Some time back in the year 1993, the situation for the Amiga looked somewhat worse than usual (doesn't it always?) and some Amiga fans got together and argued about what should be done to increase the acceptance of our beloved machine. Immediately the main reason for the missing success of the Amiga became clear: it was propagation - or the lack thereof. The Amiga should get a more widespread basis to make it more attractive for everyone to use and to develop for. So plans were made to reach this goal. One of the plans was to fix all bugs of the AmigaOS, another was to make it an OS of the 90s. AOS was born.
But what is a bug? And how should bugs be fixed? What are the features a so-called modern OS must have? And how should they be implemented into the AmigaOS?
Two years later, people were still arguing about this and not even one line of code had been written (or at least no one had ever seen that code). Discussions were still of the pattern "we must have ..." and someone answered "read the old mails" or "this is impossible to do, because ..." which was shortly followed by "you're wrong because ..." and so on.
In the winter of 1995, I (Aaron Digulla) was fed up with it and I posted an RFC (request for comments) to the AOS mailing list in which I asked what the minimal common ground might be. Several options were given and the conclusion was that almost everyone would like to see an OS which is compatible to OS 3.1 (Kickstart 40.68) on which further discussions could be based upon to see what is possible and what is not.
So the work began and AROS was born.
AROS' goal is it to create an OS which
The goal for the first final release is to make development easier and faster. For later versions its possible to increase the compability to AmigaOS if its necessary.
To reach this goal, we use a number of techniques. First of all, we make
heavy use of the Internet. You can participate in our project even if you
can write only one single OS function. The most current version of the
source is accessible 24 hours per day and patches can be merged into it at
any time. A small database with open tasks prevents all developers from
doing something twice.
Currently, there are 71 developers registered and a group of about 5 people commits code regularly. When I write this, the code is about 29MB, about 56% have been finished and 13% is in work.
We are working hard on finishing graphics, layers and intuition based on our new HIDD system. The HIDD system will allow us to put any kind of hardware below AROS with only very little effort for people who want/need/have to write a driver for a new piece of hardware. This work is mostly done by Nils Henrik Lorentzen and Stefan Berger. The main goal is to have a version of AROS which can open a screen, create windows on it and let the use manipulate them. Currently, X11 is used to create the windows and X11 also allows to manipulate them. The time-frame for this is May.
Michael Schulz has committed very promising code which can be compiled, saved on a floppy and then be used to boot a small part of AROS on a bare IBM compatible PC. When the new graphics system is ready, writing a VGA driver should take a day (if not less) and AROS should be able to open a workbench screen and windows on it. No workbench yet, though :-) AROS does already boot but it doesn't do much, yet.
The next step would be to finish the console.device (half of it is already finished) and put the input.device on top of the HIDD system. This should integrate X11 (or whatever is below AROS) nicely into the AROS system allowing to open a shell on an AROS screen. This will be done by Nils and Johan Alfredsson, who is also working on the Commodities. This should take about one month.
When the HIDDs for PC hardware are finished (mostly done by Michael Schulz), it should be possible to boot AROS from a diskette and use a shell to enter some commands. An FFS file system also seems at the brink of completion, so creating an Amiga partition and installing it on the hard disk will then soon be possible. No time frame has been set for all this, yet. It mainly depends on how much (correct) information books about PC hardware contain, how the code for Linux looks like and such.
Branko Collin is meanwhile working on improving the documentation (the web pages are not exactly ... "sexy" :-), fixing the many typos and writing documentation about how to write documentation.
Kars de Jong is working on the Linux/m68k port of AROS which allows to run binary Amiga software. Some simple programs already work. When not working on Layers, Stefan Berger tries to compile programs from Fish disks and AmiNET work on AROS.
Bernhard Fastenrath and Aaron Digulla work on merging AROS with UAE which should allow to make UAE faster and run native Amiga programs in an emulator when AROS is able to run standalone (that is, without the help of Linux). We will investigate the necessary steps and do tests during March. At the end of the month, the next step will be decided.
Przemyslaw Szczygielski and Claus Herrmann are looking at PPC support. The former for a port to LinuxPPC and the latter for making Exec on the Amiga work on PPC (thus using the power of the PPC to run Amiga software on an Amiga). Przemo hopes for some results in May and I must convince Claus to merge his code with ours :-)
What the future might bring, who knows. The main goals right now are making AROS work as a standalone system which can boot from a hard disk, writing some drivers for the HIDD system so we can see if everything really fits to where it belongs, rewrite the most complex part of the AmigaOS: Intuition, Graphics and Layers.
The future will hopefully see a commercial version of AROS (so that we can pay Amiga Inc. some money so they keep calm), a free version of AROS (just like Linux: free without support or commercial with support), many more HIDDs, a port of AROS to PPC.
But that doesn't mean that you can sit back, relax and watch our progress (which is now updated daily on http://www.aros.org/progress.gif). We need you !
Currently, more and more parts of AROS are being finished and we need some real-world code to test AROS. What code would that be ? Here are the criteria:
If you happen to have such a gem, then send it to:
Please make *absolutely clear* if you mind if your code is published with
the source code of AROS or not. We promise that we won't touch any
copyright notices (except for adding them if they are missing, so you get
your credit). It's just that you might not want that the code is revealed
for any reason which would be ok for us. We just need to know :-)
What everybody wants to know: What does Amiga Inc. think about AROS or Gateway 2000. I have spoken to Jeff Schindler, Bill McEwen and Petro Tyschtschenko. All of them have been quite positive about the project but the legal issues involved are tricky. In Europe, there is no legal way to attack AROS besides using trademarked names. Outside Europe, algorithms may be patented and since we are trying to emulate the AmigaOS as close as possible, we probably infringe some patents and that could make AROS vulnerable.
For now, we have agreed on this: We continue the development and tell everyone what we are doing but the source code is only accessible by people who register with us. If you think, you can help and want access to the sources, read the section about CVS on our WWW pages.
Besides that, we are thinking about ways which would allow us to publish the source of AROS just like the Linux sources without tripping on Gateways' toes.
Links: