservatism in other circumstances.But uniformity becomes boring on usage. We know how long time Windows users discovering Linux find the experience "refreshing". Geeks and creative people precisely hate Windows because of its uniforminity, its monoculture and unique thinking. In conclusion the spectrum of people is broad enough for needing both fully "integrated" and multi-cultural computers.
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 7, 2003 18:53 UTC (Sun) by tjc (guest, #137) [Link] This author praises uniformity à la Windows as an ultimate goal, a credo close to a religion. The most surprise free the system, the better, each change or irregularity is viewed as a disturbance, a potential menace.There's a lot to be said in favor of expected bahavior in user interface design. http://www.joelonsoftware.com/uibook/chapters/fog0000000057.html
Ximian XD2: The Way Things Ought to Be - Uniformity? Posted Dec 8, 2003 10:13 UTC (Mon) by flewellyn (subscriber, #5047) [Link] Are we talking uniformity of appearance, or of behavior? In terms of uniformity of appearance, no question that THAT is boring as hell. Moreover, users don't want their systems to look the same, obviously: we have backgrounds, themes, skins, and such for all kinds of GUIs. As for behavior, that still needs a bit of work before it would be worth really standardizing. I'm still not happy with the way many GUIs require you to use a pointing device, for example.Plus, there's the issue that different models of UI work differently. The Macintosh model, which Windows sort-of tries to follow except when it would be most annoying not to, is one way, and we all know how it works. Then there's the NextSTEP/AfterSTEP/WindowMaker model, which is rather different. There's the model that GNOME and KDE use, which is rather like Macintosh, except with multiple desktops, which adds both complications and capabilities. There's the Ratpoison/Screen model, which is completely different. There's the command line model, of course, and the menu-driven interface model, which many "text-mode" NCURSES based systems use. And there's probably others that I missed, either because I forgot or because I never encountered them. Now, I can certainly see how it's good to have some standard expectations of what a behavior will mean, in a given desktop model. For example, in the Macintosh/Windows/GNOME/KDE model, drag and drop means something in particular, and it should act the same on all UIs that use that model. But in a system that uses the NeXT model, does it necessarily mean the same thing? Can you even HAVE a conception of "drag and drop" in the Ratpoison or NCURSES models? My point in talking about all this is, if we make things too uniform and standard, we run the real risk of stifling innovation in user interfaces. And, since I think UIs are still a bit, well, primitive, we shouldn't be doing that just yet.
Ximian XD2: The Way Things Ought to Be - Uniformity? Posted Dec 8, 2003 16:37 UTC (Mon) by tjc (guest, #137) [Link] My point in talking about all this is, if we make things too uniform and standard, we run the real risk of stifling innovation in user interfaces.I agree completely, while at the same time recognizing that uniformity has substantial benefits. The real problem, in my estimation, is that it's usually a small group, or a single individual, trying to impose their idea of The Way Things Ought To Be on the rest of us for the sake of "World Dominion." Example: I used to be a Gnome user, for several years, until someone decided that all the OK/Cancel buttons on dialogs ought to be backwards, like a Mac. So I'd try to exit Gnome and sit there for 10 seconds wondering when something was going to happen, until I eventually realized that I had clicked the wrong button. Again. I did this for months, and was told in no uncertain terms on several occasions that this was The Way Things Ought To Be, despite the fact that 95% of the world does it different. Gee, I must be stupid of something, along with millions of other people. I eventually did what any rational person would do -- I stopped using the software. I started hacking window managers, which eventually led to user interface testing. My approach is to observe a small group of people using an interface and ask them questions. What I've found is that a majority of semi-computer-literate users (all Windows users) have a common set of expectations:
1. clicking on a window titlebar (or client area) will raise the window
Most people don't know that double-clicking a titlebar in Windows will maximize a window, so I can change the default behavior to "shading" a window and most people won't accuse me of designing a "bad" interface. I've got a lot of liberty with what I do with the middle mouse button, since most people don't expect it to do anything. On the other hand, changing the function or position of the minimize-maximize-close icons is so disruptive that no matter how superior a scheme I think I may have, most users will think it's a "bad" design. It's like trying to move the turn signal stalk in a car to the other side of the steering wheel. My goal is to design a general-purpose interface that meets most peoples expectations while at the same time integrating more advanced features in an unobtrusive way, with a minimum of configuration options. But it's a pragmatic exercise; I have no delusions that it's The Way Things Ought To Be. You may happen to prefer a 400-line fvwm2rc file, which is fine. But mine's only 200 lines. :-) OK, back to work...
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 8, 2003 17:18 UTC (Mon) by XERC (guest, #14626) [Link] Well, I can tell you a "secret": SOME PEOPLE H A T E COMPUTERS!!!I guess that that's one of the main things that most of the Linux hackers don't get. For instance, it's dificult for a child, who's all relatives are musicians and who's relatives' friends are also mostly musicians, to understand, that not all people are capable of playing a piano. As an example, my stepfather and his colleagues prefere to spend an hour every week on something arcade and uneficient, rahter than spend only one hour to learn new and better way to use their computers(instead of packing image files and sending them by E-mail, they burn them on a CD in an unpacked form and send them with DHL<www.dhl.com>). That also explains, why anything unstandard, any change at all, no matter how technically sound and effortsaving, is considered a very bad and unwanted thing and that's why "the holy Windo_s" is the ultimate goal for many people. In case of developers, who think the same, it could be that they just take Windows as the standard, no matter how bad or good it is. As the management(read:"people with nontechnical backgroud")
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 7, 2003 18:30 UTC (Sun) by davidl (guest, #12156) [Link] You have to bare in mind that Ximian Desktop is not really free software (and just a desktop at that) and adds its fair share of proprietary stuff. Although the Red Carpet stuff is certainly very good, is it really the only thing worth paying for? Why on Earth should people be paying for that along with a commercial distribution to run it on? Ximian Desktop isn't even a full complete distribution, nor do you get lots of added free software goodies as in a full distro, not to mention the densely packed licensing minefields that this entails for many organisations.You might as well use something like Xandros. It has a dedicated desktop, tools and is based on Debian so you can install third-party Debian packages without worry. Ximian Desktop is a total waste of money in this regard as it basically is (or certainly should be) Gnome anyway. Here we have the free software/commercial interests quandry which is going to have to be solved. I think the concept of Ximian Desktop is a bit dangerous for free software as a whole because it is basically Gnome with some proprietary bits stuck on it. Now and in the future, politically, this is going to be very difficult for the Gnome project and others. In response to KDE's Kiosk, Ximian talked about a proprietary solution, at Comdex I think. Services, support and customisation - yes. Pointless proprietary bits and bobs on top of free software - no. That creates the sort of lock-in people are trying to runaway from with Windows! We've seen this with so-called Enterprise Linux; whatever the hell that is. That's the motivation behind Bruce Perens' idea for, provisionally named, UserLinux, and it is something I agree strongly with. This is why we have seen the popular rise of free software in the first place, and we must not jeopardize that. The default setting of the Ximian Open Office of Microsoft Word format is just stupid because it then depends on the quality of the Word filters of the recipient to open it. As we should all know, MS Office formats are not open, and the formats are now so horrid it sometimes takes a bit of luck to use them! Ximian should be pushing open standards and really communicating those issues with people, not continuing to add to the MS Office format problem. Saving to MS Office formats and opening them should be possible, but it should not be encouraged. We are never going to get one desktop, one toolkit or one programming infrastructure to get everything to integrate, whatever today's definition of integration happens to be. We need to accept the differences that exist (and they are the source of great diversity) but come to a consenus on how software should interact with each other. Despite talk of how integrated Windows is, this has come about because Microsoft have basically said this is how it is going to be. Despite what anyone thinks they have not solved this problem - merely side-stepped it. What we need is a way for different software projects to communicate together without really imposing anything, and this is why I find the work of freedesktop.org interesting, because it's not just about KDE and Gnome. The KDE project seems to be favourable to a lot of this stuff, like D-BUS, and look at it on the basis of supporting good technologies. Unfortunately, I've seen a lot of public support for stuff like D-BUS from various Gnome people who work commercially, but private disdain on the mailing lists. We have comments to the effect of "Well people can just use ORBIT...", "I know you may have your own interests regarding Freedesktop...", "I don't see the point in it", "It is just reinventing the wheel" etc. etc. This conveniently paints over the fact that it certainly could enable Gnome apps to communicate more effectively with non-Gnome apps vastly increasing Gnome's range, but maybe that just isn't important to some people. A lot of this stuff comes from the people who have outside commercial interests (look at the e-mail addresses). Perhaps because they feel it will hurt their commercial products? I don't think I need to tell anyone who they are. I laugh really. People poke some FUD at KDE's licensing and say "Oh, Qt is tainted", "You have to pay for commercial development" etc. etc. etc. However, at least KDE have their potential commercial influences out in the open so people can see them. The licensing issue took a while to resolve, and Trolltech was understandably cautious, but there is now something that works pretty well. KDE is a free project and completely free community based on a commercially developed, but GPL'd, toolkit. The two don't mix but complement each other, and I have never seen any major inexplicable comments that were negative to any particular point of view or a project like Freedesktop on KDE's lists. This is not the case with Gnome, as the commercial pressures look very apparent in certain individuals' guarded comments and postings. The notion that somehow Gnome is not commercially tainted is naive, because you have developers working on Gnome who have interests that do not necessarily impartially reflect the interests of the wider community. You can see it in the product of stuff like Ximian Desktop and the situations with Ximian, Eazel and Gnome in the past. From a hard-nosed business perspective, Ximian is at a stage where it has to really prove itself to Novell, and the worries of people at KDE being completely replaced in Suse etc. etc. are a bit premature. The real dangers about these commercial influences are not that they are there, but that they are unseen and not out in the open. You could argue that Xandros, Lycoris or Lindows build proprietary stuff on KDE, but they do not have developers involved with KDE who are talking down utilizing Freedesktop work or meshing KDE development and commercial interests together. They are companies that build on free software, but don't try to influence and possibly change it for their own interests. Linux has ridden the storm of accusations of fragmentation extremely well through the open source, community and open standards uses and processes. It is a battle Linux has won comfortably. Open Source desktops have not yet ridden this storm, and at the moment we are in danger of not being able to ride it out because of the commercial interests and proprietary extensions of many different commercial organisations. We must not let things fragment as various Unices and free software has done in the past. We've let Microsoft and others in before; we must not let it happen again. I know people will dismiss this as a troll and FUD-spewing without reading what I've written at all, which always seems to be the case in these things. Oh well, *sigh*.... Nevertheless, the facts are there if people look hard enough and don't get blinded by religious or political aspects. I have written this to point out some of the pitfalls Open Source desktops still face (and many of them are not technical), and they need to be crossed together as a community. Perhaps some people will agree with at least some of what I've written. No doubt I'll get some colourful comments from some Gnome-oriented people but I won't be reading them because I didn't write this to start a stupid flame-war. Ximian Desktop is most certainly not the way things ought to be.
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 8, 2003 3:53 UTC (Mon) by elanthis (subscriber, #6227) [Link] What the hell are you smoking? There isn't anything proprietary in XD2! Red-carpet is GPL, all of the changs made by Ximian are GPL/GPLd, all the source is available, and as much of it as possible is contributed back into stock GNOME. The only reason XD2 exists is because stock GNOME doesn't have uptake of UI tweaks and improvements fast enough - Ximian doesn't sell a desktop, they sell support and services for the desktop and apps. The *only* proprietary product that they offer in any way shape or form is the Evolution Connector, which *isn't* a part of Ximian Desktop. Check the facts before you rant.The default for the Word documents is do to customer request - over here in reality, less than 1% of people can even *open* OpenOffice.org documents, and neither a business nor a home user is going to want to have to deal with the inconvenience of using Save As over Save just because of a (relatively) small number of open format proponents. It's a pain when I use Abiword, because if I'm not careful, I end up sending documents to people they have no realistic hope of reading (telling them to download software to read it isn't an option, either). I don't like using Word documents, but Ximian's move was a smart one for everyone involved, except the zealots... The D-BUS comments are completely bogus, I don't know which orifice you're pulling that out of - the proponents and opponents of D-BUS are spread across the "home" hackers and commercial hackers; the arguments, if you actually read and understand them, are to do with changing underlying architecture - some people think binary compatibility is *good*, since that makes life easy for developers, distributors, and users- and these people don't think D-BUS offers anything wroth distrupting that compatibility. It's not any kind of corporate conspiracy.
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 8, 2003 12:27 UTC (Mon) by davidl (guest, #12156) [Link] There isn't anything proprietary in XD2!What on Earth are people paying for it for then?!
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 8, 2003 16:56 UTC (Mon) by Nikolokolus (guest, #9916) [Link] Well, I'm sitting at my linux box using XD2 as I type this and I never paid a red cent for anything. I could have subscribed to their support services but chose not to.Mainly I downloaded it just to see what the big deal with it is. I can say that after a couple of weeks of use I pefer it to stock gnome and KDE (which I used to use most of the time). I don't know if there is anything proprietary or not, but it does just work; and that I like.
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 8, 2003 13:36 UTC (Mon) by anselm (subscriber, #2796) [Link] I'm told that while the client side of Red Carpet is GPL, the server sideisn't. Any truth to that?
Ximian XD2: The Way Things Ought to Be (OfB.biz) Posted Dec 7, 2003 18:35 UTC (Sun) by penguinroar (guest, #14460) [Link] I would rather see a stabilization in the api's than on the desktop. Maybe its to soon but i hope its not to far away. A stabilization in the api's would make it easier to make 3rd party apps for all dists. As of now its pretty much a running target to make binaries.Maybe all that we need is a standards base for desktops that is followed by the distributors? A standard set of API's would also attract developers to those dists that conforms as a developer wants to reach the biggest possible audience. If that isnt plausible why not make it easier to run api's of various versions side by side? Anything as long as the base system isnt rendered ancient to fast.
But the APIs are standardized (and have been for some time) Posted Dec 8, 2003 20:37 UTC (Mon) by hazelsct (guest, #3659) [Link] That begs the question: how different are the XD2 APIs from stock GNOME 2.4? For that matter, how different are any of these APIs from GNOME 2.0, which after all had the goal of stabilizing the GNOME API? My limited impression from just a little bit of hacking is that the APIs have been quite stable since 2.0, with the only changes relevant to 3rd-party app developers being the growing list of bonobo modules officially included in each subsequent release.To address your second point, of running different APIs side-by-side, that's already made possible via shared library versioning. So if app A links with libgnomeui.so.32 (GNOME 1.4), and app B links with libgnomeui-2.so.0 (GNOME 2.x), then "apt-get install A B" will drag in dependency packages libgnomeui32 and libgnomeui-0, installing both APIs, and A and B will both run just fine, everybody is happy. It all "just works", and has for many many years. (IMO, though the APIs have been quite stable, the desktop itself and apps have not really been "stable" until 2.4, corresponding to XD2. That gnome-pilot still hasn't released anything with working memofile conduit, and gnucash hasn't forward-ported to GNOME 2, says that parts are still not stable, at least for my own use. Which gives some justification to Debian's glacial release cycle: woody released with complete GNOME 1.4, sarge will have complete 2.4. But all that's off-topic for this thread.)
|
|
||||||||||||