YellowTAB made a number of announcements regarding additional features to be found on Zeta 1.0 and the future: Java 1.4.2 support, C# support via Mono and support for thousands of scanner models via SANE. Additionally, Bernd Korz, CVO of yellowTAB, sheds his view and knowledge on recent developments involving BeOS and Zeta, and discusses the past and future of computers in general.
have you please an answer
thanks
HH.
I wonder why some people are so “glued” to the kernel thing. It is not just the kernel that it is the same in Zeta as in Dano’s 15 Nov 2001 leaked version, but other libraries and kits as well, from what I observed and/or told.
“In order to succesfully port J2EE v1.4.2 and Mono/portable.net, we had to add functionalities that BeOS never had. Memory management and network are the main black spot of BeOS for such porting, so we had to fix that.”
Although some people knock net_server, I thought it was replaced with BONE. Whats wrong with networking then?
BeOS seems to manage memory well enough for my liking. Again, wondering why they are making this statement.
I do like that we will have Java and C3, though. Java especially is an important feature I find BeOS lacks.
“Zeta is not a dead OS, it can and has to support such platforms instead of being stuck with the lack of OpenOffice, Eclipse and other killer apps.”
I, for one, would rather see native apps being produced for the OS. OpenOffice is ugly and bloated. Feature-rich, yes, but not very fun (more importantly not productive). Why not put effort into polishing AbiWord, which would be much more responsive. After all, that is one of the best things about BeOS – UI responsiveness.
I’m starting to agree more and more with Eugenia in that Zeta is feeling like a Linux distro.
Right now YT’s fourms are flooded with the kernel question. I personally think they dont have the money right now to hire kernel programmers to work on it. Once they get R1 out and get money flowing in, you will probably see a updated kernel.
Sandman
>BeOS seems to manage memory well enough for my liking
Mmm, actually, no. Gimp, Mozilla and WINE’s ports comes to mind…..
>I, for one, would rather see native apps being produced for the OS.
Be tried, it didn’t work. The best bet YTAB has, is to get good ports.
>I thought it was replaced with BONE. Whats wrong with networking then?
BONE was never finished. No one is talking of replacing BONE, but merely enhance it further.
Due to the cagey way he says it, it looks like they don’t have the source after all. Which is bad. Then again, they could figure out some binary patches to fix the memory thing and the thread limit.
Also, could someone ban Arkady’s IP? He’s obviously a troll who doesn’t contribute anything.
What do you think about this speculation that YT will start working on the kernel when they get some money from 1.0?
I don’t even think they have the source to it or maybe even anything at all.
If they claim to be working on it for years, you think they would have done something to some part of the OS.
and another thing bothers me…
The whole idea of Zeta News. Why can’t there just be BeOS new sites like beos journal or iscomputeron ?
The fact that it is run by someone who sells beos software also bothered me, and the no critisism review, and now we find out that they are partnering with YT.
They say Technix is against Zeta, and Zeta News is all for it, so it looks like we can only trust this site and iscomputeron to give us a fair judgement.
These are all great pieces of news. I may finally be able to use Beos/Zeta as my main work OS (currently running RH9). I should be able to run JEdit, Eclipse and several other smaller Java apps with R1 or slightly later. All I need now is an office suite which supports either recent Office2K/XP or Open Office file formats and I’m happy!
I think Zeta has a promising future ahead of it. Enough people and, more importantly, computer manufacturers are fed up with Microsoft and looking for an alternative. Linux just isn’t there for the average consumer/small business desktop and (I say this with a genuine tear in my eye) probably won’t be for several years. Mac OSX is a work of art but they miss out on the all-important white box market. Enter Zeta – a cheap, user-friendly alternative to windows which covers the X86 market.
I personally can ignore the Kernel issue for the moment. This release gives Beos its legs again and gets developers back on track for the future – either Zeta with an improved kernel or OBOS. Yellowtab should definitely stick with the use of open source ports where they can to improve functionality without excessive work. This is the true advantage of open source which Apple appears to have grasped. If Yellowtab can follow their lead and put a user-friendly interface on open source products while simultaneously helping to develop them then they will be on to a true winner.
Keep the home-fires burning Yellowtab/Zeta!
A couple of points about the kernel:
1) BeOS was never that great at memory management. Unlike most modern OSs, it lacks a unified block and page cache. Currently, 1/8 of a machine’s memory is statically reserved for caching the disk. So even if you have tons of free memory, the disk cache might still have to throw away cached data because it hit its 1/8 limit. Heavy I/O never performed very well under BeOS for this reason. Its VM algorithms were also never terribly advanced.
2) BONE was pretty good, but it was never completed. They never had a commercial release with BONE, so its bound to be less polished than other aspects of the OS.
The comments about improving VM and networking make me curious. In Dano, VM and networking were in the kernel. How do they plan to improve it if they don’t have the kernel source? I remember some comments made in a BeNewsletter once that the BeOS code was very difficult to build. Might it be that YTab has the code, but hasn’t been able to devote the resources to figuring out the build system and compiling new binaries?
I don’t want to sound like a wowser but on their site they state that they have ported J2EE 1.4.2 to Zeta, which isn’t possible as J2EE hasn’t even come out of beta testing yet. I am assuming, however, they’re referring to J2SE 1.4.2 π
Rayiner Hashem (IP: —.res.gatech.edu) – Posted on 2003-10-13 00:40:35
A couple of points about the kernel:
1) BeOS was never that great at memory management. Unlike most modern OSs, it lacks a unified block and page cache. Currently, 1/8 of a machine’s memory is statically reserved for caching the disk. So even if you have tons of free memory, the disk cache might still have to throw away cached data because it hit its 1/8 limit. Heavy I/O never performed very well under BeOS for this reason. Its VM algorithms were also never terribly advanced.
A while back I suggested that the best thing Zeta could do would be to adopt the FreeBSD 5.x kernel, which has a larger mindshare, port the API’s from BeOS and base Zeta on that.
Compared with the amount of work that is required for Zeta, FreeBSD 5.x would be alot easier to complete and bring full POSIX compliancy with it as well.
>>BeOS seems to manage memory well enough for my liking
>Mmm, actually, no. Gimp, Mozilla and WINE’s ports comes to mind…..
Wha? The GIMP is very fast on Xbeos, I don’t think that is such a great example, but yes there is room for improvement it’s obvious I agree with you there. Particularly Mozilla/etc. the WWW browsers are pathetic, but most likely the design of the port itself is the cause.
In regards to the this bit of news:
“…new libraries and tools appeared, for example:
glib2, pk-config,libgettext, libintl and libiconv…”
New?! uh, NO. ALL of those can be found at bebits, and were originally ported by members of the BeOS community. These are not new ports, they are not ‘originals become from YT engineering’. This is exactly the kind of thing that makes them not just look innacurate and misrepresented, but also IT IS OFFENSIVE TO THE DEVELOPERS WHO PORTED THESE IN THE FIRST PLACE.
Basically, what this says is ‘we don’t need you, I bite my thumb at thee’ when that is the last thing you need to aim for when trying to build an enterprise.
Maybe I will cut them slack, but I don’t get how the webmaster or whomever authorizes press could so obviously post such incorrect and poorly worded material.
If you want to show what you are doing and promote yourselves, you don’t have to lie to do it If you made the packages work better or more natively integrated, then why not just say so? Give credit where credit is due, that press is just plain false.
Hey YellowTab: don’t piss on independent developers. TYIA
>Compared with the amount of work that is required for Zeta, FreeBSD 5.x would be alot easier to complete
Please don’t talk complete crap over here CooCooCaChoo.
A while back I suggested that the best thing Zeta could do would be to adopt the FreeBSD 5.x kernel, which has a larger mindshare, port the API’s from BeOS and base Zeta on that.
Compared with the amount of work that is required for Zeta, FreeBSD 5.x would be alot easier to complete and bring full POSIX compliancy with it as well.
Why would one want to make Zeta Posix compatible? The only feature (not a benefit) is that we can port more applications over from Linux/Unix. BeOS is about a superior user experience and currently Unix/Linux apps are below Windows apps in ease of use and are quite rough around the edges.
BeOS DOES need Java! If only for OpenOffice. People need to manipulate docs in the evil MS Office formats. OpenOffice is not perfect but there is currently nothing better at opening .doc, .xls, other than it (except for MS Office of course).
>Why would one want to make Zeta Posix compatible?
Every OS needs POSIX in order to get ports. Even NT is POSIX. Zeta NEEDS applications, so POSIX _is_ important.
However, the above claim of using freebsd/linux/whatever kernel is frivolous. The whole point of Zeta is to be unique, as BeOS was. Plus, the amount of engineering required to make it binary compatible with the old BeOS apps is 10x fold the amount the Zeta kernel needs fixing.
“Right now YT’s fourms are flooded with the kernel question. I personally think they dont have the money right now to hire kernel programmers to work on it. Once they get R1 out and get money flowing in, you will probably see a updated kernel”
I think there are a lot of factors at play here. Just let me say i don’t think it is fair to assume that this is the leaked version of dano sans update, which has become the defactor assumption. None of us ultimately know what the agreement between palm and yellowtab entails. none of us know what palm is doing with the remnants of beos. none of us know how much overlap there is between palm os 6 and beos.
That is my position. My understanding is that yellowtab had to negotiate with palm to get this done. regardless, you don’t know. Palm is not stupid. there may be a chance down the line to push beos or palm os 6 onto the mainstream desktop (x86, powerpc, even ARM) market. Palmsource has some bright people there and they gain nothing from allowing yellowtab to go off and add changes to the core of the beos/zeta that might one day compromise a migration of palm os to laptops (which they have already announced) or to other things.
Yellowtab is a godsend to palm. they are going out there and seeding a market. That means developing applications for a desktop (large screen) and gaining driver support. Palm would be positively stupid, and i don’t think they are, if they didn’t set themselves up to benefit from this if it succeeds. That again means that palm may put limits on the changes that they allow yellowtab to put in it, though palm might itself be adding some changes to components that are in beos and palm os 6.
regardless we just don’t know so i see no point in assuming that yellowtab spent money on an os they can’t upgrade. I believe bernd when he says they will be able to evolve the os. I just can’t see how they’d spend money without being able to do that or migrate to openbeos, something they say they won’t do.
“currently Unix/Linux apps are below Windows apps in ease of use and are quite rough around the edges.”
If you’d said some I could go with you on that, if you said most I could at least see your point, but labeling every single Linux application as unpolished and not user friendly is foolish.
But as to your question “The only feature (not a benefit) is that we can port more applications over from Linux/Unix.”. Because it’s easier to build a nice GUI over an existing application than it is to write a program from scratch, then design, then write a nice GUI as well?
Wasn’t one of BeOS’s strong points the large degree of POSIX support? So why do we need a Unix kernel?
>Why would one want to make Zeta Posix compatible?
Every OS needs POSIX in order to get ports. Even NT is POSIX. Zeta NEEDS applications, so POSIX _is_ important.
However, the above claim of using freebsd/linux/whatever kernel is frivolous. The whole point of Zeta is to be unique, as BeOS was. Plus, the amount of engineering required to make it binary compatible with the old BeOS apps is 10x fold the amount the Zeta kernel needs fixing.
Not to be rude or anything but why are insisting on beating that dead horse and re-inventing the wheel again? Zeta/BeOS strength was never in the kernel but in the API kits that surround the kernel.
The API kits was the most power feature of it. Bringing it to a FreeBSD kernel would enabled the coders to have all the POSIX features at their disposal as well a the rich BSD API as well.
No one said they had to use the X server or any other UNIX’y thing. Just the kernel and build upon it.
If the continue down the same path, they’ll doomed to the same fait as Be Inc. The world is split betweeen Windows and UNIX-Compatibles. Apple have realised this and that is why they chose NeXT over Be Inc. Programmers aren’t going to go out have have to flog another dead horse just to get their application running. Programmers want minimum fuss and bother when porting, the only people who are anti-UNIX and anti-Legacy are the techno-illiterate end users who think that anything that is different is inheriently better, which is patently false at every level.
Wasn’t one of BeOS’s strong points the large degree of POSIX support? So why do we need a Unix kernel?
The amount of POSIX support is so low, it is almost regarded as pathetic. I remember Mozilla trying to get ported and the number of basic POSIX and other features that were missing. It too Be Inc almost 6months to a year to come up with a work around for one problem only to result is crappy performance.
Some people just love to re-invent the wheel, I say, get a good based and build something better on it.
> BeOS DOES need Java! If only for OpenOffice.
Uhm, OpenOffice is written in C++, not Java!!
OOo has some parts that work only with Java. But Java is not required, it is optional.
From what I remember, BeOS was quite POSIX compatible, it was just missing two critical things that large apps depended on (mmap() and sockets as file descriptors). And POSIX compatibility is important. Its a lowest-common-denominator API that most OSs support. Even if you think that UNIX software is unpolished (which I contest) you can’t deny that there is tons of important UI-less UNIX software that Zeta could use. For example, there are parsers like libxml2, or support libraries like glib, or apps like Apache that can use POSIX to be very portable.
PS> Eugenia, NT’s POSIX support is really not production-quality. I have to maintain some software simultaniously on NT and Linux, and Cygwin is my best friend.
OOo has some parts that work only with Java. But Java is not required, it is optional.
Java is required to build OpenOffice.org from source.
Compared to when R5 came out, this time the OS has the ability to play DVD’s, will ship with a form of Java, has better hardware support, has a decent browser (Firebird BONE), is easier to port to (BONE and other libs needing sockets/file descriptors). Plus a newer Media Kit, improved messaging, themes, old kits improved, newer kits, this looks to be the best BeOS release yet.
Hopefully yellowTab will generate the finances it needs to further push the OS. Even if they dont have access to kernel code, if Zeta generates enough sales, they might even earn enough to buy the sources from Palm, or even inspire Palm to continue developing BeOS (yeah, right…). Big if.
So, if you want to see a form of BeOS survive, evolve and grow, I guess you’ll have to bite the bullet and buy Zeta. Geez, never thought that I’ll end up promoting a commercial product…
“Zeta is our only hope”.
“No, there is another…”
Go Zeta! Go!
Java is a great start! Let’s bring Gobe, Opera and so on back to BeOS again!
Join the BOIND Campaign
(BeOS Is Not Dead)
π
Good news is that yT is actually doing something, however the Bad news is that they decided to walk their on path.
BeUnited has been working on the port of Java together with Sun (yes Sun) for some time now, and wouldn’t that have been a wise way to work together about something?
Assuming that you can actually buy Zeta now after years of development on Apps (but not the core OS????) is this something worth its money? Some claim it is, others don’t.
I’d feel a lot more encouraged to buy something if the company I bought something from said “this is on our priority list, but we have frankly not just had time”. This comes to mind regarding Real Audio for instance.
Java is good to see on BeOS, but it’d be worth 50β¬ if it actually was done WITH BeUnited rather than against BeUnited.
The whole point of Zeta is to be unique, as BeOS was.
I don’t mean to sound disrespectful or anything, but… Isn’t the whole point of an OS to be good? I mean, BeOS & Co. must be the only OS family where people tend to put ‘uniqueness’ as a most important feature.
Well,
I think I have just decided to go ahead and purchase zeta. Probably deluxe edition in fact. Being able to use java/eclipse is definately a plus since i use ’em often.
One thing that really bugs me (according to their site), if I buy rc1, I have to pay $10 more for the finished release later. I suppose they don’t want to lose money on shipping and manufacturing costs… So I will have to think about this. Probably, I will just wait.
With bone and it’s much more complete and accurate implementation of the BSD sockets api, I’ll be able to actually use SSH2.
BeUnited has been working on the port of Java together with Sun (yes Sun) for some time now, and wouldn’t that have been a wise way to work together about something?
Well, I guess that yT are updating the version of Java BeInc implemented with the help of Sun. No disrespect to BeUnited, but I kind of expect BeInc’s version of Java to be more inline with the OS than anything which BeUnited can dish out (but I have been proven wrong before). Since yT are continuing with the Java port, I guess they must have access to the source code of Java at least, and since they’re under licence with Palm/BeInc, they cannot reveal the source to 3rd parties, which is what BeUnited are. So give yT a break when it comes to Java and the BeUnited duplicated effort.
yT have improved their game quite a bit since first public appearance. Even though they’re still fumbling, they are progressing, so by Zeta R2 timeframe they should get their act together and be a decent OS vendor. I only want to see BeOS live (and advance), so I guess I’ll have to support the only viable game in town (until OBOS is ready).
i always loved beos. recently installed it on an old laptop – it was even faster than my 1600xp machine. i think i’m gonna buy zeta – especially since http://pixel32.box.sk/ ‘ll probably run too.
“Memory management and network are the main black spot of BeOS for such porting, so we had to fix that.”
How are they gonna change the memory management if they don’t have the kernel source ?
“I don’t mean to sound disrespectful or anything, but… Isn’t the whole point of an OS to be good? I mean, BeOS & Co. must be the only OS family where people tend to put ‘uniqueness’ as a most important feature.”
Of course an OS has to be good, but having uniqueness is what can make an OS actually stand out. I mean, what do you think Apple’s marketshare would look like if they didn’t have a unique UI and unique design (for all their products, from notebooks to mp3-players)?
Umm, all you guys talking about wanting to use Eclipse (a wonderful app) do realize that this would require a port of SWT, right? Even the Mac barely has a working port of SWT, and it’s mediocre at best. Even with Java, you probably won’t be seeing Eclipse on the BeOS for a long time.
After the announced partnership of Yellowtab & Xentronix some people are in doubt of the credibility of ZetaNews since Frans van Nispen was working for Xentronix as well as ZetaNews.com.
However the reason that ZetaNews.com hired me is because Frans van Nispen will now focus on his work at Xentronix.
Joyce and myself will do all the news and reviews at ZetaNews.com
Right now we are in negotiation with another editor / reviewer. He or she will be soon announced on ZetaNews.
comments or questions?
feel free to contact me via e-mail or leave a message in the forum here.
Kernel source:
We don’t know shit about the kernel source, everything is speculation. So why say that it’s impossible for yT to change the memory management when you don’t know anything about the sources?
Uniqueness:
As someone said earlier, what gives apple the marketshare they have? Their uniqueness… they have something special. How else can BeOS take a position in the market when there is competition from Windows, Apple and Linux???
BeOS has a unique GUI and a unique user experience I haven’t found anywhere else, that is what is worth keeping. And that is why I agree that the more native software there is for BeOS the better, but in the absence of that I am glad that ports of OOo will become more viable, and that Mozilla stuff is available is good too… And when it comes to development tools like Java (and C#) they’re a necessity if BeOS ever gonna have a chance…
Of course an OS has to be good, but having uniqueness is what can make an OS actually stand out.[…]
That’s not what I meant. I’m perfectly aware of that.
What I meant was that Eugenia seemed to be saying something like “Kernels? Who cares about the kernel?? The point is being unique!”. And that is something I’ve heard too many times about BeOS. “Ok, it doesn’t have [this or that feature – you name it], but BeOS is unique. That’s what matters.”
You can have truly exclusive and unique crap. Being unique doesn’t make it less crap. Being unique doesn’t magically fill all voids. (Note: I’m not implying BeOS is crap, ok?)
I may or may not agree with the original poster about the benefits of a *BSD kernel, but I think that it is a valid suggestion. Simply that.
but kernal source is welcome to know. i have think that important answer that is and if is knowledge for the developers of the zeta, it is good thing else bad because the most important thing the kernal is. or isn’t it? I havewritten email to the developers of the zeta (they speaking my German language) and if they send reply, I translate and post here for your reading.
bye for now and standby maybe exiting news in the future!
In other words, yT still doesn’t answer any of the critical questions which were raised and rather hides behind a smokescreen. Thanks, that does it for me, especially after the nice little incident with pWarp. yT is not to be trusted in my opinion. Hope noone comes running and cries about having wasted money on Zeta. If they ever manage to release a product that is.
Troll? no. Just my observation of the current situation.
You can have truly exclusive and unique crap. Being unique doesn’t make it less crap. Being unique doesn’t magically fill all voids. (Note: I’m not implying BeOS is crap, ok?)
I may or may not agree with the original poster about the benefits of a *BSD kernel, but I think that it is a valid suggestion. Simply that.
When I was suggesting the *BSD kernel, I mean the whole *BSD base that makes up the iso-mini distribution. From there, build the BeOS API’s on top, like how Apple has provided the BSD user space ontop of the Mach kernel.
You have the feature complete API from BSD, which opens up a HUGE box of available software, then add on the BeOS layer which enables one to run most available BeOS binaries.
The fact I was point making is that Zeta is not a organisation to spur programmers creative juices, it is there to turn a profit for the investors. Be Inc. let down their investers and if people with ideas keep coming up flawed plans, investers are going to be less and less trusting in people who may come up with any ideas in the future, even if it does have a sound business plan.
Uniqueness doesn’t sell, sexy, cool, funky and buzz word compliancy, that is what sells. People want low cost, easy to use and feature rich. Provide that trifecta and you’ll get the people.
I think you underestimate the magnitude of what you’re suggesting. The amount of changes needed to be able to mesh BeOS’s kits/userspace with FreeBSD’s kernel would be a huge task, that’d take months if not years even for very experienced devs.
“In other words, yT still doesn’t answer any of the critical questions which were raised and rather hides behind a smokescreen. Thanks, that does it for me, especially after the nice little incident with pWarp. yT is not to be trusted in my opinion. Hope noone comes running and cries about having wasted money on Zeta. If they ever manage to release a product that is.
Troll? no. Just my observation of the current situation.”
you really don’t know the terms of yellowtabs relationship with palm or terms of a potential NDA. He may not be able to tell you that without breaking an NDA. Stop speculating.
I’ll believe it when I see it. I have been hearing about Java coming to BeOS for years, but it never happens. I’m not talking about personal java, but the J2SDK. Java is a difficult undertaking to say the least, and nothing that YellowTab has done leads me to believe that they have the ability to do this anytime soon.
-G
“you really don’t know the terms of yellowtabs relationship with palm or terms of a potential NDA. He may not be able to tell you that without breaking an NDA. Stop speculating.”
If that is the case, then why would I buy a product from a company that is not allowed to reveal SIGNIFICANT details about the product I am purchasing?
-G
especially since http://pixel32.box.sk/ ‘ll probably run too.
Is Pixel32 usable nowadays? …last version on their download page definately isn’t. Has Pavel upgraded the Pascal port? (this was the source to much of the instability of pixel32 in BeOS IIRC)
Personally I thinks its great that YellowTab is out here.
Any improvements are welcomed. I am not a programmer
an dont claim to be one it is good that beos users have a chance to get a improved version of beos. Lets face you cant expect them to give it away for free……….
CooCooCaChoo are u are afraid to identify yourself on
you seem to make statements when you dont know what your talking about. Unless you have a programming degree. Get your facts straight before you speak.
By the way when i post i always leave a valid email address
Eugenia keep up the good work……..
I think you underestimate the magnitude of what you’re suggesting. The amount of changes needed to be able to mesh BeOS’s kits/userspace with FreeBSD’s kernel would be a huge task, that’d take months if not years even for very experienced devs.
How long has Zeta been in development for? how much effort has been put into flogging a dead horse in regards to creating new device API’s, new POSIX api’s, porting and sprucing up existing ports? how many of these device API’s already exist in FreeBSD? every last one that was created to get Zeta on par with other operating system was available already, fully stress tested and optimised by the FreeBSD community.
How about a x86-64 port? when is it going to be done? How long are we to wait for that to appear? will it ever appear? Again, had they embraced the FreeBSD core, we would have Zeta rolling around the corner on the coat tails of FreeBSd 5.3, with x86-64 and i386 support, out of the box. All Yellow Tab would have to do is spruce up the various kits that would ride upon FreeBSD.
Smokescreen. Totally. Zeta is just refusing to answer the questions because if they truthfully do so, a significant amount of people won’t purchase Zeta. I certainly won’t based on what they’ve said. They don’t have to reveal the details of what they have specifically, all they have to do is tell us whether or not they personally have the ability to modify and fix any bugs found in the Zeta kernel. That’s all. If they say they can’t answer that because of “legal obligations” they’re full of bullcrap.
btw, I don’t check it because I get Windows users spamming me with virus’s claiming to be “Microsoft security” updates. I *WOULD* put my real address but I don’t want the crap spammed out of it.
Also, why should I list my new one when you could very well sign my new email address up to 100s of “free sex mail” services just to get back at me in the most juvenile fashion.
“you really don’t know the terms of yellowtabs relationship with palm or terms of a potential NDA. He may not be able to tell you that without breaking an NDA. Stop speculating. ”
And maybe they have a Windows XP Kernel powered by 100 gerbils in a treadmill. As a matter of fact, there are numerous important things I’d like to know to be able to trust them and to know if the product is worth my money and have to determine for myself if I think that the product has a future.
If they can’t provide that information, the reason doesn’t matter, I don’t give them my money.
if you live in .de you should know austria is .at, .au is australia.
How long has Zeta been in development for? how much effort has been put into flogging a dead horse in regards to creating new device API’s, new POSIX api’s, porting and sprucing up existing ports? how many of these device API’s already exist in FreeBSD? every last one that was created to get Zeta on par with other operating system was available already, fully stress tested and optimised by the FreeBSD community.
Try learning a bit more about what you’re talking about. You’ve mentioned on another comments thread that you aren’t actually a programmer, and it shows. You can *NOT* just slap on another OS’s kernel and get a free ride. It’d take massive rewrites of either the userspace or the kernel, especially considering BeOS isn’t even vaguely architecturally similar to FreeBSD. This wouldn’t be a trivial undertaking even if the OSes were both fully POSIX compliant UNIXes, and thus far more similar in overall system design than BeOS is to such an OS. So please, get a bit more informed before spouting this kind of drivel.
“As a matter of fact, there are numerous important things I’d like to know to be able to trust them and to know if the product is worth my money and have to determine for myself if I think that the product has a future.
If they can’t provide that information, the reason doesn’t matter, I don’t give them my money.”
wow. Harsh. Those are pretty tough standards. There are somethings corporations simply can not say. I’d base my purchasing decision on the need of the pleasure it brings. There is a conclusion to yellowtab anyway you cut it. It is called OBOS. I hope yellowtabl sticks around but if they don’t well you’ll have an open source alternative.
heh.. whether they hexed the date or not, who cares. It’s STILL a new kernel for all of those who only have R5. Yeah it might be 2 years old, so is XP’s kernel. blablabla. I can’t believe people are wasting so much energy arguing about this, it’s funny.
Try learning a bit more about what you’re talking about. You’ve mentioned on another comments thread that you aren’t actually a programmer, and it shows.
True, I don’t programme in C++/C, but I do programme in Java hence as I said previous, I don’t know the exact specifics of C++/C. Sure, I know them at the most trivial level and given a few months, I could improve upon what I learnt in my C++ classes when I was at university.
You can *NOT* just slap on another OS’s kernel and get a free ride. It’d take massive rewrites of either the userspace or the kernel, especially considering BeOS isn’t even vaguely architecturally similar to FreeBSD.
IIRC, the kits that made up the BeOS API were designed to be portable from the ground up and are 100% C++.
This wouldn’t be a trivial undertaking even if the OSes were both fully POSIX compliant UNIXes, and thus far more similar in overall system design than BeOS is to such an OS. So please, get a bit more informed before spouting this kind of drivel.
How about learning a little more politeness or are you just another cockney with a bloody huge chip on your sholder? how about this sunshine, leave the computer alone, join the Barmy Army and do a pub crawl. Atleast then your brain cells would be getting used for something more constructive that getting on my case about an idea I was floating.
We have a specific set of methods used to accomplish our goal of makeing what should amount to BeOS R5.5’s equal as Be planned it, except with important technologies fully supported to open up our market.
The extra items of support are intended to allow the user to run more programs, more reliably than ever. And at the same time, our integrating of the API in so many of these ports will make it nealy impossible to tell the difference between one type of an app or the other. Which is good of course! A java app will work, look, and act (to the user) like a native Zeta application.
Database management tools and connectivity open up the non-M$ controlled sectors. Ever go into HEB? We want Zeta to run on those machines one day. We must be able to talk to their existing network, use their existing framework and be able to run their applications, or at least be able to easily duplicate / port them.
The disk I/O cache and page cache is modifiable in Zeta, thank you :-). I am currently experimentin with different configurations and complex algorithms to discover the best performance for different situations. Developing?? You need performance for tons of small file operations. Web Server? You need flexibility, small files, and large files. Database Server? You need a system that monitors what you need and adapts. π Zeta already is capable of all but the last. The last is a task of the databasing management software being used (just tell a special library ZHDDManagement->SetAverageFileSizeUsed(int32 kbytes / int32 mbytes / int32 gbytes); And ye, GBs are a valid need for a FS with support for MASSIVE files.
Hmm… I am getting near .5 TB now on my Zeta machine… π And it is nothing special.. just 4 HDDs on IDE ATA 133 RAID π I can’t wait until I can afford to test out Serial ATA… Gonna be fun π
The kernel does not need to be completely rebuilt. It is modularized to the point internally that we can allow interchangeable (run-time) VMs should we so desire.. and never even need to touch kernel code (if we have it). I give no answers in the regards of what sources we do and don’t have… I don’t know the answers, I am just a petty little developer who has been very busy with life, and out of the circle a little lately.
As for y’al saying that Zeta is sounding like a Linux distro.. yes, it sorta is sounding that way. But it will not be that way. We polish our apps before a real release. This is why we have blocked so many release dates before, we find a problem, we fix it. However, we don’t have the Be luxury: we can’t just ommit a feature because it isn’t polished yet and release… We have to prove ourselves, to the world.
Oh well, that is enough of my rant on the one page of this stuff that I read… π
–The loon
Hey Looncraz,
In what way is the ZHDDManagement class connected to the rest of the system? Is it a high-level interface configuration for your ODBC library, or does it actually change any parameters of the FS sub-system?
Regards,
— tic
I found the report by Bernd to be quite encouraging. While all the rest of you nay-sayers are Doubting Thomases, people like me can see the validity in what what said.
An NDA (Non-Disclosure Agreement) is designed to make sure you keep your trap shut about certain things someone (company or otherwise) doesn’t want the general public (or their competitors) to know about.
As long as they provide what we want (updated BeOS and more 3rd party support resulting from it’s revival), who da (!!!) cares whether or not they can disclose every detail YOU think is important?
Y’all didn’t pony up the big bucks to make it happen. YellowTab did. They obviously got SOMETHING accomplished for their money, because stuff is being added and they’re showing Zeta at places/events.
If you’re gonna go jump on someone’s backside and start knawing, why not go over to OpenBeOS.org? Look at the difference:
Zeta = real tangible product, nearly available
OpenBeOS = potential product, still a long way off (if ever)
Before Yt came on the scene with Zeta (or before Zeta was seen as *authentic* BeOS, with a different name), OpenBeOS was our only real hope. But that “hope” is seriously overshadowed by the efforts of Yt and Zeta. Why have 2nd best when you can have “The One”?
Now, there is still a chance Yt could go B.U.D (Belly Up Dead) on us. And, thus, OpenBeOS should be kept going, if not at least in a holding pattern until such time as it is seen as truely needed, should Zeta fail.
But, then again, in a realm run by MacOS X 10.2 and beyond (yea!), and Windows XP (X=non-existant and P=Privacy; you knew that XP is Microsoft’s SpywareOS, right?), the likes of Zeta and OpenBeOS really don’t have much room to flex what muscle they have, if any remains. π
I won’t buy Zeta just to have BeOS again. I’ve moved onto MacOS X 10.2.8 on my “new” G3 B&W/350. But, should Zeta stir up a bunch of 3rd Party support and it looks like Zeta is truely reviving BeOS in a whole new way, THEN I might reconsider. Until then, I’m staying with an OS that is sooooo here and now!
Latre!
Luposian
As a former BeOS User (still have BeOS installed on an AMD K6-500mhz machine, but haven’t booted it in ages…)…
All I want to run BeOS again as a daily driver is:
– A Modern Web Browser (A port of Netscape Phoenix would be fine)
– A Modern Mail App (A port of Thunderbird would be fine)
– An office Suite (I own GoBe Productive, and it’s good enough)
– A Checkbook program (Even a copy of Quicken for DOS on Bochs would be fine… Or Quicken for Windows running on a BeOS port of WINE)
– Drivers for my new Printer and scanner
– Drivers for newer hardware
– Things like Flash and Shockwave so I can go to new sites.
I need my OS to be transparent to me. It needs to do what I need to do.
It doesn’t need to be different.
It does need to work, and work fast.
Al
Something can be 100% C++ and still not be portable. For example, most of the kits use the underlying BeOS messaging API — you would have to replicate that in the FreeBSD kernel to even have a hope of porting the kits. Also, consider Apple. The major differences between the OS X and NeXT kernels are just an updated Mach, a move from 4.2BSD to 4.4BSD, and some NetBSD kernel components. Also, the OpenStep APIs were already designed to be portable, and had been ported to many UNIX OSs. Even with all that, it took Apple a long time to get OS X working, and they’re developer team is much larger than YTab’s. It just wouldn’t be feasible for YTab to do something like that.
i agree with everything in the post except the references to openbeos. I don’t see openbeos as secondary to zeta. to me zeta will keep us going until openbeos comes along, something i hope yellowtab will eventually support.
It is a matter of open source vs. proprietary for me. I just don’t trust proprietary OSs that much any more since they tend go belly up unless they are MS and i think the same goes for developers. My guess is developers would prefer to code for open source all else being equal. All else is not equal though (user base) so coders code for MS and apple.
Open source to me is about keeping the OS alive regardless to what happens to any individual company.
Yellow Tab have not chosen an easy path to success.
What the odd BEOS enthusiast thinks wont matter much after version 1. Right now the so called ‘BEOS community’ is a handy place for YT to scrape up a bit of revenue.
If every active BEOS user upgrades to Zeta it wont make YT a huge success, they need the corporate and educational markets. Hence .NET and Java are important. OpenGL unimportant. Perhaps YT will announce a CITRIX CLIENT for Zeta soon?
IT departments are looking for a way to reduce the burden of patching desktop PC’s every month.
Companies need something (anything) that works quickly on old machines, has lots of drivers, has a decent network stack, and runs CITRIX CLIENT without running WORMS. Something Fast, cheap and simple to support that looks a bit like windows. Perhaps this is where YT are going.
Sounds as if you are talking about linux. Look, it’s the same again as some years ago with Be Inc.: a small company with a too small budget hyping a product. Only that there are alternatives which are either further developed, have more hard and software support, are cheaper or fomr Microsoft (or any combination of the aforementioned examples). Be Inc. failed. Now tell me how a company which builds on top of a failed product, has less consumer backing and is more secretive than Be Inc. was while noone even knows if they have any sourcecode at all to make some serious decelopment should be able to make any impact in the current OS landscape? Hell, even SkyOS which is mostly a one-man-show seems to move faster than Zeta.
IIRC, the kits that made up the BeOS API were designed to be portable from the ground up and are 100% C++.
Ok, you’re somewhat confused here. There are two different kinds of “portable”: 1) can be recompiled on a different CPU architecture with no changes, or 2) can be recompiled on a different OS platform with no changes. BeOS’s kits do indeed confirm to #1, but #2, not even remotely. They’re very heavily tied to BeOS’s message passing infrastructure, as well as to various other architectural differences BeOS has when compared to a UNIX system. End result, no, moving to another kernel would be at least as much work if not more as compared to fixing the existing one.