The Syllable team makes strides these days on their AtheOS fork, and most of the AtheOS developers have joined Vanders and Rick to their quest of a better operating system. OSNews reader Daryl Dudey joined the Syllable team recently, and he already wrote a nice networking preference panel for Syllable. The team is still looking for a kernel developer though, or C++ developers who like writting low level system software. On a related hobby-OS note, SkyOS made its first steps into the SMP world.
Syllable Team Marches On
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 the ported & modified source code of ABrowse* available? Can’t remember it being part of the standard AtheOS source code… If not, is this the reason why the Syllable team is relluctant to move to GCC 3.x?
IMO, Syllable is new, and now it is probably the best time to move to a new version of the compiler that might introduce binary incompatibility (which does not matter at this point, as everything is being recompiled for Syllable atm – but is ABrowse’s code (the most important app on Syllable/AtheOS) available to recompile?)
* (ABrowse is the AtheOS/Syllable web browser, based on KDE’s Konqueror/KHTML 2.1 (coded as of April 2001))
And to add to that question, has any effort been made to sync up with the current KHTML code base? I don’t know if the KHTML engine has been updated, but the Konqueror browser has gone through several rendering and functionality updates since Kurt originally ported KHTML for AtheOS.
It is not easy to upgrade the KHTML. The way it was ported was literally changing the Qt API over to the AtheOS one. Syncing today it means pretty much, re-porting the whole KHTML Kpart over again, which is something that is pretty difficult to do and requires quite some time…
And there are more pressing matters today for Syllable as an OS, than try to upgrade the browser. The OS doesn’t even have swap or IDE/CD-Rom support yet… ABrowse/KHTML 2.1 is more than enough for Syllable as it stands today. It is just important to actually have the porting code readily available, so if the team goes to GCC 3.x, to be able to recompile it, as it would most probably will break binary compatibility.
So, it is much more easier to use Gecko?
Why would it be easier? We are talking about a full porting again, as in the case of KHTML. And in fact, Gecko is even bigger, more complex and moreover, it is not pure C++ as KHTML and AtheOS/Syllable is. So, I don’t see why Gecko would be easier to port than KHTML in this case.
To quote my husband, who frequently philosophizes: “There is no portable code. Only code that has been ported.”
OT. has he ever heard about NetBSD?
You don’t need to “port” Gecko to make mozilla run on Atheos/Syllabe, just other much simpler code that makes Gecko work (low level system lib, basic gfx and widget functionality, etc). That’s why Mozilla runs today in more than 25 SO. For example, AFAIK, to BeOS basic port was made in less than a week.
The ABrowse source is in the CVS Repository, however I admit I didn’t spend any time trying to get it to compile. CVS can do many wicked and subtle things to source code, not to mention we do not know the original runes used to compile it in the first place! The ABrowse package we provide is simply a gzipped tarball of the ABrowse directory from Atheos 0.3.7
Getting ABrowse up to date would be a “full time” job, in the sense that anyone taking on that work would wouldn’t be able to simply do the work and then move onto something else. The code requires a dedicated maintainer. Its not just porting/updating khtml but it is also the GUI & things like cookies that also require work.
Its the same situation for Glibc, in fact. I’ve spent weeks on the Glibc source from the atheos.cx CVS, and it still won’t compile. Something as large as Glibc really needs a full time hacker to keep it current. If anyone has experience of hacking Glibc, I’d be interested in any help you could offer!
As far as GCC 3.x goes, anybody is free to do the port, and I would happily keep GCC 3.x as an official package. However, nobody really has the time to fix the GCC 2.x C++ (We already know there are some problems, as Bill Hayden went through them with Cosmoe). Everyone is rather more interested in bringing the GUI upto date and getting all of their ideas out and into the open.
Looking at the current work being done, what will probably happen is that we will continue to work on 0.4.x releases that will fix the various missing peices (The VM, GUI classes etc.) and then start some “modernisation” work for 0.5.x, E.g. moving to GCC 3.x, bringing Glibc & the development tools upto date etc. etc.
the only two ways i can think of would be porting the app (khtml, whatever) to the BeOS api, or retargetting (think wrapping) the entire kde api.
both of these would take ages to do!
khtml has got the cleanest implementation and least amount of nastiness of any browser code i have seen, though. its streets ahead of mozilla in those terms.
NO KDE! no no never!!!! Not on syllable.
Given that Atheos now has a new fork & energy, what is the future goal?
Would it be reasonable to borrow some of the Cosmoe effort to allow BeOS APIs to compile & run or does the Syllable team want to keep the Atheos APIs?
Syllable is intended to be a good, Free, usable desktop operating system. I’m sure you’ve heard all of the comments about Linux “never is no good on the desktop”? Well, I’m with them! I want Syllable to fill the desktop space in a way that Linux currently fails to do. Because we’re aiming for a well defined target, we can concentrate on the things that are important to desktop users, and hopefully that will help us to focus on doing things the Right Way.
We are not trying to clone BeOS. There are four or five different projects out there already which are all trying to do that, and I see no reason why we should tie ourselves down to somebody elses design. Curently Cosmoe and Syllable share enough common code that patchs from one could apply to the other with a little work, but thats simply not going to last. I should imagine that within two or three releases of Syllable we will have extended the API’s enough from the original AtheOS API’s that Cosmoe and Syllable will no longer even be realisticaly source compatible.
Essentially Syllable has two overiding rules:
1) Design it for the Users
2) Keep It Simple, Stupid!
Hey, my first front page mention!
Thanks Eugenia.
Syllable is starting to gather some real momentum now, we have a wiki set up and the ideas are really flowing now, the mailing lists are reasonably quiet at the moment because everybody is coding away (hopefully!)
As far as my networking app goes, the code I am working on now is working now and coming along very nicely. I have got some ideas that should make it a core part of the networking subsystem, mainly that it will be capable of bringing up interfaces on boot and detecting new interfaces.
I have to say, to those who haven’t tried it, Syllable feels very nice to actually use. Boots fast, runs quickly. There are a lot of areas that need work but thats just a matter of time and hard work.
It may have been ported fast, but it was around for many months before it was USABLE…
And in fact, Gecko is even bigger, more complex and moreover, it is not pure C++ as KHTML and AtheOS/Syllable is
…and that’s probley one of the reasons Kurt used KHTMl and not gecko.
[i]”There is no portable code. Only code that has been ported.”[/i[
I agree!
Does anyone have a VMWare Image of Syllable? I’d really like to try it out, and that would be the easiest way for me to get it up and running. Atheos runs under VMware, so Syllable should as well
Thanks.
I really feel that though you guys are not aiming for BeOS compatbility, their are tons of projects that are going to be having similar APIs thateither working togetehr onm APIs or even combing the GUI efforts would be best for all.
As for browsers, Cosmoe shows how close AtheOS is to beOS, why not try out the Mozilla port to BeOS, BeZilla. Though its lways ncie to have a lightweight alternative like Galeon for GNOME, K-Meleon for Windows, the OS X one, etc
Kurt’s KHTML port was done using wrappers around the Qt classes. He intentionally touched as little of the original KHTML code as possible in order to make it easy to synch with future KHTML releases. Updating to a newer version of KHTML would be much, much easier than porting Geko.
I don’t understand… why do you link to OBOS website on your homepage when you are in the team of another OS?
Why would it be easier? We are talking about a full porting again, as in the case of KHTML. And in fact, Gecko is even bigger, more complex and moreover, it is not pure C++ as KHTML and AtheOS/Syllable is. So, I don’t see why Gecko would be easier to port than KHTML in this case.
Okie dokie. But I would rather have Gecko instead of KHTML… mainly because Gecko supports more standards than KHTML. But right now, I don’t think ABrowse is that important to bother in Sylable, a lot of other things need more attention than ABrowse.
NO KDE! no no never!!!! Not on syllable.
LOL, read the comments above… nothing to do with porting KDE over, but making an implementation of KDE’s APIs so porting KHTML would be easier.
Given that Atheos now has a new fork & energy, what is the future goal?
To became bloated 20 years after the 1.0.
It may have been ported fast, but it was around for many months before it was USABLE…
And still isn’t usable….
Updating to a newer version of KHTML would be much, much easier than porting Geko.
No one wants to port Geko. We are talking about Gecko! Okay, i’m getting lame.
> I don’t understand… why do you link to OBOS website on your homepage when you are in the team of another OS?
History lesson:
Well, the Atheos and BeOS developers have a lot of common history. Most of the AtheOS devs are ex-BeOS devs. The two projects are not so far away from each other, and you will see many of the same developers working or encouraging both the Atheos/syllable/cosmoe and cosmoe/beos/openbeos projects. They are pretty much the same people we are talking about…
Yes, Eugenia is right. The other reason is that my page is pretty new and I am not finished putting links etc up there yet.
I am glad you reminded me actually, I am going to put links to all the related projects OpenBeOS/Cosmoe/Syllable. They are not sister projects, but they are all going in similar directions, but with different ultimate goals.
Anyway, I can’t wait to see a releasable version of OpenBeOS. I was going to get involved with BlueOS (BlueEyedOS? now) but couldn’t find a point of entry as such. Anyway of all the above OS I think Syllable has the best chance of being successful, if only because it is so useable already.
Daryl.
I tried at one point to get ABrowse to compile (from the source tar ball on atheos.cx) but configure errored out. I tried feeding different opions and same result each time. As Vanders said, we need a full time maintainer of ABrowse, or someone to port a different browser engine. As Jon said, the source is just wrappers to KHTML, and the entire source for the browser (not counting the KHTML wrappers, and KHTML) resides in just 1 file and would be easy to improve, once someone can get a compile.