[openbeos] Re: Issues

>
>Hey all...
>
>John Gabriele has been working on an updated FAQ for the website. He 
>came up with a few "issues" that I don't have a definitive answer for. 
>So I'm going to post them here and see is anyone else has some good 
>answers.
>
>* What is the "sockets-as-files" issue?
>
>I think this is another way of saying will OpenBeOS support true 
>Berkeley sockets? Ok, net guys, are we? Can we support this and still 
>retain source compatibillity for R5 net programs?

This is really a posix issue. The POSIX calls (select, among others) are the
ones that need this implemented. The POSIX interface to networking would have
to support this. The networking stack can live anywhere.
I can't think of any reason that this would break R5 programs.
I don't know if this is the time to build it in or not. It is really most 
useful for
porting Unix apps. I don't know that that is sufficient for us to go beyond the 
"No new features, just R5".


>* What is the "memcpy" issue?
>
>Dunno, you got me! I've never even heard of this one. What's up with 
>memcpy?

I think this might really be the mmap issue. I.E. BeOS doesn't support it.
NewOS will support it shortly, so it is really a freebie for us.

>* What is the issue with BONE and the Printing Kit?
>
>Well, I know that the BONE beta that's been floating around screws up 
>printing somewhat (well, I've read as much on BeGroovy threads). But 
>probably a better way to ask this is "Will OpenBeOS support BONE?" Now 
>BONE itself is an upgrade to the R5 kernel. AFAIK, our net guys intend 
>to keep networking in userland. This is correct, no? However, they 
>could design it so that it provides BONE-like services. Are there any 
>plans for this? (this relates to the sockets question above)

If I understand the issue between printing and BONE, the problem was that
printing stuck its hands in Networking's shorts to do some network printing
stuff. BONE broke that, not being in userland anymore and having a different
API. This should not be an issue for us. No, we will not support (or need)
BONE. The fact that we are doing a kernel replacement sort of prohibits that.

>* What happens when www.be.com is no longer available?
>
>Actually, I don't think this will happen for awhile -- a year or more, 
>to take a wild guess. Still, something to think about. I know that all 
>of the old Be Newsletters are available as one big zip (which I've been 
>meaning to grab). Stephen van Egmond has the annotated BeBook online 
>(and available as a download), so that's taken care of. Beforever (and 
>several other sites) have mirrored some of Be ftp stuff including the 
>Personal Edition download. Ryan Christiansen just recently zipped up 
>the entire Be Bug Database (up to Oct 9 2001) and made it available on 
>BeBits (baby, it's big!). There are probably still some bits and pieces 
>that may need to find a new, safe home.

I don't know that we care all that much. I have the newsletters already.
I think we could legally make them available - they were on Be's web site
for public consumption. We all have a copy of the BeBook, and the annotated
one isn't related to Be.com. The FTP site is mirrored. I don't know that we
care about the bug database.

I guess, what would we want/need be.com for?

>
>* Can you recommend any books to help me get started with OS 
>development?
>
>Uh...uh...
That is a big question. OS development is a big question. I would recommend
different books to the Networking team, than I would to the Kernel or 
(theoretical)
OpenGL team. 

I would recommend some good C++ books. For Kernel, I would say the 4.4 BSD book.
I picked up a Linux Kernel book that is OK (although Linux kernel internals 
don't hold
a candle to NewOS's internals). 

>* What happens if Palm agrees to license the code to BeUnited?
>
>Oh boy, the big one! Well, for my part, I plan to continue working on 
>OpenBeOS for as long as it takes until it's finished, regardless of 
>what happens with any other initiatives. Hope I'm not the only one!

That is probably the most *F*AQ of them all.
I will say what I have always said. 
It depends.
Since the license doesn't exist, discussing its terms and the future is more 
speculative than the stock market. Under some circumstances, I would want 
nothing to do with the venture (i.e. a very low funding company that can't open 
any of the source and doesn't have the means to improve much of anything). Under
some circumstances, I would JUMP at the idea - BU opening the source to 
some/all 
people in OBOS (maybe under NDA) to help us/help them.

I guess the bottom line is that I will do the best thing that I know how for 
the Be community.
And I will encourage us all to do so.

Other related posts: