An interview with Aaron Digulla who established the Amiga Research Operating System(AROS) in 1995.
Hello Aaron, I really appreciate you taking the time to do this interview. Could you give us a history of how you became interested in Amiga's?
I started with a C64 in a big warehouse. At the time, I couldn't afford one, so I had to stand all day (or a large part of it) in front of a shelf, with my neck craned because the monitor was too high up and the keyboard was too low.
My second computer was an A500. I've always admired the open attitude and the simple complexity of the C64 and found the same features in the Amiga. With time, I've bought some extensions and finally an A1200 whichI even built into a tower case.
Unfortunately, at that time (1994), it was already obvious (even if many people didn't want to realize it), that the Amiga was doomed. For a development job, I was given a PC to take home and the raw power of Intel was beginning to outweigh the beautiful design of the Amiga.
Today, I'm mostly using x86 PCs with Windows at work and Linux at home.
You are known for establishing the Amiga Research Operating System (AROS). Could you explain how and why you started the project?
Even before Commodore (remember, the company which sold the Amiga a long time ago?) folded, there was a lot of discussion about changes, modifications and improvements to the Amiga Kickstart. Most discussions went like this:
1. "We must have feature XYZ"
2. "you can't do this, it would break blabla. And it's impossible."
3. "How about feature ABC? That would be great"
4. "We must have feature XYZ" etc.
I wanted to create a playground where developers could create patches for the Kickstart ROM, a place where changes could be implemented to measure the impact of the changes on the system instead of discussing what might or might not happen if when why what.
What was it like developing AROS in the beginning?
It was the typical Amiga thing: A lot of enthusiasm and a lot of "you can't do this". As always, the expectations won over the offers to help by several magnitudes. ;-)
What was the greatest struggle you had in the beginning?
For many years, AROS simply lacked the manpower. Over several years, we had three or four active developers, each did something and then faded back into the noise of the 'net.
That made it impossible to tackle big problems like DOS Packets. Also, I made my usual bunch of design mistakes (Exec based DOS handlers, for example).
But there was also a lot of stuff we did right. The build system was very solid and could cope quite well with all the changes which we implemented in the last 12 years. There was a core set of C header files which defined a lot of macros (some of them so big that some compilers ran out of memory trying to build them up) which allows us to port AROS to new systems with very little effort.
People have ported AROS to a new system within a day or so. After that, they could compile and Exec would work. Of course, writing the IO drivers takes more time but the code is very simple to do.
How many developers did you have helping you in the beginning? What were their names?
Oh dear. There are currently 370 developers registered and many of them have contributed a lot to AROS. Who to mention and who to annoy?
The first developers were Iain Templeton (from Tasmania, even!), Sebastian Rittau, Lennard voor den Dag, Morten Holm, Tommy Johansson, Hans Van Ingelgom, Martin Steigerwald, Peter Bortas, Johan Alfredsson and Kars de Jong (from the CVS log; my memory for names is ... uhm ... well).
Did they volunteer to help you?
Luigi went around the globe to break their legs so they had to sit in front of a computer and type ;-)
Honestly, I never really was the big PR guy and it was astonishing how many people noticed my project just because of the AmiNET releases and word of mouth.
Do you still keep in contact with all of them?
No, I'm not the type, really. Actually, they keep in touch with *me*, and every now and then, their names pop up in the commit mails.
What was the community surrounding AROS at that time like? Did you notice a lot of support from Amiga fans?
At the time, the community had already learned to distrust. Companies have to pay salaries and the bills, so some very unpopular decisions were made. Therefore, the reactions and the support were quite varied. Some saw AROS as another threat to the Amiga, ruining more companies who supported the system, others were ready to send anything which even smelled a little bit commercial to hell and thought AROS as the Amiga savior.
Unfortunately, we didn't manage to attract many of the professional Amiga developers, so it took AROS many years to be able to compile an Amiga application. By that time, Amiga was quite dead and most Amiga fans had become PlayStation(tm) fans, or Mac fans or whatever. Some (like me) had even given in to the delicious but poisonous lure of the x86 PC. Smokers and drug addicts, you know what I'm talkin' 'bout.
So all in all, there was a lot of "please help the Amiga" kind of support but little in terms of development time. The few developers which we had did great work but there is only so much you can do with a few people who work on such a project as a hobby.
Still, I believe, we all had a great time and it is fun to watch a program grow, suddenly beautiful icons pop up, Eric W. Schwartz donates a hot (*drool*) logo for out website, some designers gave my crappy HTML layout the well deserved beating and suddenly, we have a bootable live CD. Life is great :-)
What do you feel is your greatest accomplishment while developing AROS?
I managed to keep a cool head in times of crisis. Once, a well renown Amiga professional found some of his code in our source and wasn't really happy with it. The situation could be resolved without lawyers which I think is the only way to handle this.
At another time, many developers were about to leave the boat because someone posted endless ... uh ... rants on our mailing list rambling about whatnot. It wasn't trolling, this guy just had a completely different view of the world.
Sometimes, email is just the wrong medium but when emotions are high, it's much more simple to cool everyone down with a carefully worded mail than with anything else.
Could you describe some of the work you have completed for AROS?
I designed the build system (MetaMake), I'm in charge of the version control system and mailing list and I iron out the eventual kink in the real world interface of the project. I also try to be the unprejudiced arbiter when something important has to be decided.
There is probably also some code in AROS which I wrote but I'd have to dig out the CVS logs to find out ;-)
Are you currently still involved with the development of AROS?
No. I would have to look into the Subversion log to see when I committed something last time. I'm happy to see that the team can do it's work without me.
In what way and do you plan to continue to be involved with AROS in the future?
For the next years, I see no changes coming. Unless someone steps up and wants to replace me as project leader but if there was someone who wanted this ... well ... job, I guess he or she would have contacted me by now :-)
Are you happy with the progress of AROS up to this point?
Yes. My rants above might give a different impression but I don't think that more could have been achieved. Being in "charge" of on OSS project gives an interesting contrast to the usual high-pressure projects which I'm paid to enjoy. With OSS, you cannot force things. Either someone does it or it simply doesn't happen. It also helps to separate important from unimportant: If nobody does it, then it's probably not quite the issue some people would like it to be.
People have expressed that the AROS project needs to be organized better. Do you think the project needs better organization? Please explain.
Sure but who is willing to invest the time? Step up or remain silent forever.
Right now, we're doing most of the stuff ad-hoc as we need it. Of course, that means some lost energy but creating processes and responsibilities also doesn't come for free. Many of the big projects in the industry fail because the management part of the team takes over and brings everything else down.
In your opinion, what feature could greatly improve AROS at this point?
We don't have a compatible system to handle DOS packets and a lot of Amiga code depends on that. Especially the stuff which isn't games :-)
And it would be much more simple to maintain project quality if we had an automated test system.
Currently, I'm working as a Java developer and I've simply stopped writing code which isn't or can't be automatically tested. It's just not worth the few minutes which I could theoretically save if I wouldn't have to fix all those bugs and problems later.
Taking a step back, being able to compile AROS in AROS is probably the next key feature which will bring more developers to the system. Eating your own dog food is always a great way to improve the taste.
What application would you like to see available for AROS?
Well, I've always used the Amiga for games and to edit code. XDME has already been ported to AROS (even though there are a few bugs left to iron out *ahem*) and for games, there is AROS-UAE, the UAE port for AROS. For me, that's all I personally would need.
Have you been or are you currently involved in any other projects besides AROS? If so, what?
I'm losely involved in the Eclipse project (www.eclipse.org), Maven (maven.apache.org) and I'm planning to start a twin of AROS for Java (a playground for experiments on the JDK to improve the language) next year and I'm toying with a lot of stuff related to Linux.
Do you use any other operating system? If so, what type?
At home, I have a computer network with a server and several PCs running SuSE Linux (10.0, 10.1 and 10.2).
At work, I'm stuck with Windows but when Firefox and Eclipse hide the desktop, it's bearable.
What were your favorite things to do with your Amiga?
Develop software, render 3D images and play games.
How many Amiga's have you owned and what models?
Two, an A500 and an A1200.
Do you currently own any Amiga's?
No. I gave my A1200 and the 12'000 floppy disks away and I don't think I'll be ever going back. The Amiga hardware was great to look at but never that important for me.
Do you use Amiga OS4 or MorphOS at all?
No. Michael Battilana from Cloanto was so nice to give me a copy of the Amiga Forever DVD (which I can really recommend! It contains a five hours of videos like the Amiga launch party with Andy Warhol and the infamous Amiga Deathbed Vigil Video PLUS an Amiga emulator with all the kickstart ROMs!).
No AROS on the DVD, yet, but Michael offered us some space on the DVD when we have something which we not only can be proud of but which is actually of some real use for the buyers. Maybe we'll just put something on it which we only can be proud of. We'll see. If someone finds the time.
What are your thoughts on the Amiga community these days?
I think we're seeing the end of the valley of tears. All the whiners are gone and now, a lot of the old fans, supported by jobs and experience, come back to have a second look at a source of joy from their past. I think the future of the Amiga has only just begun.
Is there anything you would like to add?
Of course no interview is complete without the Request For Help. Can you read English and write in your mother language? Why not help translate the AROS website or the developer documentation into your language? It's simple to do, just takes as much time as you can spend and might help someone out there to join the team and work on AROS.
Or do you long for the Good Old Times? Make a dream come true and help us build the fastest Amiga that ever was.
Thank you Aaron. Again, I really appreciate you doing this interview for The AROS Show!
Hello Aaron, I really appreciate you taking the time to do this interview. Could you give us a history of how you became interested in Amiga's?I started with a C64 in a big warehouse. At the time, I couldn't afford one, so I had to stand all day (or a large part of it) in front of a shelf, with my neck craned because the monitor was too high up and the keyboard was too low.
My second computer was an A500. I've always admired the open attitude and the simple complexity of the C64 and found the same features in the Amiga. With time, I've bought some extensions and finally an A1200 whichI even built into a tower case.
Unfortunately, at that time (1994), it was already obvious (even if many people didn't want to realize it), that the Amiga was doomed. For a development job, I was given a PC to take home and the raw power of Intel was beginning to outweigh the beautiful design of the Amiga.
Today, I'm mostly using x86 PCs with Windows at work and Linux at home.
You are known for establishing the Amiga Research Operating System (AROS). Could you explain how and why you started the project?
Even before Commodore (remember, the company which sold the Amiga a long time ago?) folded, there was a lot of discussion about changes, modifications and improvements to the Amiga Kickstart. Most discussions went like this:
1. "We must have feature XYZ"
2. "you can't do this, it would break blabla. And it's impossible."
3. "How about feature ABC? That would be great"
4. "We must have feature XYZ" etc.
I wanted to create a playground where developers could create patches for the Kickstart ROM, a place where changes could be implemented to measure the impact of the changes on the system instead of discussing what might or might not happen if when why what.
What was it like developing AROS in the beginning?
It was the typical Amiga thing: A lot of enthusiasm and a lot of "you can't do this". As always, the expectations won over the offers to help by several magnitudes. ;-)
What was the greatest struggle you had in the beginning?
For many years, AROS simply lacked the manpower. Over several years, we had three or four active developers, each did something and then faded back into the noise of the 'net.
That made it impossible to tackle big problems like DOS Packets. Also, I made my usual bunch of design mistakes (Exec based DOS handlers, for example).
But there was also a lot of stuff we did right. The build system was very solid and could cope quite well with all the changes which we implemented in the last 12 years. There was a core set of C header files which defined a lot of macros (some of them so big that some compilers ran out of memory trying to build them up) which allows us to port AROS to new systems with very little effort.
People have ported AROS to a new system within a day or so. After that, they could compile and Exec would work. Of course, writing the IO drivers takes more time but the code is very simple to do.
How many developers did you have helping you in the beginning? What were their names?
Oh dear. There are currently 370 developers registered and many of them have contributed a lot to AROS. Who to mention and who to annoy?
The first developers were Iain Templeton (from Tasmania, even!), Sebastian Rittau, Lennard voor den Dag, Morten Holm, Tommy Johansson, Hans Van Ingelgom, Martin Steigerwald, Peter Bortas, Johan Alfredsson and Kars de Jong (from the CVS log; my memory for names is ... uhm ... well).
Did they volunteer to help you?
Luigi went around the globe to break their legs so they had to sit in front of a computer and type ;-)
Honestly, I never really was the big PR guy and it was astonishing how many people noticed my project just because of the AmiNET releases and word of mouth.
Do you still keep in contact with all of them?
No, I'm not the type, really. Actually, they keep in touch with *me*, and every now and then, their names pop up in the commit mails.
What was the community surrounding AROS at that time like? Did you notice a lot of support from Amiga fans?At the time, the community had already learned to distrust. Companies have to pay salaries and the bills, so some very unpopular decisions were made. Therefore, the reactions and the support were quite varied. Some saw AROS as another threat to the Amiga, ruining more companies who supported the system, others were ready to send anything which even smelled a little bit commercial to hell and thought AROS as the Amiga savior.
Unfortunately, we didn't manage to attract many of the professional Amiga developers, so it took AROS many years to be able to compile an Amiga application. By that time, Amiga was quite dead and most Amiga fans had become PlayStation(tm) fans, or Mac fans or whatever. Some (like me) had even given in to the delicious but poisonous lure of the x86 PC. Smokers and drug addicts, you know what I'm talkin' 'bout.
So all in all, there was a lot of "please help the Amiga" kind of support but little in terms of development time. The few developers which we had did great work but there is only so much you can do with a few people who work on such a project as a hobby.
Still, I believe, we all had a great time and it is fun to watch a program grow, suddenly beautiful icons pop up, Eric W. Schwartz donates a hot (*drool*) logo for out website, some designers gave my crappy HTML layout the well deserved beating and suddenly, we have a bootable live CD. Life is great :-)
What do you feel is your greatest accomplishment while developing AROS?
I managed to keep a cool head in times of crisis. Once, a well renown Amiga professional found some of his code in our source and wasn't really happy with it. The situation could be resolved without lawyers which I think is the only way to handle this.
At another time, many developers were about to leave the boat because someone posted endless ... uh ... rants on our mailing list rambling about whatnot. It wasn't trolling, this guy just had a completely different view of the world.
Sometimes, email is just the wrong medium but when emotions are high, it's much more simple to cool everyone down with a carefully worded mail than with anything else.
Could you describe some of the work you have completed for AROS?
I designed the build system (MetaMake), I'm in charge of the version control system and mailing list and I iron out the eventual kink in the real world interface of the project. I also try to be the unprejudiced arbiter when something important has to be decided.
There is probably also some code in AROS which I wrote but I'd have to dig out the CVS logs to find out ;-)
Are you currently still involved with the development of AROS?
No. I would have to look into the Subversion log to see when I committed something last time. I'm happy to see that the team can do it's work without me.
In what way and do you plan to continue to be involved with AROS in the future?
For the next years, I see no changes coming. Unless someone steps up and wants to replace me as project leader but if there was someone who wanted this ... well ... job, I guess he or she would have contacted me by now :-)
Are you happy with the progress of AROS up to this point?
Yes. My rants above might give a different impression but I don't think that more could have been achieved. Being in "charge" of on OSS project gives an interesting contrast to the usual high-pressure projects which I'm paid to enjoy. With OSS, you cannot force things. Either someone does it or it simply doesn't happen. It also helps to separate important from unimportant: If nobody does it, then it's probably not quite the issue some people would like it to be.
People have expressed that the AROS project needs to be organized better. Do you think the project needs better organization? Please explain.
Sure but who is willing to invest the time? Step up or remain silent forever.
Right now, we're doing most of the stuff ad-hoc as we need it. Of course, that means some lost energy but creating processes and responsibilities also doesn't come for free. Many of the big projects in the industry fail because the management part of the team takes over and brings everything else down.
In your opinion, what feature could greatly improve AROS at this point?
We don't have a compatible system to handle DOS packets and a lot of Amiga code depends on that. Especially the stuff which isn't games :-)
And it would be much more simple to maintain project quality if we had an automated test system.
Currently, I'm working as a Java developer and I've simply stopped writing code which isn't or can't be automatically tested. It's just not worth the few minutes which I could theoretically save if I wouldn't have to fix all those bugs and problems later.
Taking a step back, being able to compile AROS in AROS is probably the next key feature which will bring more developers to the system. Eating your own dog food is always a great way to improve the taste.
What application would you like to see available for AROS?
Well, I've always used the Amiga for games and to edit code. XDME has already been ported to AROS (even though there are a few bugs left to iron out *ahem*) and for games, there is AROS-UAE, the UAE port for AROS. For me, that's all I personally would need.
Have you been or are you currently involved in any other projects besides AROS? If so, what?
I'm losely involved in the Eclipse project (www.eclipse.org), Maven (maven.apache.org) and I'm planning to start a twin of AROS for Java (a playground for experiments on the JDK to improve the language) next year and I'm toying with a lot of stuff related to Linux.
Do you use any other operating system? If so, what type?
At home, I have a computer network with a server and several PCs running SuSE Linux (10.0, 10.1 and 10.2).
At work, I'm stuck with Windows but when Firefox and Eclipse hide the desktop, it's bearable.
What were your favorite things to do with your Amiga?
Develop software, render 3D images and play games.
How many Amiga's have you owned and what models?
Two, an A500 and an A1200.
Do you currently own any Amiga's?
No. I gave my A1200 and the 12'000 floppy disks away and I don't think I'll be ever going back. The Amiga hardware was great to look at but never that important for me.
Do you use Amiga OS4 or MorphOS at all?
No. Michael Battilana from Cloanto was so nice to give me a copy of the Amiga Forever DVD (which I can really recommend! It contains a five hours of videos like the Amiga launch party with Andy Warhol and the infamous Amiga Deathbed Vigil Video PLUS an Amiga emulator with all the kickstart ROMs!).
No AROS on the DVD, yet, but Michael offered us some space on the DVD when we have something which we not only can be proud of but which is actually of some real use for the buyers. Maybe we'll just put something on it which we only can be proud of. We'll see. If someone finds the time.
What are your thoughts on the Amiga community these days?
I think we're seeing the end of the valley of tears. All the whiners are gone and now, a lot of the old fans, supported by jobs and experience, come back to have a second look at a source of joy from their past. I think the future of the Amiga has only just begun.
Is there anything you would like to add?
Of course no interview is complete without the Request For Help. Can you read English and write in your mother language? Why not help translate the AROS website or the developer documentation into your language? It's simple to do, just takes as much time as you can spend and might help someone out there to join the team and work on AROS.
Or do you long for the Good Old Times? Make a dream come true and help us build the fastest Amiga that ever was.
Thank you Aaron. Again, I really appreciate you doing this interview for The AROS Show!
posted by Paul J. Beel at 10:30 PM
11 Comments:
Thanks, Interesting Read!
6:53 AM
I've enjoyed reading this interview. Thanks!
8:02 AM
I'm glad everyone is enjoying it!
8:41 AM
Its always nice to hear something new about AROS, and developers working on it :]
2:26 PM
Several years ago you introduced me to open-source programming by getting me involved with AROS.
I did very little for the project, but I've learnt quite a lot from it.
Thanks Aaron!
9:15 PM
"I did very little for the project, but I've learnt quite a lot from it."
great, now that you learned a lot from it, go out find 2 mates and all 3 of you can put someting back and help everyone else in the process, AND DO IT NOW
good on ya.....
1:03 AM
> Several years ago you introduced
> me to open-source programming by
> getting me involved with AROS.
Hello Bernardo! :)
4:45 AM
Nice to hear from Aron, our Holy Spirit =)
I hope AROS will rise higher than before these days, we got a lot of new people, ideas and so. DOS.lib code was spotted on a mailing list and with a a permission and guide of an author (interviewed here =) ) we can get it integrated (aAron, please, heat it up if you can - seems that others have no extra feeling to do it). And i can agree - the thing we maybe lost is a
just a little more strict guidance
and management. Not a schedule, but a detailed plan.
Thnx !
4:14 PM
I am glad this project is going usable, years ago i made my best works ever on amiga, then i got stuck in a mac and at last i bought two pcs, the laptop I using now and another at home. I see the end of this piii/900 close to the horizon and the idea to work on vista does not attract me too much. I really wish to see things as openoffice gimpinkscape and scribus (and hopefully vmware) on AROS that i think look much easier to handle for me than a linux, might be the desktop operating system alternative loth of lusers (me too despite all) wait for, who knows?
1:50 AM
Another good interview by Paul.
But at the same time, it's kind of a "downer" to see the originator of the project lose interest in Amigas (the physical entity) and sort of drift off the project (contrasting this with Linus and Linux) into something that's not as interesting (Java/Eclipse, etc), granted that this is my opinion, but I believe there's more intriguing technology out there.
Of course it's his life and his right, and he is who he is, and things do "happen". Can't do anything but wish him best of luck in his endeavours and hope to see him turn back at some point.
At the same time I have a feeling this (Aaron's public image) is perhaps related to a seemingly partial lack of organizational focus and future-bound direction for the project. I'm new to AROS - although I've been hearing about it for years, I never looked into it as seriously as these last few days - and I must say that thankfully the main site, aros.org, does seem to be quite well put together (informationally, organizationally, documentation wise) for such a small project (compared to other open source projects). So that means a big congratuations to the AROS team. But like I said about organization and future-bound direction, I didn't really have something jump out at me (I had to search for milestones and roadmap, and no prominent "manifesto" exists, other than the typical "about", which is very much anchored in 1995's "about")
At the same time I think fortunes will change when more audacious and forward looking leadership takes the helm. The question that will be answered will be: What is AROS's aim for the future?
I understand that what is happening right now is "catch-up", in the sense that it's the time when the OS is being built and trying to get any decent "modernity" in it, but future aspirations are always a super-magnet for incoming developers. It also helps focus and hone current directions, in lieu of the future.
Anyways, enough mumbling. Next step is to actually code something, coz talk is cheap. Again, congratulations for the team that has made AROS what it is today - more impressive than expected!
7:22 AM
To offer a rebuttal to my own post (one just before), I stumbled on this very relevant thread in mere minutes from my last posting!! (talk about fate at work????)
http://cataclysm.cx/2007/01/20/optimise-for-fun/
And to follow with a counter-rebuttal-rebuttal, my emphasis was on "aspirations", and not making AROS a "serious OS".
To be honest, I see things from this angle:
I *hate* Windows with a passion. As a software engineer I detest every piece of it (OK - not every - some are OK, but not most).
I quite like Unix and close derivatives - Linux offered a sanity to the x86 world, when Windows was the only other choice - no thanks.
Yet, I believe Unix is not good enough any more (X11 was cool at a time, but having had a taste of Apple's Cocoa and Core Image, Core Animation and other techs, X11 is crap. Apple's got the future in their pocket). Some of its (Unix) concepts are good, but overall, we need to break away from similar systems. We have to think different, because computers are changing, especially with the massively parallel CPUs and GPUs that are starting to appear everywhere. Computing power allows for more powerful, yet simpler languages, and thus different "songs" (programs) to be written about. New paradigms. Hardware is also becoming reconfigurable at quite a fast pace. It might even be possible that software becomes "hardened" again (think anti-machine and data-centric computing). Even UIs are going in for a major change now that we have multi-touch and multi-dimensional spatial sensing devices.
So I guess I was hoping (like a typical undying die-hard Amiga hopeful), that the Amiga part of AROS would have such futuristic aspirations. Sure, rocking the server market and taking a pie from Linux/*BSD would be great, or the embedded world from Symbian, etc, but that's not the future I meant - I meant future as in "it's not here yet", and who says *that* can't be fun? :-)
Enough talking to myself in this old old thread!
Cheers to all!
7:46 AM
Post a Comment
<< Home