Norman Feske at the technical university of Dresden, Germany, is developing a new windowing system for the experimental Drops OS, called DOpE. DOpE runs under Linux as well and it can run X11 on a window. Read the proposal explaining the project (in german) and check out some screenshots.
How does this compare to http://www.berlin-consortium.org/ ?
—
http://islande.hirlimann.net
Use http://translate.google.com/ for those of us that don’t speak deustche!
OSnews does not support url UBB tags
(yawstlla) yet another windowing system that looks like ass. just what linux needs to conquer the desktop. Oh well.
Hmm, this looks pretty cool. The fonts look ok, and yeah, I was mainly looking at the fonts lol…me and those fonts Where can we download DOpE from? I didn’t find a “download” section.
Well, as I don’t read German and Babelfish isn’t really doing that great of job it’s hard to tell if this is aiming to solve some of the underlying issues of X or not.
But at any rate, I don’t ever forsee Berlin evolving into a usable window system. CORBA was not built for the kind of performance necessary for a display server (then again, neither was most any form of IPC in existance today)
From a design standpoint, Berlin is plagued by design decisions which will forever hinder performance. It’s a great thing to look at from a research standpoint. I just don’t ever see it becoming something that would see day to day use by a widespread userbase.
Wow! I did not expected that OS-News dedicates a news
entry to DOpE. Thanks.
Just to clear something up:
DOpE is primary meant as a lightweight user interface for
the Drops research project. Drops is a realtime operating
system, based on the L4 micro kernel API. Consequently,
DOpE has to be ‘realtime capable’ – an issue, that is
covered by only very few windowing systems (e.g. QNX).
The most interesting fact about Drops is L4Linux – a
Linux-kernel running at user-level on the top of the
Drops-kernel. So Linux is just an application, which
can be executed concurrently to other tasks, especially
realtime tasks.
The screenshots show DOpE running under Drops, which
is running L4Linux.
The work does not intent to solve any problems of
linux (honestly, I can live with xfree4 very well).
The visual appeariance of the user interface will
most probably change with its further development.
matt: I really like to see a suggestion, how a gui
should look like.
At the moment DOpE is at an early development stage.
When a stable version is ready, it will be made available
to the public.
D3wd, this thing is like,d0p3!
D3wd, this thing is like,d0p3!
Lol!
Does this mean that after using DOpE, you will either
a) Get the munchies and migrate to Pizza-Box Linux or
b) Get really paranoid and migrate to Tinfoil Hat Linux?
YAONJWDGI (Yet Another OSNews Junkie Who Doesn’t Get It). This is an experimental GUI for an experimental OS that is only tangentially connected with Linux. This has nothing to do with Linux konquering the desktop!
is berlin still actively developed? I check their website from timt to time, but the last news item is from february …
fred
Well this is looking good, progress for an X replacement. Yes many people like X because of it’s client aspects but I would have to think the majority of people who would like to use linux or do use linux only want/need a desktop for the computer their at, not so they can use a computer 12k miles away. If this is much faster and more stable this could do well. The screen shots don’t look bad, nothing amazing compaired to others but not bad. Hopefully it doesn’t turn into a mess of differant API’s and desktops and configurations. Just have one setup and some flexibility on sizes and colors and everything will be much nicer.
As long as it is NOT the following:
c) Getting a Dell.
–JM
goddamnit, the problem with having such a common name (and using it) is that somewhere there is a weenie using the same tag…
Norman: you obviously speak good english, any chance you can translate the document? i have tried google and altavista and they both peter out before the Tcl/Tk section. unfortunately my (very old) school german isnt good enough to read the rest of the article, which looks very interesting.
having used RiscOS for quite some time i think its nice that someone finally recognises the usefulness and innovativeness (is that even a word?) of their interface. it would be good if the file save interface was extended to a system like graphical pipes, so i can have a file that i just wrote and ‘graphically cat’ it to a printer, spell-checker, MUA, etc…
i think 8 1/2 tried to do this, but without much success…
How does this compare to http://www.berlin-consortium.org/ ?
I don’t know how it compares with Fresco (Fresco is the new name, the name would change when 0.3 is released) and DoPE because the specification is in German, and both Google and Altavista only translate the beginning (there is a word limit).
(yawstlla) yet another windowing system that looks like ass. just what linux needs to conquer the desktop. Oh well.
At an alpha stage, if looks is your #1 goal, you wouldn’t succeed.
is berlin still actively developed? I check their website from timt to time, but the last news item is from february …
Yes, it is being actively developed. The team is transitioning to a new website (http://www2.fresco.org) that changes the layout as well as the backend, and therefore the site remains outdated. The best way to find out the latest news about Fresco is by signing up from “fresco_devel” mailing list on their site. Or go to #fresco on OpenProjects.
0.3 is coming out soon, but the developers is trying to remove some major bugs before releasing it. Plus, they want the new site up prior to the release.
Well this is looking good, progress for an X replacement.
Maybe it is an X replacement on Drops, but not on Linux 🙁
matt: The proposal is rather old. I wrote it before
I started the practical work. It is just a
summarisation of my ideas. Currently, I am working on
a new (english) document describing the concept. At the
moment it covers mainly the technical stuff, like
redrawing strategies, client-server communication etc.
In the past there were a lot of great ideas around
user interfaces. E.g. I love the way, how RiscOS let work
its applications together. Graphical piping is actually
not far beyond that. As stated in the proposal, my work
is very much influenced by RiscOS, Nextstep and Tcl/Tk.
rajan r: In fact, a x-replacement for Linux had to
implement the x-protocol for not spoiling the big
amount of existing x-applications. There are no
applications for a completely new approach (like DOpE).
Thus, there is no argument to use it instead of X.
On Drops, the situation looks differently – there is
no established user interface and I have the chance
to establish a mine 🙂
I was just about to point out Fresco’s change of name from Berlin, but you did it first.
I can’t see any fatal design flaws in Fresco. I mean, what’s wrong with vector graphics (and therefore consistent widgets), native alpha transparency, CORBA, native Unicode, device independence, and BeOS-style multithreading? The only thing there that could possibly pose a problem is CORBA. But CORBA has less overhead than X, because its use of vector graphics allows it to be much more efficient with CORBA than X could be with the X protocol, so the only thing that Fresco wouldn’t work well in is embedded devices. And they don’t need a GUI in the first place.
Of course, what we really should be asking is, what do we need a GUI for in the first place? Early interfaces based on recall had the problem of using cryptic commands that no one could possibly remember. But if somebody could type in (or handwrite or speak) the word “Email”, and have their email program come up; or the word “Web”, and have their browser come up; then that completely changes GUI concepts. And people generally only run one application at once, anyway. So, make a GUI that only has one application open at once, no titlebars, always maximised… essentially, what Palm OS does.
We’ve got to remember, always copying Windows ain’t gonna work. We need some good ideas of our own once in a while.
Run one GUI app at a time from a command line?
I don’t think this is a good idea for productivity, but better use of the run box is a must.
I would say any GUI needs at least needs some sort of graphical menu to sift through to find apps, and some sort of pallet to see and select apps that are already open. Even blackbox and fluxbox offer that much. My personal preference is for these two things to consume as little space and CPU as possible. This is a very basic concept and MS was not the first to come up with it. A menu is a menu, there is very little that can be changed there and some sort of 3-dimentional cube thing would be cumbersome and CPU intensive. The portion of this where you _can_ be inventive, is on where to store minimized apps so they are both easily accessible and out of the way.
rajan r: In fact, a x-replacement for Linux had to
implement the x-protocol for not spoiling the big
amount of existing x-applications.
Personally, I wouldn’t mind ditching all the X applications and starting a new window system from scratch. Fresco did so because of a lot of reasons. Here are them
1) X applications aren’t consistent with other X application. KDE apps look different from GTK+ apps that looks different from Motif apps…
2) Due to X11 itself, most X apps are written to be raster based (the UI again). This would be a problem for Fresco which is trying to make it mostly vector based.
Fresco has a long way to go. But if GTK+ manage to get a lot of developers (mostly C developers) to it from QT and Motif, I think Fresco could do it. Besides, it may not be in the near future the thing that replaces X, but a clean break is better than continuation. Besides, what X apps that are so important? If apps where the issue, I think writing something like Windows’ graphics and window system would be better, no?
On Drops, the situation looks differently – there is
no established user interface and I have the chance
to establish a mine 🙂
Good idea, but Drops has nothing special over X using Hurd :p Besides, I would just like to know the technical ideas behind the windowing system, not the ideas of an UI not yet implemented yet.
jim:
At the current stage, the windowing system is just
technical – not a fully functional user interface.
Surely, stuff like menus and icons will be added later.
rajan r:
> Personally, I wouldn’t mind ditching all the X
> applications and starting a new window system from
> scratch. Fresco did so because of a lot of reasons.
Your a certainly right, but when aiming at the desktop
‘market’, the existence of applications is all end
everything. So it would be silly to try competing
without a broad base of existing applications.
I will inform you when my document about the technical
ideas behind the windowing system is ready. This will be
ready end of september (sorry, I cant provide it earlier –
it is my diploma).