Home | Forums | Submit   Haiku Generator | Quotable JLG | The Icon Tarot 


A Thought About A Path To An Open BeOS

Filed under the:  department.
Posted by:Deej on Wednesday, 09 May, 2001 @ 12:45 AM
 
Submitted News

I’m putting this under the “submitted news” category, as I guess I’m submitting it. I’m bored tonight, and have been reading so much arguments for/against the whole OpenSource BeOS issue. Anyways, not trying to start yet another flame war… we all know the issues behind it, and that it’s not going to happen. Just a bored thought follows, read it if you like, but please don’t flame me about it. Long day… :P

What if… ah, the everlasting “what if” scenario… what if it was an undoubted fact that BeOS was to cease to exist. And another “what if”… what if Be signed on a semi-large group of dedicated BeOS developers, guys that have been around the trenches for a while. Put them under NDAs, real heavy, real strict NDAs that would reach far beyond Be, Inc’s demise. Hand them the code to BeOS, and let them strip out the licensed stuff. No charge to Be, Inc. or it’s shareholders. These devs could work on getting rid of the licensed code, and put together a detailed requirement sheet of what would be needed to get BeOS operational again, in an open source version.

Obviously, this depends on all those licensed parts and the wording of such. It also depends on whether BeIA is still a viable product for Be, as they couldn’t plan to open source their core business. There are many other dependencies, and whether Be, Inc. would even consider it a possibility, being the “leaked” stuff already. Not to mention it would be a ton of hard work for those devs working for free, I sure wouldn’t voluteer for the job. :P

But like I said, a bored thought. Take it, leave it, but, by all means, don’t flame me for it. Just me ranting about a mindstorm a bit. Anyways, cheers, and goodnight all. :)



22 Responses to “A Thought About A Path To An Open BeOS”

  1. KAMiKAZOW Says:

    Do it like AROS

    Why not going the same way as AROS? [url]http://www.aros.org/[/url]
    An OpenSource BeOS-Clone…
    And call it B-ROS, BeOZ (or something like that)…

  2. Anonymous Says:

    Why Open Source?

    I don’t get the obsession with open sourcing BeOS. If you’re good enough to know what to do with the BeOS source code, write your own operating system and open source it. What if Be spun off BeOS into a separate company whose primary focus is BeOS and nothing else? It doesn’t seem to be important for BeOS to stay synchronized with BeIA, so there’s no need for one company to handle both. Create a BeOS company that can look at the current market and tailor an operating system to fit that market (for a change). What I mean by that is, there is a class of hardware that this OS should run on. That’s the first priority. Then, there’s a huge library of software out there that people will want to use (that software being Win32 based), and the OS will have to allow people run it. That’s the second priority. After that, they can promote their own native API, write their own apps, design their own hardware, whatever they want to do. But they need to kick ass on fronts other than just engineering.

  3. chrish_work Says:

    One problem…

    One flaw to this plan… there isn’t a “semi-large group of dedicated BeOS developers” anymore.

    Another problem is that working on operating systems is “hard”, and “most” developers are much more interested in working on [i]applications[/i] instead of libraries and OS foundations. You don’t want a bunch of “Learn C in 21 Days” people romping through the code (note: not to disparage anyone who learnt C via one of those books, I’m suggesting that writing OS-level code is a very difficult skill to learn and master… it’s not something you can pick up in your “spare time”).

    - chrish

    [Edited at 9:35 May 9 2001 by chrish_work]

  4. Deej Says:

    In Response To chrish_work @ 05/09/2001 09:32:24 AM

    Re: One problem…

    Yeah, thought of those points too. But good points nonetheless, Chris.

    Like I said, it was just some late night thoughts, that probably shouldn’t have been posted on the front page of BeGroovy, but I did. :P

    Anyhoos, to counter that… If Be was to meet a nasty demise, who’s to say many of the pro-EX-BeOS engineers wouldn’t join in and help on this idea. But it is super-very-not-a-chance unlikely to be sure.

  5. Anonymous Says:

    In Response To chrish_work @ 05/09/2001 09:32:24 AM

    Re: One problem…

    >it’s not something you can pick up in your “spare >time”.

    Heck, I thought all great programmers picked up their programming skills in their spare time.

  6. Anonymous Says:

    Wait a second…

    Hey, if you’re going to be hiring a corps of free engineers to open source the BeOS, why not hire a corps of free engineers to fix the media kit, make BONE and OpenGL work, write kick-ass inhouse apps for the BeOS? Hell, if I knew enough C (and C) I’d do it.

  7. Anonymous Says:

    In Response To Anonymous @ 05/09/2001 09:00:42 AM

    Re: Why Open Source?

    The reason that Open Source is the best solution right now is because it is BeOS’s only hope. Hear me out. The only reason Linux is surviving is because it is not dependent upon the economy. Nerds program for it and it gets better. Be Inc., which is dependant upon its (slimming) bottom line has control of the core of BeOS. Because Be Inc. has (allegedly) frozen development on BeOS, it’s core is frozen. The only way for the BeOS core to advance is for:
    1) Be Inc. to resume development on it
    2) it to be released to a group that is capable of continuing development.

    Solution one is not a current possibility for Be Inc. If Be Inc. can keep it’s head above water for a while in a new market, I think we will see BeOS work continued (albeit far less than that of BeIA).
    Solution two is very possible. It could be released to another company but (because of Micro$haft) that would enter the same bind. The only smart solution is to put the core (the parts that Be Inc. can) into the hands of those not bound by a bottom line.

    How could this happen? I don’t know. I’m not a business major. For all you anti-OSS people out there, I hope this has shed some light on why we OSS guys are screaming OPEN-SOURCE IT!

    I recognize that Open-sourcing it is a difficult route both legally and developmentally, but we OSS guys see it as the best option (except of course an economic miracle for Be Inc.). =)

  8. Anonymous Says:

    In Response To chrish_work @ 05/09/2001 09:32:24 AM

    Re: One problem…

    I’d do it… if I knew C . I think you underestimate the love people have for BeOS. The engineers were the kinds of people who stayed at work for 10 hours a day, some slept in cubicles (read the interviews in the BeOS Bible). I don’t see it outside the realm of possibility that a few of them would see BeOS.Sourceforge.net and say “Woah! I wrote that part.. lemme see here…”

    As for it being a lot of work to take those licensed parts out, check out AtheOS. One guy wrote it. You think a small group of dedicated people couldn’t do it for BeOS?

    If Be doesn’t survive (which I really hope they do, because they did a great job on BeOS when it was still theirs), I hope it DOES goes OSS… then at least there’s a freakin chance of it ever being updated.

    -SirNickity

  9. bman08 Says:

    [No Subject]

    It was always my impression that a lot of the developers jumped ship on BeOS because they were unsure about it’s future. It strikes me that an OSS BeOS would alleviate that problem.

    I don’t think it would be hard to find people for the proprietary code removal squad in Deej’s proposal. OSS zealots would line up if only for the opportunity to say “i told you so”.

  10. Anonymous Says:

    [No Subject]

    I think that same approach (dedicated volunteers under NDA) may be applied to publish half-existing BeOS updates - network, media, GL, drivers, font-angine etc. Be guys claimed about of lack of resources to do such bussiness for such side-product as BeOS is.
    Despite i’m not “trusted” Be-developers, i’m ready to spend a big amount of my free-time in this project if it will happen.

    Let’s go step-by-step.

  11. Anonymous Says:

    [No Subject]

    A path to an open BeOS is a deadend.

  12. rusty Says:

    Project requirements…

    I realize that most people think that the “requirements” of a project such as this is that it do what BeOS does, without the licenced code.

    Actually that should be the goal of such a project, and in reality, based upon the current state of BeOS, might not even be a good goal. A better goal would be to create a stable OS that executes BeOS binary executables without re-compiling. Of course others will say it must have Bone or OpenGL Hardware support, or (insert favorite speculative enhancement to BeOS).

    Personally I think the minimum requrements should be that it run OpenTracker and OpenDeskbar. Minimum distributable requirements should be that it have a Network Stack, rudimentary e-mail client that supports attachements, and has a web browser that doesn’t sit around waiting for all the graphics to finish downloading on a web page before allowing you to step back a page.

    Everything beyond that in my opinion is gravy.

    That said, I would recomend starting by figuring out what needs to be where in the Operating System. Basically the reverse of developing a driver, since you know the interface for a driver, how would you use the resources a driver provides. Etc.

    I suspect that both the boot loader in the form of Bootman and the from windows boot loader are both licenced code, so you will need to come up with alternatives to those. Lilo is an acceptable starting point for me at the moment, however long term I would recomend a program similar to Bootman such that you can create the boot menue by selecting the drives to boot, labeling them, and storing the menu and boot instructions to the boot sector of a hard drive.

    Next up would be a paired project where there are two “independent” groups working. The first group documents the libraries. I.e. what the functions in each library are, as well as what parameters are required and which are optional. Obviously an Object Oriented OS has to provide some of this in the form of programmer API’s, however some of the libraries are going to only be used by the kernel, and may not be documented. Someone will need to be willing to dig into the binary code and figure out the equivalent of the .hpp files.

    The second part of the pair would be resonsible for replacing those functions and objects. I am presuming that the documentation of which code is licenced and which is Be, Inc. souced is not generally available to either team, as a result you will probably have to re-write all of it.

    Hmm.. That presumption is based upon the possibility that Be, Inc. goes under and does not release any of the code to general public.

    You may be wondering why you would want a paired group for this. Go back to the days of the original IBM PC, and the original clones. Phoenix, Award, Compaq all had to have one team that did not code their final bios chips go through the documentation that IBM released on the original Bios (note that because of the anti-trust procedings they were under, this was actually the source code) and build a set of specifications for each bios call, that would then be coded by a bunch of coders who had not seen the IBM Bios code. Only in this way were they able to sell their own products. If the BIOS chips had contained IBM Bios code they would have had to pay high licence fees to IBM.

    Likewise if you are going to reverse engineer licenced code, you will need to go to a higher layer of coding, specifically the specifications of what the code does, without the how of that specific code, then someone else can treat that specification as psuedo code.

    Unfortunately the specification extraction portion is very time consuming. If you have the header files, as well as the commentary for each function or class, without the code itself, that can simplify things a bit. However it still will take time.

    Saddly I suspect that there will be other Operating Systems that are going to come along and provide better results in the interim, causing a significant loss of brainshare.

    -Rusty

  13. Zaranthos Says:

    If we wanted it open sourced we could pay for it.

    If the community really wanted it open sourced we could start a fund to pay for Be to do the work. The essential work would be take out the licenced code and document the “missing” parts. Even releasing it in non-working condition would be fine as long as the code was available to work on.

    I’m hardly a fan of open sourcing BeOS but vs it dying I would pay a large sum to save the wonderful code that makes BeOS.

  14. amien Says:

    In Response To Anonymous @ 05/09/2001 7:19:25 PM

    Re: One problem…

    This is something that I find frustrating. I program professionally, but there seems to be this assumption that people out there who can program will be able to program an operating system, in C. There’s quite a gap between what you get taught in a course at uni, and the select area beos users might program in in their work or studies or hobbies, and being able to write an OS. I don’t feel there’s enough support around in the BeOS community to get people going like this. I wish I knew what special skill it was that people who program on open source kernels have, the way they can pick htings up, etc. I don’t have it, and I don’t feel particularly lonely :) The community needs to be structured to be more supportive to undertake a project like this. (I bought the O’Reilly be programming book, and it’s shocking. It’s got mistakes all through it and I think it’s not as well written as it could be)

  15. Duffahtolla Says:

    In Response To Zaranthos @ 05/10/2001 00:15:34 AM

    Re: If we wanted it open sourced we could pay for it.

    It would be nice, but I have this hunch that..

    Be wont take our money.

    I don’t know why, but I think its true.

    They may release the unrestricted source on their own, or not. But I think our money will not sway them.

    They could make a good chunk of change by selling a 5.5 with the beta OpenGL,etc.. They don’t.

    It just doesn’t make sense to me.

    I wish one day when this is all over (for good or ill) that Someone could explain to me What in hell were ‘they’ thinking.. =(

  16. Anonymous Says:

    In Response To Duffahtolla @ 05/10/2001 01:39:14 AM

    Re: If we wanted it open sourced we could pay for it.

    Shame on you for doubting Be.

    Check back in 1 month.

  17. chrish_work Says:

    In Response To Anonymous @ 05/10/2001 05:57:17 AM

    Re: If we wanted it open sourced we could pay for it.

    Shut up, H-Kon.

    :-)

    But seriously, maybe Commodore will release some new Amiga hardware, too.

    - chrish

  18. Anonymous Says:

    How some companies have gone to OSS

    Digital Creations went open source and now Zope, their main product, is seeing a new release almost every month.

    (Zope is an application server for those curious.)

    You can find an article describing why they went to OSS here:

    http://www.zope.org/Members/paul/BusinessDecision

    It’s a great article.

  19. Anonymous Says:

    In Response To chrish_work @ 05/09/2001 09:32:24 AM

    Re: One problem…

    If you really DO think that you are that good and that dedicated, go work on Atheos, a functional, OS that looks and smells quite a bit like BeOS.

    It is already OSS and the head developer seems to have a good idea of what is going on. They need all kinds of help, from drivers to a new desktop interface (the one that they have is UGLY).

  20. Anonymous Says:

    In Response To chrish_work @ 05/10/2001 09:38:38 AM

    Re: If we wanted it open sourced we could pay for it.

    they are, and no not that AmigaOne thing.

    They also going to re-release the c64, but with 64Mb RAM not 64K.

  21. Anonymous Says:

    it’s not really open source…

    as you can’t redistube the source.

    however I’d like it.

  22. KHP Says:

    Tips for open-sourcing BeOS

    Looking at the BeOS about box, this is what I see:

    RealPlayer G2
    MP3 licensed from FHG (but based on Blade?)
    RSA
    USB support from Intel
    Indeo from Intel

    I believe the font renderer, also, is licensed from bitstream?

    Anyway… not that I’m a fan of open-sourcing BeOS, but it *could* be possible given the following:

    * RealPlayer could be scrapped (or you could talk to RealNetworks and work out something maybe)
    * MP3 could be re-implemented using LAME libraries?
    * I don’t know much about RSA, but perhaps something like OpenSSL could help here? (don’t have a clue)
    * USB seems to be pretty open, Linux source could probably get you going.
    * Indeo can be scrapped, probably. (Or maybe Intel would let you use it in the form of an .so?)
    * For the font rendering, you could use FreeType 2. Its an excellent OSS Font Renderer, and I think it has better quality and wider support than what BeOS currently uses.
    * There are a number of OSS bootloaders that could probably be adapted.

    Using these things would involve writing a lot of wrapper classes, probably. And you’d have to pay close attention to licenses and linking rules. And you would *REALLY* need to know what you’re doing.

    All I am trying to point out, is the licensed components *could* be replaced, or (harmlessly) eliminated. Of course it may take an unrealistic, super-human amount of work. I don’t know.
    [Edited at 14:01 May 11 2001 by KHP]
    [Edited at 14:02 May 11 2001 by KHP]
    [Edited at 14:03 May 11 2001 by KHP]

[powered by WordPress.]

Random Haiku:

Install BeOS
It is like the beauty of
Orange Blossom

Since 1998 - Until the Last User Leaves...
BeGroovy, established 1998

search BeGroovy:

BeGroovy Archives:

October 2017
S M T W T F S
« Jan    
1234567
891011121314
15161718192021
22232425262728
293031  

other:

23 queries. 0.083 seconds

[powered by WordPress.]