Syllable developer Daryl Dudey is hard at work the last few months creating new preference panels for the Syllable operating system. Check out the screenshot. Additionally, Syllable is still looking for kernel & driver developers to join the enthousiastic team. Also, here’s an updated installation guide for Syllable.
What do all these new desktop-orientated operating systems have in common? None are being made by commercial entities. They don’t like the water.
Here is my top 10 of things I would like to see Syllable to have before by the end of the year:
1. IDE driver for hdds and CDs.
2. Some fixes on the widgets, GUI and default coloring.
3. Switch to GCC 3.2 and build for mcpu=i586 march=i686.
4. MTRR support, and a faster VESA driver. A lot of people that use AtheOS or Syllable are resorting to VESA, and currently that driver is just terribly slow.
5. Swap support and better VM. A better kernel Scheduler.
6. Some fixes and more functionality in the app_server and interface API.
7. A “Media Kit” in C++, that would allow a good framework for drivers and sound apps too.
8. A file manager that works. OpenTracker would be great btw…
9. An easy Installer, similar to BeOS, that could let you install Syllable at any time, by any supported source, even when you are booted to it.
10. The fixes needed in the kernel to allow PostgreSQL support. Port the latest PHP and Apache 2 too.
Dear Santa, this year, I’ve been a very good girl…
Hey Eugenia,
Sounds like your asking for BeOS!
To be honest, It has crossed my mind a few times to simply take the Atheos code, make the changes to talk about, and integrate OpenTracker into it. Then when the other ‘kits’ that are in development now at OpenBeOS are ready, we may actually have another BeOS clone that works!
Hmm, maybe you’re doing this yourself! I would not put it past you, you crazy porting genius you =)
Actually, it is not BeOS at all. Syllable lacks a lot of things currently, that _all_ OSes have, like swap or a Media Kit of some kind.
The only BeOS-like features I would like is the Installer, and possibly OpenTracker. But Kurt Skauen tried to port OpenTracker a year ago, and it proved very difficult, as OpenTracker is a “client” of the BeFS, not just BeOS desktop…
Dear Santa, this year, I’ve been a very good girl…
I think 10+ million greeks may disagree with that ๐ Wow, how good were you the year before ๐
Sorry, I couldn’t resist, and please don’t take offense.
I still believe that I was even more kind than I should have been…
Oh, I’m (mostly) on your side (and I’m also greek btw).
Send me an email to get together for a drink if you are also in the Bay Area.
Hey, I’m thinking of throwing a party soon (within the next couple of months) — I’ll send you an invite!
I just installed Syllable a couple days ago… Now if only I can ssh in… For some reason, sshd refuses to authenticate any user that I try to log in as. When I run in debug mode, it keeps telling me that the user “adamk” is an illegal user (same if I try to ssh in as root), despite the fact that I’m definately not an illegal user ๐
And, unfortunately, no one on the syllable-developer mailing list has responded to my pleas for help… Can anyone here help?
Adam
I had this similar problem with ssh under Slackware 7.x. I think I added ‘-lcrypt’ while running configure script. And it was fixed…
%> LIBS=’-lcrypt’ ./configure –prefix=/to/some/where
NOTE: I run it under Slackware Linux, not sure what shell Syllable is using, the way to add more to LIBS maybe different.
Good luck…
Well,
I suppose now that OpenBFS is available the porting process for OpenTracker could be simplified if you just replaced the FS in Syllable…
Eugenia?
Yeah… thing is that Syllable has its own filesystem. While the two filesystems are similar, Kurt had made a few changes in the architecture of AFS, to be faster… I think he writes what he did, in his site.
I think the Apply and Close buttons make a lot more
sense than Ok, Cancel and Apply.
But I would Prefer a layout like this:
Default Revert (some space) Apply Close
Not sure about changing the button positions, they are defined in the Styleguide and we wan’t to keep a consistent look.
BTW: The screen prefs app is not complete yet, I have actually been working on it in the last couple of days and have made good inroads. Don’t assume that is the finished app!
RE: ssh problems, the mailing list is quite at the moment, I am hoping it is because people are busy coding!!
Eugenia has made some suggestions on the apps, so I shall be updating them today. Obviously I wan’t Syllable to look as good as it could be, and the prefs apps are a big part of that, being part of the system dealing with possibly difficult choices (i.e. configuration)
Also, on the way after the screen prefs app is done is a colour/appearances (which includes title bar theming) and a boot manager app to configure grub.
Lots of interesting development ahead, if you are looking for an interesting sensibly designed OS to work on, dive in and join the mailing list. There is plenty to be working on right now.
Why try to do a BeOS clone of every OS out there?
Ive seen OSX shot with BeOS themes, Windws XP! <- Ugly as hell
Syllable, Leonardo, B.EOS, Cosmoe, AtheOS, OBOS every of them have simularities with BeOS (some more than others)
I have to say, Be.Inc did success, they did a change in the OS world, atleast afterwords….
Please dont turn Syllable into another Clone project, instead lets talk and exchange ideas for your clones (Syllable doesnt inteand to be a clone).
I think an “Attack of the Clones” site would be nice, there we all can follow the progress from every OSS project?
Any Ideas? (daily updated)
/Konrad
thing is that Syllable has its own filesystem. While the two filesystems are similar, Kurt had made a few changes in the architecture of AFS, to be faster… I think he writes what he did, in his site.
So then the idea would be to change the AFS API to be compatible with OpenTracker… or the other way around, either way I think this would be hard, and if Kurt found it difficult, then it must be…
=(
> 3. Switch to GCC 3.2 and build for mcpu=i586 march=i686.
Shouldn’t that be the other way around, i.e. “-march=i586 -mcpu=i686”?
Eugenia do you have any screenshots of this OpenTracker?
( I have heard about it but since I am not a Beos user I haven’t used it)
What kind of filemanager does Syllable use right now?
P.S. Did you notice that I spelled your name right this time.
>Shouldn’t that be the other way around, i.e. “-march=i586 -mcpu=i686”?
Yeah…
>Why try to do a BeOS clone of every OS out there?
Syllable is not a BeOS clone. It has some architecture similarities to BeOS, but it ends there. It is like saying “don’t turn Linux or OSX into a unix”. They are Unixes deep down. And Syllable has some similarities to BeOS already.
>Eugenia do you have any screenshots of this OpenTracker?
If you see ANY BeOS screenshot with a filemanager, that’s OpenTracker. OT also includes the Deskbar, which is the BeOS menu/taskbar system. Here is one:
http://www.operating-system.org/betriebssystem/bsgfx/beinc/beos-scr…
But the power of OT is in the way it works, not on the way it looks.
>What kind of filemanager does Syllable use right now?
None particularly. There is an included “demo” fm app, but it isn’t really useful.
Konrad: Please dont turn Syllable into another Clone project, instead lets talk and exchange ideas for your clones (Syllable doesnt inteand to be a clone).
Kurt once said style and looks isn’t important now so early in the game. Syllable *looks* like BeOS, but is nothing like BeOS (yes, a lot of similarities, but every new OS built now have similarities with BeOS :-).
The only two projects that plan to clone BeOS’s interface to some extent is B.E. OS and OBOS.
i think everyone should be happy with the progres they made, the only thing we could do is help them with developing or designing some gfx, (and give them tips)
one thing i would love to see is “the dockbert” Idea in the beos Land to syllable
because its straight forward and its elegant to
“The PC is a horrible mish-mash or hardware and standards. I am involved in a little group trying to code a hobby OS for our own use. We come across the problem of drivers so many times. **WHY** can’t this be handled by the BIOS or hardware like it would have been on older machines. **WHY** can’t there be a translation layer in the hardware, so you would send an API call for draw line to the BIOS and it would just happen irrespoctive of the card installed. **WHY** do all the different cards have to have different calls. I see *NO* reason for the current mish-mash of different calls and standards. It may add a couple of pounds to manafacturing costs but would make the whole PC world a much more sensible place and a much more inviting place for people interested in writing hobby software.” – http://homepage.ntlworld.com/daryl.dudey/Rants/rant1.htm
Daryl, I agree with your grips. Hopefully the new Itanium PC specs will fix some of these things, but that won’t help us with IA32 architecture OSes. The most obvious way to address the current driver situation is to write drivers for the most common hardware first and to create a device driver architecture that enables others to quickly implement drivers for other devices.
NeXTstep had the best driver system that I’ve ever seen. It had an object oriented (Objective-C based) driver kit that enabled you to sub class the portions of the driver objects that you needed to specialize. This shaved off at least half of the driver work for people since they could reuse a lot of the code via the kit’s classes, which were comprehensive. I haven’t looked to see if Mac OS X uses this but I suspect that would since it was well implemented.
The NetBSD (http://www.netbsd.org) project has an excellent driver design as well, although it’s not object oriented. It has a nice factoring of devices and busses which enables a driver for a video card to work on the PCI or AGP buses. This saves writing two drivers for one video card. Very nice design.
Daryl, I too love 6502 assembly work. I spent years building 6502 programs for fun and profit. It was a lot of fun, but hard work requiring a lot of instructions to move tiny amounts of bytes around. This doesn’t seem to have changed with the IA64, although you have over 300 registers, programs and data bases are much large now… so it’s all relative.
One of the main, if not the major, technicial barriers to entry for operating system authors, commercial, hobby or otherwise, is the large number of drivers that need to be written to support even a small swath of available devices on the market. You’ve got keyboard, mouse, video, scanners, audio input and output, hard disk (IDE, EIDE, ATA, Serial ATA, SCSI, …), floppy, CD, CDRW, DVD, timers, network cards of all kinds, RAID cards, serial io, parallel io, modems, busses (PCI, AGP, …) and the list goes on. Each of these categories can have a multitude of devices that are sold in the market and that need to be addressed if any serious market penetration is to be achieved. Solving the technicial challenge of the right complement of drivers is one of the key milestones of any mature operating system. A further challenge is the wide range of combinations of motherboards and cards that people actually use. Also the market and technology don’t sit still as Moore’s Law is a blessing and a nightmare at the same time. The tradeoff is always between how many and which drivers to support and test. Testing alone chews up an enourmous amount of time and is very boring and repetive work.
It would be interesting to have a program that people can run that would report the configurations of machines and post them to a central web site. This program would need to run under Windows and Unix systems since they are the more common systems.
It would be very good to pool resources and compile a list of all the devices that need drivers for an OS to make it in todays market. Maybe there are such lists and market demographics already kept and available somewhere.
In undertaking the design of a new operating system that is intended for a serious wide spread auidence the key quesions of driver architecture and which devices to support is of prime importance.
My copy of BeOS, the last made, only works with a limited number of CD-ROM players. This required me to search for over a week to find an “old enough” CD-ROM device that would actually work, and that was just to install the OS. I’ve had similar experience with OPENStep devices. (I still run OPENStep in my production work).
An open driver strategy is likely to be the most important part of any operating system to be open. NeXT was wise in keeping theirs open. This is easier for completely open systems since the whole system is open.
Aside from the technicial challenges of building a new operating system there are the challenges competing Microsoft (97.5% market share in one recent report) and with a whole range of other OSes, such as Linux that have now gained significant corporate support.
Do we really need another “clone” os? Another unix clone? We have plenty of these and they work great and have excellent developer communities.
I put forward that anyone intent upon building a serious OS needs to create something that users really want. Something new and different enough that it’s fresh and addresses real needs. Standards are important, but which standards do you adopt? There are many. How much of the legacy solutions do you support? Support too much and you’re just a “clone” of something else, so why bother (unless you are trying to clone a previously closed source system as Linux did and GNUStep and OpenBeOS are attempting).
Of course another reason to build a working OS is for learning, for fun and just because you can.
Funny and appropriately enough, while writing this a friend called me with questions regarding how to configure his video card for a freshly installed FreeBSD system. XWindows… Arrrggggg…
> NeXTstep had the best driver system that I’ve ever seen. It had an object oriented (Objective-C based) driver kit that enabled you to sub class the portions of the driver objects that you needed to specialize
BeOS can do that too. Actually, yes and no.
Yyes, because the driver model was object-oriented and there were samples. No, because it was written in a procedural language (C), so you had a little bit of work to do. Yes, because you didn’t even need to implement functions which you wanted to keep empty (extremely few people implemented support for select). Overall, if half the driver code is the same across all drivers, I’d say it wasn’t the best possible driver model because specialized features wouldn’t be able to work… ๐
>Do we really need another “clone” os? Another unix clone?
Syllable, AtheOS and even BeOS *ARE NOT* Unix clones. They are somewhat POSIX compliant, but that does not mean they are Unices. They are not.
WindowsNT+ is also POSIX compliant, and that does not make Windows a Unix.
BeOS is a dead OS! Why the heck should opentracker work on syllable or any other remnants of the old and dead BeOS. It is time to innovate and create a newer more modern desktop operating system without the ugly looks, slow speeds, and general overall horrible performance of BeOS. There is already an official openbeos project, and that is where that crap belongs. Keep BeOS and its failures away from projects like atheos, and syllable which hardly need the influence of BeOS to become popular and highly usable.
>Why the heck should opentracker work on syllable or any other remnants
Because OT was and STILL is the best damned file manager out there. It lacks 1-2 features, sure, but compared to ALL the filemanagers I have used, it works the best for the job it has to do. Oh, and its attributed-oriented queries ROCK THE WORLD. Recreating it (because Syllable also supports attributes), it might be more difficult than porting OT.
>without the ugly looks, slow speeds, and general overall horrible performance of BeOS
You have absolutely and utterly, no idea of what the heck you are talking about.
> Keep BeOS and its failures away from projects like atheos
BeOS failed in its business and programming model (mt). It NEVER failed on the user experience. The user experience was and is the most STRONG point of BeOS.
No one said Syllable to take the “bad parts” of the BeOS history. Only the good ones (OT, installer, integration…)
Joe Camel, you must have a hobby in creating flamebait because your statements have no place in reality. If you had used BeOS for more than 5 minutes you would not have posted anything that resembles your posting. If you are going to comment about any OS I highly suggest that you install the OS and use it for a month or so before you judge.
Regards,
Jason VanDerMark
Proud BeOS User
“I will stop using BeOS when they pull the keyboard from my cold dead hands.”
I did use BeOS for more than five minutes, and the user interface was inconsistent, awkward, and highly overrated. As for opentracker, yes it was a good filemanager, but I don’t see what would be so hard about creating a native clone, rather than porting it. Everything BeOS deserves to be buried with the OS, where it belongs. It is time to start fresh and leave the old leggacy baggage behind and improve ideas, and start totally open from the beginning rather than hanging on to old code and old ideas. Anyway, this isn’t to start a flame war, it is to expose a point of how biased people become over using an OS simply because it is the hip thing to do rather than the actual benefit of sticking with something and making it great simply by attempting to do things differently than they had been done in the past. Becoming so damn fanatical about old technology is simply a way to ensure that nothing newer, better, and different would ever become a reality. I say it is time to say good bye to BeOS, and focus on taking a different path with AtheOS and Syllable. And those people who love BeOS so damn much can stick with OpenBeOS.
Just a comment.
I thought this story was on Syllable and not BeOS.
> I did use BeOS for more than five minutes, and the user interface was inconsistent, awkward, and highly overrated.
Haha.. yes… now, you convinced us all.
> It is time to start fresh and leave the old leggacy baggage behind and improve ideas
If BeOS did one thing right, was just that. But it paid this “no legacy/new ideas” thingie by not having programmers who understood how to write complex applications for BeOS.
> old technology
Most of BeOS’ technology has still to be surpassed by any other OS.
I would say QNX surpassed BeOS in more ways than could ever be imagined. BeIA didn’t have a chance against QNX and nor did BeOS. Anyway, end of discussion, stop destroying AtheOS and Syllabus by pushing old BeOS ideas on it. Go to OpenBeOS and destroy that OS.
Yeah, and you have no idea what you are talking about, about any OS. And if I see such trolling around here again, you will get booted.