Kansai Open Source Forum: Day 2 Report

Blog post by koki on Tue, 2007-11-20 07:13

Demo machine at KOF Haiku boothDemo machine at KOF Haiku boothNot surprisingly, it took me much longer than originally planned to find the time to write about the second day at the Kansai Open Source Forum conference (KOF). But no worries: memories are still quite fresh, as it's usually the case when things go well and you have fun. The second KOF day started earlier, especially for me. The exhibits were scheduled to open at 10:00AM, so with Momoziro we decided to meet at the hotel lobby at around 9:00AM. But I was up way before that, at around 5:30AM (compliments of my very jet-lagged old body). So I sat in front of my laptop to write some emails and then went through my Haiku presentation slides again, making little changes here and there, changing the order of a few slides and even adding a couple of slides based on some of the questions that I had received during the first day at KOF.

Later on I had a nice breakfast, then read the Japan Times (like I used to during my good old times in Japan) and watched a bit of the news on TV. A bit before 9:00AM I went down to the lobby where I met Momoziro, and after a short subway ride to the convention center, we were both at the Haiku booth on the exhibit floor, setting up our table-top Haiku gear. As there was sometime left before the start bell, I went to see if I could test my laptop with the projector used in the conference room where I was to give my Haiku presentation that afternoon; I was asked to come back during the lunch break, and that's what I did (more on this below).

It was easy to tell from the amount of people walking around the ATC complex that this was going to be a much busier day, and we were excited that it turned out to be as expected. As soon as the exhibits were open, people started flowing into the floor, and we were again showing off Haiku to the crowds. There is nothing more comforting for somebody who has crossed the ocean to present something (Haiku in this case), to see that the effort was worth it, and the visibly increased traffic gave me that warm feeling. According to the organizers of the conference, attendance was higher than last year, and way over 1,000 people (unofficial count).

Not long after the conference started, we were pleased with the presence of Hiron-san, the developer of CoveredCalc, a very cool multi-platform skinnable calculator also available in a BeOS version. I had communicated with Hiron-san for quite some time over email (since my times at yellowTAB and JPBE.net), but this was the first time we met in person. And as it's usually the case, meeting someone in person is so much better and rewarding than exchanging emails. Hiron-san has recently been very intrigued by Haiku, and is following the project closely (he recently reported a BMenuItem-related bug that was promptly fixed by Axel). CoveredCalc for BeOS does run in Haiku, but it still has some unresolved issues (not sure what they are). Nevertheless, Hiron-san intends to eventually release a Haiku version of CoveredCalc, and hopefully other apps at some point in time. I want to thank him for showing his support by coming all the way by ferry from Tokushima, Shikoku, just for the day (I believe it is a 3 – 4 hours trip). Hope we can meet again Hiron-san!

As the day went by, Momoziro and I kept busy showing Haiku to the visitors, explaining different aspects of the OS and/or the project. We also handed out a lot of flyers; in fact, we had printed about 50 of them, and they were all gone by around 2:0PM; in a way, that's a good sign (that there is interest to know/learn about Haiku), but it also meant that we printed too few; gotta do better next time. :)

Momoziro's laptop running HaikuMomoziro's laptop running HaikuAround noon, I went to check my laptop with the conference room projector. My laptop has an analog monitor output, and at home I was able to have Haiku drive an external LCD display through this output; so I was somewhat confident that I could run Haiku natively and display using the projector hooked to the external monitor output. After some trial and error, I was able to use that setup, but I had to switch the video of the laptop to the external output only "before" booting into Haiku. I could certainly live with that; all I had to do is change the angle of the podium, so that I could see the screen while I was talking. This worked well in the end.

I then went back to the Haiku booth, where I spent most of the time running Haiku demos. So, what did I demo, you may ask? Actually, it depended on who the audience was. For those new to Haiku and BeOS in general, I would give a quick overview of what Haiku is, focusing mainly on some of its unique points; then I would actually do a hands on demo on my laptop running Haiku natively. This usually included a quick overview of Tracker/Deskbar, a bit of OpenBFS magic (live queries, attribute display and editing in Tracker, custom attributed using FileType preference app, etc.), and then running various applications, to show both how far Haiku has come, and to showcase the snappiness of the OS in general. To those who knew or were familiar with BeOS (or Zeta for that matter), I went straight to the point, telling then “Hey, look, we have come this far!” and running various apps like WonderBrush (thanks for the artwork Stippi!), BeShare, Vision, BePodder, and Opera among others. I would then put the icing on the cake by showing off some old but memorable apps such as 3DMov to appeal to their (dormant?) BeOS hearts. :)

Unfortunately, I could not run my favorite Haiku demo with 7 – 8 videos playing at the same time (see video here), as since about a month or two ago Haiku will start dropping both sound and video frames after the second instance of the media player. I am sure the devs will take care of this at some point (which reminds me that I need to file a bug report about this). By the way, this is really not a complaint, as we all know that Haiku is still not even alpha, and as such, it runs very well at times.

My Haiku presentation was from 4:00PM; since there were ten minutes between presentations for preparation purposes, I went to the conference where I was presenting about fifteen minutes in advance. The previous presentation (about Nadeshiko, a programming language in Japanese) ended early, so I had plenty of time to prepare. By the time I was ready, at around 15:50, I had an empty room before myself. The room had a capacity of about 50 – 60 people, and it felt quite lonely. Fortunately, people started to show up, and I am happy to say that by the time I started, there were approximately 30 people in the room. This was a little above average, according to what I heard, so not bad. :)

My presentation was simply titled "What's Haiku" and was meant to provide general information about the project in general, including key highlights of the OS, a brief history of both Haiku in itself and in the context of a BeOS timeline, the Haiku code base (ie., a fork of the NewOS kernel, OpenTracker, OpenBFS, AGG-based graphics system, Freetype as a font engine, etc.), current status of the project, and recent developments (SATA support, OSS port, Webkit port, GSoC projects, etc.). I had 16 slides total, that took me about 25 minutes to go through. I then went into a hands-on demo, and wrapped it up with a Q&A session, for the full 50 minutes that I was allowed.

I have a confession to make. I prepared the slides in OpenOffice and created a PDF with the intent of doing the presentation running BePDF in Haiku natively. I had tested this setup (actually inspired by François Revol), and it did work. I installed the same Japanese fonts on both Windows (where I was creating the slides) and Haiku, and the results were good. BePDF displayed the PDF nicely, albeit slowly, with Japanese characters and all. But Haiku had not been behaving very nicely on my laptop lately, and in the end I did not have the guts, and went with plan B, which consisted of running the presentation in Windows, and then rebooting to Haiku to do the hands-on demo. Ironically, during the hands-on demo, which lasted about 20 minutes, Haiku behaved quite well and did not crash (I did have to restart Tracker a couple of times). Shame on me...

Overall, my presentation focused more on general facts about Haiku the OS and the project, as there is very little information about Haiku in Japanese. I did not even try to dive into technical details that I would most likely not be capable of understanding or explaining anyway. In spite of the lack of depth from a technical point of view -- and my dishonorable decision to present Haiku using Windows -- I think the presentation went well, was well received and gave the audience both a good insight into the project and some food for thought at the very least. Unfortunately, there was no video feed, and I don't even have photos of the presentation, as Momoziro had to stay at the Haiku booth; but at least I hope you can get a feel of how it went from this brief overview. The slides I used for the presentation are available for download (527KB PDF). Needless to say, they are in Japanese; maybe I can find some time to translate them into English, so that they can be used at least as reference material for others.

After the presentation, I went back to our booth, to find Momoziro surrounded by quite a crowd. Several of those who had attended my presentation went down to the Haiku booth to ask more technical questions (I had advised them to do so). During this last stretch of the exhibits, I spent some time with Daiki Ikeda, editor of Software Design magazine, who showed interest in Haiku and is considering some coverage in the future (perhaps when the first alpha is released), and maybe even interviewing one or more of the Haiku developers. I also had a nice conversation with Suguru Hamazaki, a former BeOS fan who could not help but be impressed by what Haiku had achieved. Suguru-san, who works for IBM, attended my Haiku presentation and was the individual who posed the most interesting and challenging question of all during the Q&A session (he also blogged about Haiku in his personal KOF report). I am sharing his question at the end of this blog entry, as I found it to be very thought-provoking and something that, as as community, we may want to ask ourselves as a way to become more aware of the future possibilities of Haiku. More on that at the end of this post.

Back to the booth, the remaining exhibit time went by very fast. The people that came to our booth after the presentation for additional information kept us quite busy, and before we could realize it, it was closing time. My overall impression of the conference is positive. Mind you, there is always room for improvement (more local support would have been nice). But the audience was the right one, I think, so from both the perspective of spreading the word, I personally feel attending KOF was worth it. I also had a lot of fun, not only showing off Haiku, but also meeting people from both the local community as well as other projects. I would probably do it again next year if I had the chance. I do hope that we can see more proactive participation from the local community by then.

We then packed our things, said sayonara profusely to everyone around us, including the organizers and some of the booth neighbours we made friends with during the conference, and left the convention center. We went back to the hotel to leave all our stuff in my room (Momoziro had already checked out), and went for a Korean BBQ dinner. It was an all you can eat restaurant, so needless to say, I ended up like the snake that had swallowed an elephant in The Little Prince. ;) At around 9:00PM, Momoziro headed to the train station for his return home to Hiroshima, and I went back to my hotel room for a much needed rest.

That sums up what happened on the second day of KOF. As I mentioned, I wanted to finish this blog entry with a question that Hamazaki-san made during my Haiku presentation Q&A session. So, what did he ask that was so thought provoking and worth sharing? It was a simple question: he asked what our target market was. Ehem... Well, I started giving the "we are focused on the desktop" canned answer, but I could see in his face (and that of others) the next question coming: where does that leave Haiku in the world of today's computing? I felt I had to answer before they had a chance to ask, and so I did. :)

My longer answer was this: while Haiku was still a long term proposition (as in, it will not become usable by the average Joe in the next six months), in addition to the desktop, it also had future potential to address niches similar to those where BeOS excelled, and to illustrate I mentioned some of the commercial applications where BeOS was/is being used, such as ROLAND Edirol's dedicated video editing system, the TuneTracker radio automation system, in-store kiosks, etc.. Following hints from people in the business that I have met in the last few years, I also mentioned the potential suitability of Haiku for certain embedded applications, such as set top boxes and dedicated media centers. I also expressed my personal view that, at some point, Haiku could make a great personal OS for the small form factor desktops, laptops and other smaller modern computing devices that are becoming more and more prevalent in today's digital world (François, port Haiku to the Nokia N800!). I wrapped up by giving my view that the future of Haiku is wide open, and that while the project is focused on the desktop, this is open source, so nobody can discount the potential use of Haiku in personal computing in a wider sense of the word.

Why do I bring this up here? Because I have been actually asking the same "what need does Haiku address?" question myself. I am not suggesting anything like a focus shift; I am pretty happy with the direction Haiku has taken, and the progress that the dedicated Haiku developers have been making lately. But I still wonder what everybody else thinks in this respect, more as a way to explore what sort of possibilities Haiku may have in the future. So I will leave you with this question, if nothing as food for thought and/or discussion.

I hope you all enjoyed the report. :)


Re: Kansai Open Source Forum: Day 2 Report

I think that Haiku would fit nicely on a system like ASUS eeePC or other UMPC and Haiku would be an awesome mediacenter pc running on a small itx.

Re: Kansai Open Source Forum: Day 2 Report

I think I like the direction that question is headed, because it begs the point, "Is Haiku destined to be the 'generic desktop OS of hobbyists,' or will it *also* have a noted area of strength that gives it special appeal to a particular audience?"

Windows is a dominant, general purpose OS that's used for everything. The Mac has appeal to creative people like graphic artists, publishers, videographers, etc. Linux, appeals to geeks, programmers, and businesses looking for cheap servers and office software. What does that leave us?

Let's go back to our roots to see if we can find the answer. If Haiku were to follow in BeOS' footsteps, it be a desktop OS indeed, but with a specialty in the handling of high bandwidth media; especially audio and video. I think it could be again, and with much greater success this time around.

Major audio developers who were sick of Windows were lining up to write software for BeOS at the time of the great "focus shift." Be, Inc. had many of the top names in audio right at the brink of porting important, serious stuff to the platform when Be pulled the rug out due to financial trouble. If a respected, growing, open-platform OS were available that offered good hardware support and a strong base of dedicated volunteer developers (showing it has a future), I'm confident audio companies would begin lining up to create products for it.

At its peak, lebuzz.com, which focused on BeOS audio development, was getting over 1,000 hits a day, and I believe a lot of the enthusiasm for using BeOS (Haiku) for audio is still there, ready to be tapped. Just the existence and hope of Haiku and its growing prospects of successful emergence have already kindled fires under the boiling pots of a number of audio developers, at least one of which has interest in commercial development for the platform. Hobbyists have also been dusting off promising old BeOS audio projects and begun working on them again. Chatter in the e-mail groups is also noticably audio-related again.

And there's TuneTracker. Without overstating its role, it's probably safe to say that TuneTracker has established a certain notoriety for BeOS among audio people. Maybe we can build on that by trying to give Haiku an audio focus, as BeOS had; getting the media kit in great shape, and putting strong emphasis on driver support and enhanced audio capabilities.

At least, it's something to consider.


Re: Kansai Open Source Forum: Day 2 Report

Dane wrote:

Let's go back to our roots to see if we can find the answer. If Haiku were to follow in BeOS' footsteps, it be a desktop OS indeed, but with a specialty in the handling of high bandwidth media; especially audio and video. I think it could be again, and with much greater success this time around.

While computing has changed a lot since the BeOS days, I also think that Haiku could still have future potential in applications such as those where BeOS excelled. But in reality, it is hard to predict when and how Haiku will end up being used in the future. In my humble opinion, how fast and how far Haiku can go will greatly depend on how many BeOS/ZETA developers feel strongly enough about the project to dive into the Haiku code to actually make things happen.

I find it a bit puzzling that there is a considerable pool of developers out there that could be committing code to the Haiku repo, but for some reason choose to stay on the sideline and wait. Of course you can't force them to write code; it's their skills and time after all. But this (at least seemingly) prevalent wait and see attitude leaves Haiku with a small dev force that, while very capable, dedicated, and resilient, is limited in resources. Fortunately, this last year has seen the return of a few of these devs, and hopefully we can see more and more getting their hands dirty in the not so distant future. The core devs could certainly use any help they can get. :)

Anyway, this is getting a bit off-topic... :)

With regards to the desktop focus, I think the focus is right and I don't have second thoughts about it. I just find the classification "desktop computing" sort of misleading. Does it mean that laptops and handhelds or tablets our out? I frequently think that perhaps "personal computing" (as opposed to enterprise or business computing) may be a better definition for the sort of user segment that Haiku would eventually target.

TT is one of those rare BeOS success stories that I always bring up whenever I have a chance (you should be proud of it!). As you say, media apps would certainly make a nice niche for Haiku. But then, the media kit and Haiku in general needs to be up to par to the needs of today's computing, and that requires manpower. So while we can realize the potential, we hit the issue of limited resources again, an issue that can only be solved or at least alleviated by the commitment of those BeOS/ZETA devs out there with the skills to contribute code to the Haiku base. Every bit counts.

Re: Kansai Open Source Forum: Day 2 Report

Here just to express my appreciation for your work Jorge.
Thanks for such a nice report.

Re: Kansai Open Source Forum: Day 2 Report

Hi Koki,

thank you very much for this detailed report, I enjoyed reading it very much. Sorry I couldn't come with you.

I guess everyone, especially the developers, ask themselves the question from time to time you pose at the end of your report. Personally, I believe it is a mix of many things. I can try to give a little insight into my own motivation.

I only use Macs occasionally, but I do use Windows and Linux (Ubuntu) regularily to get things done. Windows only recently, to edit some AVCHD footage. But I still use BeOS as my main OS, I do all coding, websurfing, emailing, chatting, music and image editing on BeOS. I think the software which I use for this on BeOS is superiour compared to the software on for example Ubuntu for the features I use. Especially editing code. That being said, I am very pleased and impressed with Ubuntu. It does many things very right. Sadly, at the same time, it's developers can't seem to get some stuff working properly: automatically selecting the correct screen resolution - or worse yet - letting you configure the correct resolution yourself is still not working smoothly for me.

Using Windows is also a very mixed experience, at least for me. I can't believe that Microsoft has not managed, with all their resources, to give Windows a completely unified look. Dive a little under the surface and be presented the NT4 look... that tells you something about how important the "whole system" approach is for Mircosoft. And already a few days after installing, I am getting the first problems (I won't even go into what a pain "installing" was). One of my CPU cores is running at almost full load, the whole time, and this damn Vista won't tell me what process is using up this CPU! I feel like I don't have power over the machine. The next disappointment is software. Most of the time, the feature I was looking for does not work as advertised, if at all (for example AVCHD editing). Another thing about software is all the distracting bling. There is just too much moving on the screen. I feel like my workplace is cluttered with advertisment.

Anyways, with Linux and Windows, I have two systems which *can* work nicely, but if things need fixing, I feel completely powerless. They are both a huge mess internally, that's my impression. And from what I read, I don't think MacOS X is much better under the hood.

With Haiku, I have the hope of a sane, clean system, which lets me run my beloved Pe, Beam, Vision and other software, on modern hardware. It being open source means we can fix any issues and we can develop new features or refine any aspect of the OS. The other Haiku developers all seem to care about the "whole system" thing very much, and so I think Haiku will be great, R1 and anything beyond. The main feature of Haiku may be "potential" right now, a promise of a lean, no bloat, no non-sense system. But even beyond that, we do have great software right now, which we will be able to run on Haiku reliably, eventually. So it's not all just dreams and no getting work done. My hope is that with our first alpha release, we will attract new developers, both for applications and the OS. Haiku should be a solid foundation, at least no one can take it away from us.

Re: Kansai Open Source Forum: Day 2 Report

Hi Stippi,

stippi wrote:

thank you very much for this detailed report, I enjoyed reading it very much. Sorry I couldn't come with you.

I'm glad you enjoyed the report. And please, no need to apologize. :)

Thanks for your input; it is quite insightful. I actually agree with many of your points. I also share your hope for a lean & mean OS that makes sense (like BeOS was), and believe that Haiku has the potential to eventually provide the solid foundation that you mention.

The good thing about doing trade shows is that it gives you the chance to interact with the people that could potentially have an interest in your product (that would be Haiku in our case). This interaction can be very valuable, as it allows you to see things from the side of the fence of the people that you are trying to target, and you can learn a lot from that.

So while I understand your point that there are people that use BeOS as their main OS, I am also very cognizant that today BeOS/ZETA is not a viable alternative for many, and the feedback that I get at trade shows only reinforces that awareness. Being that Haiku is not yet ready for end users either, and that it will be some time before it is, for these people Haiku can only be a (long term) future proposition.

That's why talking about the initial motivations for the project, the history, and present status (all very important information), and I do the hands-on Haiku demos, I tend to put a little emphasis on the future potential of the OS, because I think that's what could actually make it an enticing proposition for those who can see beyond the "but I can't watch YouTube videos with it!" shortsightedness. :)

Re: Kansai Open Source Forum: Day 2 Report

Live audio and video at gigs. Simply put but that's it really. Low latency, fast respondse, intuitive UI.

another example could be, you come home from work (whatever OS using) and you can play and be creative with Haiku.

Re: Kansai Open Source Forum: Day 2 Report

Once again, another enjoyable and as Stippi said detailed report on Haiku at Kansai. It's an interesting question the focus of the OS and we all seem to have similar visions of where we want it to go.

Coming purely from my 2 cents, I work as a Sys Admin on Windows boxes all day. I can see their value, as they get the job done and can't see any of this software being replaced, particularly on a user's desktop. Business users don't embrace change, which is why for years I've wanted something to plugin and replaced Exchange at the backend, so the users can happily keep using Outlook.

At home, I have XP (I deleted my Vista), OS X and Ubuntu linux, as well as Haiku running natively on another box, and unfortunately, no luck running Haiku natively on my Sony VAIO UMPC (the video spacks out), but as stated, it's pre-alpha so you can't expect the earth. My Windows machine is used purely for games ie. the ones from the big publishers. I can't see Haiku competing in this space and I don't think we'd want to.

I mainly use my OS X machine for all the video editing, podcast making, etc ie. the creative stuff, as well as general web surfing. To be honest, getting around the system could be better and I personally believe that OpenTracker is the best file navigator out of them all.

Ubuntu works pretty well, as long as you don't start playing with things and then you get yourself into trouble. I've had issues with resolutions too. Like if I don't have my monitor on, it will only display 800 x 600 and it can't be changed later. So, it means if I reboot this machine, I have to turn my monitor on first. Stupid eh ?

So, where does that leave Haiku ? Well, Haiku as I see it wins, 1) it's small footprint. Sure you can get small versions of Linux but nothing with the grace of Haiku. 2) it's speed. In terms of speed, when I first used BeOS, I had no idea that an Intel processor could work that well. 3) it's filesytem. OpenBFS has a number of advantages, I love the fact I can start ripping a CD (dragging and dropping it's cda files to the desktop) and before they're finished I can access them. Try doing that on anything else and it'll make nasty noises at you :) This is where also the audio low latency comes into play. 4) it's open source. Developers like this, however most users don't care. But this is catch 22, without developers you'll have no users.

For Haiku to compete in today's world, it needs a modern web browser that supports the modern plugins such as Flash and Java. Having Flash on Haiku would also open the door for a number of games, sites and applications that are written in ActionScript. By the way, here's a simple flash game I started working on called Haiku Match using Stippi's cool icons - http://sikosis.com/flash/games/haikumatch/