the idea of using netscape for a linux GUI seems a bit half baked...

i REALLY like the idea of having an improved file manager (drag and drop between ftp/smb/local) built into mozilla.. basicly something like ie, but even better..

my only concern is that while nglayout seems to be quite small and fast, i'm worried that all of this XUL stuff is gonna get slow, buggy, and bloated... i'm not sure that xpfe was a good choice.

It may seem half baked because a lot of research needs to be done into how such an idea would be implimented.

However, this comes from www.berlin-consortium.org/timeline.html

--> Addition of XSL and/or Gecko Flow Object classes This is still in the "blue sky" phase, but it's something to work on. The question of whether it's worth trying to make a corba loader for XPCom objects remains to be resolved. If not, we'd be reimplementing a lot of stuff. <--

Could this be a possible route to go by having Mozilla and Berlin working together?

with berlin that might be a possibility. but it seems as though adding yet another ui layer to any os is a bad idea.

i see real use as mozilla being used as an application, but i'm not sure i like the idea of it being an interface library of sorts. sure, the idea of netapps based on HTML/XML/XGIMME2LETTERS has been around for awhile, but i'm not sure the benifits outweigh the overhead of having a browser doing the display work.

it a nice idea, but it seems like its been implemented too many times already. how many interface libraries already exist on linux? qt/gtk/swing/motiff/....

i just hate having yet another lib installed on my machine. windows suffers from the same problem, dll hell.

i say keep it simple.. pick one ui library, then build good apps on it. don't reinvent the weel 25 times over..

I can agree with that, but are any of the existing worth building on? Are they ready for the future? I keep hearing complaints that the UI of Linux sucks, and I can agree, no matter if you are running Gnome, AfterStep or whatever.

Linux is just now starting to take off, so right now is the time to come up with something good.

There are a lot of benfits to my suggestion. You point out bloat and potential problems with glitches as negatives, and that is problably true if you have 20 different interface libraries installed. A new standard is what I am talking about here.

The Linux community needs to decide now on a path to take before the platform is dismissed as a real alternative from Microsoft. Sure, you and me might be able to use a half-baked GUI and command lines no problem, but a modern OS nowadays has a full blown and easy to use GUI and graphics capabilites.

If someone does not do an XML driven GUI first, Microsoft will. And they will receive high praise for being great innovators.

People find the internet easy to use, and a highly customizable OS that works better and easier than the internet, while being integrated with it would take off like a rocketship.

Creating an awesome GUI for apps will become much easier, and developers surely would at least be interested. As they find that a XML interface is much easier to create, they will dump what they are doing and come on over. Don't forget about easy portability to other OSes.

Mozilla seems to have the clout, tools and experience to head up something like this. I hope it is at least considered, and I am only testing the waters here before I consider pushing for a proposal.

Hmmm. I'm by no means a tech guy or anything, but my initial reactions are (1) didn't 1.0 of GNOME just come out like this week? And you're talking about overhauling the GUI again? (2) Mozilla itself isn't even DONE yet-- in fact, completion is still a ways off, and it's proving to be a massive, slow-moving project on its own-- so this idea seems premature, (3) there's enough flaming and arguing over GUI/DE Linux standards between KDE/GNOME, and (4) For clarity, I like the idea of keeping the browser seperate from the UI.

This reminds me a bit of when people were saying "Let's make a MozillaOS" a year ago...

It's nice that we have these cross platform tools and stuff, but I dunno, from looking at freshmeat.net it looks like GTK is slowly becoming the Linux standard... What would be the real benefits (vs the costs) of doing something like this? Niave question but, wouldn't it be as easy for programmers just do use XUL for their individual applications?

What do other people think?

W

Linux has an ease of use problem. So what do people do? Instead of coming up with one standard way of solving the problem, everyone and their uncle tries to come up with their own solution. There's KDE, now Gnome, Corel also wants to get in the act... End result: it just gets too confusing. I think the linux community should get together and ask themselves-do we really care about making linux easy to use or not? If they care, they'll stick with one dominant standard (always giving others the right to choose alternatives) but if they don't care then might as well stick with a commandline.

oh mozilla just plugged in the expat xml parser. cool

I may be missing something, but the arguements that I have been reading in this thread sound a lot like the arguements that M$ used in there defense. NOTE: I am not advacating that M$ is correct.

Personally, I think the idea should at least be tried. It's good that people are aware of the problems of integrating Browser + OS, but I think that the open source guys are on the ball enough to deal with these problems.

The proposal is exciting and interesting enough that some work should be done on it. The end result (if it stays true to the "keep it as slim as possible and avoid code bloat" faith) could really take Linux into a different league.

Go for it!

While I believe Eric's comments have merit, what is the point in creating a whole user interface based around the web browser's architecture when two large scale efforts which are close to being complete have already addressed this? Has the man never heard of KDE or GNOME? Simply make Mozilla GNOME-compliant (offer services via ORB), and be done with it. Mozilla uses GTK, GNOME uses GTK....it seems pretty obvious to me. Integrate Mozilla into the GNOME project and you pretty much have what was being described......

I for one think it's a fantastic idea. Linux could really use a UI with NS style and quality, and since both are open source the result could truly be unified in a way that is not possible with Windows. Both projects could potentially benefit.

As an aside, check out AmigaMoz! http://home.clara.net/tom.hurst/amisite/

It is a daunting project, however, and for the time being most Mozillans are likely going to be concentrating on getting Moz5.0 out the door.

By the

way, I echo the comments above regarding "one standard." This is MS type speak. As long as the different UIs take one another into account and continue to do so in the future, there should be no "one standard" GUI--that stifles competition. I mean, people have the "choice" of alternatives now, that doesn't mean they are intelligent choices to make for anyone other than the most astute programmers.

Hello Eric,

different ideas never hurt..

http://www.nl.debian.org/berlin/arch.html

http://www.progressive-comp.com/Lists/?l=linux-ggi&m=87915016500941&w=2

http://www.progressive-comp.com/Lists/?l=linux-ggi&w=2&r=1&s=mozilla&q=b

http://www.nl.debian.org/berlin/mail-archive/msg00703.html

dunno about berlin, but GGI was also ported to OpenBSD, Amiga, you can download with CVS and port it.

There are a ton of great ideas out there but I think only a major agreement in the Linux world will sway people in a certain direction.

Any other ideas?

Any of you be interested if I put up a web page on the subject of an XML driven interface?

oops..I guess xpfe is not easy to port to GGI because of xpfe depends..maybe a GDK/GTK wrapper?

See Windows 98 for a stunning example of how NOT to do an interface. Frankly the current browser navigation paradigm is klunky and awkward. Using it to navigate around your system is excruciatingly annoying.

Software currently available has a lot going for it. Enlightenment is a nice window manager and I love its ability to be themed to anything. GTK+ also has the theming ability -- between the two I can give my OS a look and feel that applies to all GTK programs and Enlightenment. Gnome is quite nice now, and the whole system can be quite visually pleasing and laid out much better than the atrocity Microsoft has inflicted upon us.

Lets not try to emulate a bad idea and design. You shouldn't always do something for the sake of doing it.

Win98 is very bad and I'm worried when people say things like "Windows is easy to use and linux needs to catch up to it." I don't know why people should look at Windows as an example-what ever happened to the Mac? (or BeOS even though it's missing several things like a startup folder, file manager,etc)

i agree that win98 is not a good ui.. but some of the idea's aren't that bad.. using a browser to navigate your hd is not useful.. ms got it backwards, what would be nice is the ability to navigate the net as if it were local.. i want drag and drop between ftp, local and smb dirs... make local/remote all seem local, not everything seem remote...

i think a useful fm is really what linux is (as some claim) missing.. switching tasks is the job of the window manager.. other then that all that is needed is a nice file manager.. dfm is pretty nice, but still needs work, kfm is slow, and i've not touched whatever comes with gnome, but it's icons seem to be too large to be useful.. if mozilla gets a nice ftp ui and adds smbclient functionality, it would be pretty nice.... as long as they can keep the bloat down...

making the layout engine modular to the extent that NGLayout does not depend on any given toolkit (sic!) will enable someone to create new interface frontends.

and what the graphical interface look like is not necessarily all that matters..as in what buttons to click with mouse..for instance if you do not use a mouse, keyboard shortcuts might be available :-) press h for help.