OpenBeOs Developers Talk About Progress

DeltaSigma writes: "Michael Phipps, of the OpenBeos team, recently hosted a public Q&A Session where many of the public musings over a completely new open source operating system have been addressed. The answer to all the 'is there room in the market?' questions was answered in a way: 'We are an OSS project. Marketing is not our job.' Perhaps more /.ers could keep this in mind ..."

Comments

A Meta-Question

[sllort]

I've noticed that Slashdot tends to post articles, like this one, that link to "IRC logs". I'm not familiar with this particular application, but personally, I use AOL Instant Messenger, and so does everyone I know. I think the quality of these stories would be greatly improved if they could be posted in AOL-IM format for "the rest of us" to read. I know that Slashdot was recently picked up by Forbes magazine; now that Slashdot is reaching a wider audience, maybe it's time to accomodate other readers?

Other than that, an excellent technical article. Bravo.

Re:More readable version

[Narcopolo]

Thanks, and a thump on the head to the people who moderated this as redundant.

Ok, now in human readable format

[joshv]

Still not perfect, but much better:

<Captcpu> and now.. we are starting. :)

<mphipps> So I want to ask any people in the north americas to let the Europeans ask questions first. Just to be nice.

<Captcpu> good evening everyone, I'm Captcpu and I'll be you're moderator this evening :)
<Captcpu> and that's mike if you didn't know :)

<mphipps> Hi! Even though we are starting early, I will stay until 8 (my time) unless we run out of questions. ;-)

<Captcpu> the way tonight's chat is going to work is, you can private message me your questions, and I'll foward them to mike :) so, Private message me away, I'm waiting :)
<mphipps> What if we threw a war and no one showed up? :-) Someone, ask something!
<Captcpu> we'd be in trouble :)
<Captcpu> it's simple /msg captcpu Your question here
<mphipps> monolith - no estoy embarrazada. :-)
<Captcpu> Hey Mike, Monolith would like to know if you're with child.
<mphipps> Monolith asked if I was pregnant. :-) While I first said, no, I have to say that it feels like we are giving birth. ;-)

<Captcpu> <lillo> first Q: is someone already working on the kernel VM? It seems like a very urgent task to be completed before other things can be touched...
<mphipps> Yes. Next question. ;-)

<Captcpu> <Matzon> Since newos is nowhere complete, why did we fork this soon? is he willing to divulge any rough specifications on what the VM will and will not support? ie. integrated VM/FS cache and/or mmap/munmap?
<mphipps> Good question. I am wondering that myself. ;-)

<Captcpu> and to add to it, he has to say... <Matzon> Why rename all newos methods to be BeOS compatible (thus making obos kernel undiffable with newos), when we could have provided an abstraction layer?
<mphipps> Truth is, Travis et al have a lot going on in their lives and they haven't been making major changes to NewOS. If you go back even 4 months or so, there weren't a lot of check ins. And we are getting to the point where we need to spread our wings and fly.

<Captcpu> here's our next question from <BlueOS> When do you join B.E.OS? and a secondary more serious question: <BlueOS> What about binary compatibilty in the driver side?
<mphipps> As for the second part, we are doing the easy stuff ATM. There are bigger changes to be made that will make the question irrelevant. BlueOS - when will you join us? ;-) I don't see anything in our way to having driver level compatability. No bus_manager compatability, though.

<Captcpu> here's another one mike... <AnEvilYak> is he willing to divulge any rough specifications on what the VM will and will not support? ie. integrated VM/FS cache and/or mmap/munmap?
<mphipps> Yes to both of those features. I would not want to ship without. At least, that is the plan ATM. ;-)

<Captcpu> Here's a nice one from: <mwilber> Why did you decide to start the OpenBeOS project?
<mphipps> Insanity. ;-)

<Captcpu> good answer :) but wait..there's more...
<mphipps> Seriously - I had a project that I have been working on for years on BeOS. The short version of the story is that it is an object oriented paradigm in which every class is a process and every instance is a thread. It needs hyper fast messaging and process/thread swapping. No other OS will do that. Even R5 wasn't the best. So, when the Palm announcement was made, I looked at Linux and the BSD's, but none of them were as fast and as easy to use. So I decided that BeOS must continue on.

<Captcpu> here's an interesting one from: <shatty> the new font engine is going to be freetype right? when can we expect things to reach the stage where that is working? it's part of app_server, so we need to wait for app_server don't we? :-)
<mphipps> The font engine is indeed going to be freetype. Proto 6 is in development right now. I don't think that font handling is in it, though.

<Captcpu> Here's one: <macdonag79> Why are people writing custom versions of basic CLI tools when they could be obtained from, eg, *BSD?
<mphipps> Good question. Many of those custom CLI apps are little tiny things to help us develop. I doubt that they will be part of the final "release". But since we don't have bash, et al, yet, some little things to help us work and test make sense. The whole shell is scheduled to go out the door.

<Captcpu> <elver> How is the name choosing going? Will we stick with OpenBeOS (unlikely) or do we have a list of possible names already?
<mphipps> We have been cutting the list of names down. Many that we submitted were not usable for one reason or another. It hasn't been our top priority, but it is getting there.

<Captcpu> Here's an interesting one: <linn> question: who are you? you = captcpu. may I answer Mike?
<mphipps> sure
<Captcpu> I'm a web developer and college student from Las Vegas. I'm Kurtis Kopf's Business Partner in real life, the guy that's designing the new website for Open-Beos :) and I've agreed to help as I can :) besides that I'm a pretty normal guy.. I think..

<Captcpu> Someone (forgive me for forgetting your name) wanted to know about the Game Kit, what's the status?
<mphipps> Game Kit. Ahh. The poor, misbegotten step child. :-) Honestly, I haven't heard from RobMed in months. GK is one of the items on the list for the poor unfortunate soul who signed on to be the "C++ code finisher".

<Captcpu> Here's one: <Matzon> I read a review of TuneTracker, wherein Dane Scott mentioned that obos would have a first release this year - ofcourse he is only speculating... but what is our (roughly) timeframe?
<mphipps> Sigh. So many ways to answer this. There is the Commodore answer "when it is done". There is the Microsoft answer "Q3, 2002". There is the real answer... I very strongly feel that we could finish this by the end of 2002. It will mean some long nights and hard work. But I think that it could happen. What would it take? Some more *REAL* help. Some people willing to really dig in, grab some piece of work and ride it until it is done.

<Captcpu> Ooooh.. question about networking: <z3r0_one> Question: I've seen recently that the network stack is being moved into the kernel (if it hasn't already been done), and that sockets will be file descriptors. Does this mean that the net team borrowing ideas from BONE, and trying to recreate it? Has the problem with select been fixed yet?
<mphipps> Where is David when I need him? ;-) select will depend on the new kernel it's not properly fixable on R5. The networking stack, as it is today, runs as a kernel module. Is this BONE like? No. BONE had it truly built in, I think (I wasn't on the beta list). The select issue is very easy, with source code. The OBOS kernel will handle select right, out of the gate. It remains to be seen if our networking stack will work properly with R5's kernel's select. I doubt it.

<Captcpu> Here's we go...
<Captcpu> <DragonSoull> 1. I've heard requests to make server applications like mySQL run in OBOS. Are you doing any work in that direction? And if so don't you think that's outside of the "Desktop OS" focus of the original BeOS?
<mphipps> *Excellent question*.

<Captcpu> I figured ;)
<mphipps> mySQL would certainly be required on a server. But I can see, too, where it would be cool to have it on a client. For developers, for one thing. So I don't think that doing some small amount of work to make mySQL work is a bad thing. But that is different from major work or porting it ourselves. I ***STRONGLY*** believe in the focus on the desktop. BeOS wasn't, isn't and shouldn't be a server OS. Does that mean we shouldn't be able to run ftpd? No. But the *FOCUS* is on the desktop.

<Captcpu> Here's an innocent question: <x-gh0st> Will R1 support localization or is this feature reserved for R2?
<mphipps> R2. Localization is *VERY* important. And it *has* to be done right. And it really needs a whole ball of other features, like GUI with a layout engine.

<Captcpu> and now we have hit the 19:00 hour. Oficially starting =) <M_BeOS> can i ask, 'What plans have been made to succeed in the areas Be failed in, the marketing, the lack of drivers, and apps? Without these we could be in for a repeat....'
<mphipps> That is probably an FAQ. ;-) Few, honestly. We are an OSS project. Marketing is not our job. I expect that OBOS will have a "RedHat". Some company to come along, package our source with a dozen CDs of apps and sell them. Driver wise, we are helping and supporting anyone who is interested in writing them. Scott is doing an awesome job at BeDriver, and BU is working in that area, as well. I am hoping that with the shrinking of the hardware market, there will be fewer drivers to write.

<Captcpu> Here's one from: <coolbear> What are the GUI Interface plans for after version 1.x, are they in consideration now? I have some proposals, where should I send them, and in what format?
<mphipps> Yes, we are thinking about a number of things.

<Captcpu> " <coolbear> I refer to API and features. "
<mphipps> There are a few proposals out there, now, for new looks and feels. I have certainly heard about Gonx enough times. ;-) API wise, I haven't hear a whole lot. Any proposals, thoughts, ideas go to Glass Elevator.

<Captcpu> here's a long one from: <misza> Do you agree that it would be better to clone the current UI (Yellow Tabs, same functions like double click a tab minimize it) and implement all those functions exactly, and provide an interface to skin the UI(e.g change position of buttons, implement light skinning) having the current UI as the default one because that is familiar to alot of users, Rather than to create an entirely different UI that may be infl
<mphipps> For R1, there is no promise of skinning. The only "promised" feature is that it will look like and work like it does today.

<Captcpu> Here's a good one from: <AlienSoldier> As to help the community and OBOS in itself, does the OBOS team will suggest a line of hardware so that futur buying streamline the community to have easier driver transition in the first release R1.
<mphipps> Sigh. ATM, this is a tough one. We have the Matrox driver "in the bag", so for video, that is my only promise. If the kernel boots on it today, it is likely that it will tomorrow, too. As for other stuff (networking, sound, etc), we can't really say. If current R5 supports it, esp if the driver is publically available (i.e. source), we probably will, too.

<Captcpu> and now <sdrsolo> states "Out of the 214 listed programmers how many are contributing?"
<mphipps> Not enough. In fact, we are reworking the web site to make it more representitive of those contributing. Let's put it this way - there are 32 people with CVS write access. And none of the team leaders are beating me up because they have to submit so many patches. OBOS is still a place where *1* person can make a huge difference.

<Captcpu> and... <mwilber> have you been contacted by any corporate/government/educational organizations that are interested in OpenBeOS?
<mphipps> Yes. A few corporate. And we are working with those.

<Captcpu> this reminds me. there. that fixes that. anyway.. here's one from: <z3r0_one> Question: Another tough one: Is true multiuser support in the future of OpenBeOS?
<mphipps> Nice. :-) It all depends on what you mean. ;-) For example - If you mean "I want to log in and have my own email settings, home dir, etc", then that is an R2 thing. If you mean "I want OBOS to be just like my Linux box where I can have all of my 1337 friends log in and download warez", than no.

<Captcpu> Here's a good one: <M_BeOS> I can't program, and neither can many of us... How can we help on other ways?
<mphipps> Just like a kid at camp. Send money. ;-) Seriously, though, folks. One way is to volunteer at some of the other BeOS groups. Many of them are always looking for *GOOD* help. Translators can always mirror and translate our site (or others). Being helpful in the community. Being on IRC and answering questions. And also doing testing. We are (slowly) getting toward another release. We would like more people testing and looking at things.

<Captcpu> Here's one from the Dark Side of the Mac: <mdvb747> Are there any plans for a PPC version of OBOS?
<mphipps> When Steve Jobs calls me and asks for one. Seriously, though - I love PPC. I *want* to give my Mom an iMac and have it run OBOS. The question is time and resources. I can't justify dropping VM or other "generic" work for a port to a platform that doesn't really want us. If someone wants to do it, I am more than open to helping in any way. But I can't justify "assigning" people to it.

<Captcpu> and an interesting question from: <miloshe> When can we expect USB support and a nice media player?
<mphipps> USB - Hopefully with R1. I have a good USB book and I have done some preliminary work on it. "Nice Media Player" is somewhat vague. I would have to ask what is wrong with R5's.

<Captcpu> <El-Al> Is anybody working on OBOS API documentation? <El-Al> .....and if so, what tools are being used to documen<El-Al> Is anybody working on OBOS API documentation? sorry abou the repeat there at the end :)
<mphipps> OBOS API documentation has a team, yes. What tools? Not 100% sure. We are using doxygen inside the code.

<Captcpu> Here's one from down under: <SmallStepForMan> Well, OBOS has been going for just over 9 months now. How do you look back on the last 9 months, and is it according to expactations, subpar or way beyond expectations. BTW - good morning from Australia
<mphipps> G'Day to you! The last 9 months have been unimaginable. I had no clue what I was in for. None of us did. It has been a fun trip, though. My only regrets are a) that I can't do it full time and b) the price that my family (and others) have paid.

<Captcpu> and from: <el_d00d> Will there be localized versions of OBOS, and can some of us help translating OBOS?
<mphipps> el d00d, huh? :-) Nice. Localization (which we touched on briefly above) is an issue for R2+. I know how important it is to you. But it really should be done right. And that means when we have the time and freedom to do it the Right Way.

<Captcpu> Mike, Monolith has informed us that our spanish sucks.
<mphipps> I didn't make that one up. ;-)

<Captcpu> but.. onto the next question: <grim> How's progress with the change of team structure going? Are the leads working on todo-list-type pages? IMO, it would need to be pretty fine grained to allow people to pick up a smallish task, work on it, submit it and be happy:-) Rather than TODO: Finish the kernel...
<mphipps> The team structure didn't really change all that much. Adding people to the teams changed. And that went pretty quick. Some TODO lists are done. Some need more work. If you want to work in an area and can't find the todo list, ask the team lead. They will help you. And if they don't have one, they will get sick of people asking... ;-)

<Captcpu> Here's a long winded one: <lillo> currently there's a debate on wheter obos should be fully graphical or if it should have a text console as well just in case. It's sure that the second would be useful during kernel development, but it'd break the BeOS phylosophy... What's your opinion?
<mphipps> This is in reference, I think, to a conversation that we were having on the kernel list. For the average user, I think that the system should be just as it is today. I can see some value to a boot disk that is text based. For either emergencies or as a "download this, see what hardware it detects". But not for users to use every day.

<Captcpu> <M_BeOS> Are we going to have humorous API call's just like Be did?
<mphipps> I won't dictate that, either way. I certainly intend to extend the "is_computer_on" series in the kernel.

<Captcpu> here's one from: <Deris> Will there be a new OBOS update that will replace certain things (like apps and preferences)
<mphipps> I think that this is asking about an automatic update. Personally, I like that idea very much. I know that there has been a lot of interest in the community about that. I would love for someone to write such a thing. Can't promise that it will be us.

<Captcpu> Programming, Oh my!: <mmu_man> What about a BeIDE replacement ? (I suggest XEmacs >:-)
<mphipps> E$*#()#*Q)(? Begone you heathen dog! There is aught but one True Editor: VI! :-) I put BeIDE in the same catagory as a number of other apps (browsers, mail clients, etc) - every distro should have one. But they are not necessarily the realm of the OS group.

<Captcpu> <mene> Will opentracker be moved into OpenBeOS CVS..and developed under OpenBeOS name?
<mphipps> I doubt it. That would "cut off" other groups that might want to use it.

<Captcpu> here's one: <scanty> What can we expect from the OBOS POSIX-layer ?
<mphipps> Hopefully, functionality. I would *love* to see it work as well as R5, plus mmap and select.

<Captcpu> cute. fuzzy. animals? <elver> Linux has Tux, BSD has their cute little demon. Should OpenBeOS have it's own cute animal? If so, what should it be? :)
<mphipps> OpenBeOS will not have a cute animal. Whether some other, different name has one or not is up in the air. The admin staff is about 50/50 split on whether it is a good idea or not. It would have to be a good one.

<Captcpu> <Zaranthos> Question for mphipps How much time do you spend on OBOS? Do you work full time and then work on OBOS?
<mphipps> About every waking minute. I do indeed have a full time job. (And a wife and 2 kids). I work on OBOS an average of 4 hours a day.

<Captcpu> Packages? for Me?: <RageMax> are there any plans for a _standard_ package format for program distribution, possibly for R2?
<mphipps> Plans? No. This is a veyr good idea. And I think that it ties in to the installer/updater mentioned earlier. (Pardon my dyslexia).

<Captcpu> Cash? Money?: <earlcp> How configurable will the VM and DiskCache be for users? Example I would love to devote 512K to just the DiskCache.
<mphipps> Ideally, not at all. Very briefly, the ideal VM/cache system would need no tuning and would always have the information you want to access in ram. I know that we won't be *ideal*, but very good would be nice. I would like to think that mundane details like disk cache and VM size would be something that the OS could handle on its own.

<Captcpu> <Der.is> Will you release a commercial version of OBOS (for companies), so that you can afford OBOS-update servers or such? <Deris> =) sorry!
<mphipps> I think that the question here is really "How will you afford to run OBOS without any money?" Part of the answer is that we will cross that bridge when we come to it. Part of the answer is that I am hoping that distro makers see good reason to "give back to the community".

<Captcpu> <elver> What about firewalls? Do people have to code their own firewalling software (like in windows 9x) or will the firewalling be built into the kernel or the net_server? (more like Linux)
<mphipps> When you say firewall, you could mean a couple of things. BeOS, by default, had all of the ports closed. Now, I am not a security guru, but that seems very secure to me.

<Captcpu> <monolith> does he aim to try and get market share in the desktop arena, taking from MS, or does he plan on creating the best possible OS, with no compromises made for the less technologically-capable folks that make up the general public?
<mphipps> This is really a marketing question. Think about Linux for a moment. We are more like Linux (the organization) than Be. I am not here to change the world, necessarily. I am not here to sell N boxes per year. I am here to make an OS that I want to run and that you want to run. But I also want to make it work for my Mom (a non-techie). I don't believe that these two things are mutually exclusibe.(exclusive)

<Captcpu> and now.. for the last question for this session.
<Captcpu> <Zaranthos> Is there a timetable for an official OBOS name so we can start grabbing up domain names? :^)
<mphipps> I am assuming you mean "to help out the project". :-) I just got some mail about this, actually. Hopefully, we will have an annoucement of where the name issue is going soon... Folks - this has been a lot of fun. I appreciate you all taking time out to come here. I do have to go, though.

<Captcpu> I would also like to express the many private messages I've recieved thanking Mike and his team for all the hard work they have put in.
<mphipps> I think that we will probably do this either weekly or every other week. We will get the logs posted shortly... Bye, all! Thanks again!

<Captcpu> And that ladies and gentleman is the end of the session. Thanks for Attending!