Linked by Thom Holwerda on Wed 4th Apr 2007 21:29 UTC
Zeta A lot of things have happened in the past few days concerning Zeta, BeOS, and Haiku. In order to create some order in the chaos, Eugenia and I have created a rough timeline of what happened the past 6-7 years. Read on for the timeline and some more thoughts on the matter. Update: Magnusoft ceases distribution of Zeta. Update II: Access answered the questions posed in the article.
Thread beginning with comment 228331
To view parent comment, click here.
To read all comments associated with this story, please click here.
Lefty
Member since:
2007-04-05

I've answered this question at least a couple of times, most recently in email. As a time saver (for me) I'll reproduce that response and give it a catchy title in the hopes that more folks will read it:

Releasing anything to open source requires a considerable amount of due diligence to ensure that the code being released is not encumbered in any way.

The BeOS sources amount to roughly 3/4 of a gigabyte of code. Having anyone "just come in" and do whatever to it would take a considerable amount of time and effort, and I'd still have to be involved myself and involve my engineering staff, who'd have to take the time to familiarize themselves with that 3/4 of a gig of stuff which we're don't, and don't stand to, derive any income from.

Preparing code for open source release takes considerable work, and it's not simply engineering work, either. There's no quick 'n' dirty way to make it happen--not without my company taking on what I'd consider to be significant and unacceptable risk. Since I am the open source compliance officer for the corporation, I'm required to go through the due diligence necessary to ensure that I'm not placing my employer in a problematical position. That's not a task I can take on lightly, particularly when my time is quite well-filled with activities which are directly relevant to our actual business activities, something that, as I've indicated, BeOS is not. Sorry: them's the facts.

So, I won't make any statement one way or the other as to whether we'll release the BeOS sources. And--in all sincerity, not to aggravate you further--if we decide to, you'll know about it when we do it, and not before: I donít want to create expectations in the community which I can't, ultimately, fulfill. For similar reasons (and issues of security, exposure to NDA'd third-party intellectual property, and the like aside), I'm hesitant to involve a raft of engineers, unknown to me, of uncertain quality, whose work I'd have to oversee anyway.

Operating systems are complex, as I'm sure I don't have to inform people here. You can't "just" do anything. Sorry, again, but that's a fact, too. You don't have to like it, but I can't change it.

I hope this clarifies my position and situation and provides some context.

Reply Parent Score: 5

memson Member since:
2006-01-01

Lefty, all that being true, I wonder if some of the ex-Be engineers might not undertake the work as a labour of love? Surely they have the experience with the code to be able to open up small chunks at a time. Really, releasing chunks at a time could be more useful to Haiku, because it's possible to focus on the chunks that Haiku is missing or is currently "poorer" at. I heard mention of the VM subsystem, for example. Maybe by limiting the scope and by focusing on things Haiku would benefit from, your opensourcing of sections of the BeOS source could speed Haiku development up. The way I see it is that when Haiku is ready to call itself "R1", a lot more heat goes off of the owner of the BeOS sourcecode IP to do something with it publicly.

Another thought. Given the magnitude of the task you'd undertake to actually *do* anything with the BeOS sourcecode, would ACCESS ever consider licensing the code to a third party to develop? I did get the feeling that there are people out there that would love to *do* something with the code. For example, create an updated version of the PowerPC build with BONE and using GNU tools rather than Metrowerks. Though the PowerPC market is miniscule, currently probably in the region of "hundreds" of users, there are plenty of old Mac's out there that would run BeOS though, and a PowerPC version might be made to support newer hardware too. The problem at the moment is a lack of "free" PowerPC distribution.

One final plea: Is there any chance that ACCESS could release a free version of R5.03 (last release of BeOS)? Say, ISO for PowerPC and Intel? That would actually solve an awful lot of issues that have been raised. If the binaries were released under some kind of licenset that allowed "controlled" modification and re-distribution, BeOS MAX could then become legal.

Edited 2007-04-07 00:32

Reply Parent Score: 3

Vibe Member since:
2007-03-12

Lefty, all that being true, I wonder if some of the ex-Be engineers might not undertake the work as a labour of love? Surely they have the experience with the code to be able to open up small chunks at a time. Really, releasing chunks at a time could be more useful to Haiku, because it's possible to focus on the chunks that Haiku is missing or is currently "poorer" at.

This sounds reasonable.

Though the PowerPC market is miniscule, currently probably in the region of "hundreds" of users, there are plenty of old Mac's out there that would run BeOS though, and a PowerPC version might be made to support newer hardware too. The problem at the moment is a lack of "free" PowerPC distribution.

Playstation 3? There's got to be some money in that.

One final plea: Is there any chance that ACCESS could release a free version of R5.03 (last release of BeOS)? Say, ISO for PowerPC and Intel? That would actually solve an awful lot of issues that have been raised. If the binaries were released under some kind of licenset that allowed "controlled" modification and re-distribution, BeOS MAX could then become legal.

Let it die with Zeta, so Haiku can be a real beginning.

Reply Parent Score: 1

Lefty Member since:
2007-04-05

Sigh.

I really hate to have to rain on anyone's parade...

The ex-Be engineers (ex-PalmSource engineers) made their choices. If they'd been interested in continuing to work on the ex-BeOS sources, now the property of their ex-employer, I'd have been more inclined to consider it if they hadn't up and taken off en masse. As it is, I'm frankly not.

Let me explain: we had some plans to use, for example, OpenBinder more broadly as a component in the ACCESS Linux Platform. The departure of the main engineer on that project, whom--since he reported to me at the time--I can only describe as "precipitous", left us in the position of having to yank all that stuff out again since we no longer had anyone to support it. This cost some time. Moreover, we couldn't spare the even larger amount of time it would have taken to bring someone up to a level even approaching his familiarity with the code.

(Moreover, the ex-Be engineers, the main one of whom I had directed to have no other responsibilities other than supporting OpenBinder, immediately stopped doing anything on Binder once he'd left. I was doubly displeased.)

So we're not using Binder at the moment, we're using D-Bus.

So, not likely.

As I've said several times elsewhere, the legal time involved in writing a license, negotiating terms (since even in a money-free transaction, there are always terms: we have interests to protect) and then administering that license (since a license, like any relationship, requires ongoing attention) would take a significant amount of legal time, which, likewise, we can neither really justify nor spare in the immediately foreseeable future.

So, not likely.

And releasing a "free" (i.e., all we get to do is spend money and time on it) version, one that we'd have to a) compile, b) build, c) test to see if it did, in fact run on PowerPC and Intel, d) fix whatever egregious bugs we turned up in steps a, b and c, lather, rinse, repeat.

Since we wouldn't, in this scenario, be releasing sources, I'd assumed you were, in fact, referring to a "free version of R5.03" which actually worked on something like even remotely current machines, as opposed to the machines of five of six years ago.

Moreover, I'd have to evaluate the status of any third-party software inveigled in all of this, determine whether we could use any of it or not, then have engineering shore up whatever holes got left because we couldn't use this or that...

So, not likely.

I'm going to just start using the permalinks. I don't really have the time to be saying essentially the same things over and over...

Reply Parent Score: 3

lucky13 Member since:
2007-04-01

the PowerPC market is miniscule, currently probably in the region of "hundreds" of users

This is why nobody in his right mind is going to pay (1) for a license to develop BeOS, (2) pay to actually develop BeOS, or (3) ever find anyone to distribute a re-developed BeOS. There's not enough demand to justify any step in the process.

there are plenty of old Mac's out there that would run BeOS though

Wow! That must be a lucrative market to which to cater an old OS that was designed for "modern" hardware! By golly, sell a $100 OS for a $10 computer -- wonder why nobody ever tried that before.

Development and distribution license: thousands and thousands of dollars.
Costs of updating and re-developing an OS: thousands and thousands more dollars.
Getting "hundreds" of owners of older $5-10 PPCs to fork over $100: priceless!

There are some things money can't buy. Good business sense is apparently one of them...

and a PowerPC version might be made to support newer hardware too

At what price? Where's the market demand outside of the "probably in the region of 'hundreds'"?

Read what Lefty wrote again -- the part that says: "stuff which we're don't, and don't stand to, derive any income from [sic]" and "my time is quite well-filled with activities which are directly relevant to our actual business activities, something that, as I've indicated, BeOS is not. Sorry: them's the facts."

IOW, it's too much work for them that can't ever pay them enough because there's not enough interest.

Reply Parent Score: 1

tonestone57 Member since:
2005-12-31

Welcome to OSNEWS Lefty,

Where you'll get some posters asking you the same questions over and over again *and* those who are not satisfied with your answers. :-)

You provide a good justification for not open sourcing BeOS & I agree with you. I don't think Haiku needs BeOS code (getting close to done on their own), though open sourcing *missing* drivers would have helped a bit out, but that is Ok. You've already assisted Haiku out, where you could, your efforts are appreciated.

Though I've given my personal insights and opinion; The legal issue needs to be worked out between Bernd & Access. Many will be interested in the outcome once resolved, though I have a pretty good idea how things are going to turn out & I'll comment on it next week with Bernd's statement. And I gave a little hint with one of my previous posts.

One of Lefty's comments that is relevant to the legal issue:
http://www.osnews.com/permalink.php?news_id=17623&comment_id=228034

Edited 2007-04-07 02:31

Reply Parent Score: 2

Vibe Member since:
2007-03-12

If you warez BeOS you'll get burned 'cause it's illegal.

*Badoom*

Reply Parent Score: 1

fyysik Member since:
2006-02-19

Tough. From this Lefty's posting i got impression, that even GPL-ed code, if used in Zeta, cannot to be published now, which creates very bad contradiction, as it should be published by license (as GPL-licensed derived work was distributed in public).

Not to say about BSD/MIT licensed pieces which are known to be created by third parties from BeOS community and then improved by YT, like some drivers.

Corporate (ACCESS) virus which polluted all what YT/Bernd touched?

Reply Parent Score: 1

lucky13 Member since:
2007-04-01

Corporate (ACCESS) virus which polluted all what YT/Bernd touched?

Access is a very open source-friendly company. I really don't think you should blame them for Bernd's mistakes.

Reply Parent Score: 2

Lefty Member since:
2007-04-05

Hm, well.

(I'm not sure I've got you correctly, but let's run with it...)

If Mr. Korz were to have used code which is covered by the GPL as part of Zeta, as it's been sold, I'd imagine he'd have had to have adapted it for that environment in some way. If so, since the modified versions seem not to have been republished or offered for reproduction (to legitimate purchasers of binary versions of Zeta), then that would presumably put him already in violation of the licenses on, and in infringement of the copyrights of, any of the code he used in such a fashion.

Is that what you meant...?

Reply Parent Score: 1