aranym_history_1
I tried to publish a small article about ARAnyM history many times. But I never started to write it. Now I think it could be a quite nice first article on AST. In this article I'm describing my view and my memories of the beginning. I hope that Phanatic, Joy and STan will read it some day and add some comments to it. Some parts of ARAnyM history are documented in our mailing lists but for many users they aren't comprehensible because the beginnings were in strange language for them. I hope this article will be more comprehensible :-)
I remember the time when I started to think about some emulator
solution. For many years I was happy user of Atari ST 1040fm (ehm, not
so happy, I had many troubles with my first one) and because I hadn't
access to Internet and I played only with
GFA basic, I
didn't need more than it. In 1997 I came to Internet and I saw that
Atari is still alive (yes, at that time many sites and a lot of new
applications existed). At that time I was using my Atari only
occasionally and my primary platform was my IBM PC compatible toy. My
first reaction was purchase of IDE harddisk to my old Atari and the
next year I bought MegaSTE. The next step was a connection to Internet.
The cheapest one was through null modem serial cable to small i386
server (the first Atari SH which hosted Czech FTP and newsletter) in my
room in Prague. The first "implementation" was only telnet to Atari SH
through PPP :-) But with 4 MB of RAM I didn't expected much :-) and
also didn't need more :-)
Yes and now BUT. But then Thomas Berndtsson came with
Debian MiNT. Yes, I knew SpareMiNT,
but as Debian GNU/Linux user I knew advantages of dpkg against RPM. And
with MegaSTE I had only Motorola 68000, so SpareMinT wasn't important
to me. Debian MiNT was. I tried to help Thomas several times but with
MegaSTE I could be only tester not more. I needed some better machine
to help with Debian MiNT. And better machine like Falcon or even some
clone? I hadn't enough money. But I had my PC and I used
STonX quite often. And the next
"but", STonX seemed dead at that time (it was big surprise that STonX
revived during my thoughts).
At that time I already knew Joy (Petr Stehlik) and STan from Jays
(Standa Opichal) from some Czech activities. I also knew Phanatic
(Ctirad Fertr) from some discussions. During
Atariada we (I, STan and Phanatic)
discussed possibilities and finally at 23rd May 2000 I declared in
Czech Atari mailing list that we are trying to decide which possibility
is the best one. What we had?
- Virtual machine - at that time I defined it as a "perfect" emulator and I wasn't happy from this solution (I wrote that it is only a way how people are going to Linux and later years showed that it was true, but only partially).
- Ctirad's Ax86 - Ctirad came with something more closer to hardware - the emulation layer with drivers on x86 platform, without some x86 OS. This needed a lot of work and later showed as impossible.
- STan's like oTOSis - Linux with m68k emulation and GEMDOS support and X server with VDI layer and Window manager AES like.
And the reaction of Joy?
At first in Czech: "Dulezite je rici, ze "dokonaly" emulator ST nikdo nepotrebuje. Kazdy by chtel dokonaly emulator alespon Falcona, aby mohl zustat u Aniplayera. A tady je uz temer jasne, ze takovy emulator "nikdo" "nikdy" nenapise - je v tom totiz nejen 68030/68882 (vcetne MMU a cache), ale taky DSP a Videl - a to uz je zrejme moc velky orech."
And one short English translation: "It's important to say that nobody needs a "perfect" emulator of ST. Everybody wanted perfect emulator of Falcon, because every body wanted to stay with Aniplayer. And now it is already almost clear that this type of emulator "nobody" "never" will write - it is not only 68030/68882 (with MMU and cache), but also DSP and Videl - and it is evidently big problem."
It wasn't the first time when Joy said that something wasn't possible :-) And not the last time :-) When you need something impossible, just ask Joy for his "impossible decision" and you'll see somebody (probably Joy) will do it :-)
So, now we are in the situation just right before real technical discussion :-) I found in my mailbox archive the oldest e-mail about it (from 19th May 2000) and it was about improvements of STonX. To be continued...
Few comments:
Well, to be honest, I think that the STan's idea was the best for the futur.
It's like Nocrew's OSIS ( http://osis.nocrew.org/ ).
Of course I use and support Aranym but we are confined in a Emulator( I know, Aranym is not really a emulator)... no direct access to the hardware, impossibility to have native softwares, etc..
PS: congratulation for this webpage, I like the idea :)
zorro
If we created STan's idea, we would have been depended on one platform. And Atari user would migrate to Linux and use ARAnyM only as some bridge for only a few applications (yes, actual situation is partially similar, e.g. Ctirad migrated), because many original applications aren't ready for STan's idea, I think. You can see it in MagicMac. Lots of its users migrated to Macs... And don't forget I wanted to help with Debian MiNT and FreeMiNT ;-)
Milan
Hi!
Although I still like the idea of oTOSis way very much I must agree 100% that our decission was the best we could do in 2001. Hmm, I think it actually evolved to the current shape. I believe that if anyone had brought another things to the CVS we might have ended up having something of completely different nature. However if you have asked me "Why this way?" that time I wouldn't know the answer. Let me explain why in a little more detail.
We ( JAY Software) have been trying to
develop GEM programs at that time. None of us was experienced enough to
dig in depth in the operating system internals like in the VDI or AES
implementations not to mention operating system kernel. Speaking of
myself I would simply say I was too young for such things.
Our platform suffered from lack of working, reliable and modern HW at that time. Moreover the Milan II had been sent to ice rumorously due to HW chip producer cuts. So there was no option for me to upgrade my Atari and go on with the development. My F030 ran at @40MHz, but only in 256 color depth. That effectively made me using MagiC for absolute most of the time which makes one not to use and develop open-source essentials and tools at all. As I was making GEM apps my desire was much simpler: All I needed was to have MagicMac/MagiCPC quality open-source emulator for me to help with the GEM development.
At the time Joy was playing with booting to desktop in Fanda I was using STonX booting to FreeMiNT. Well, I wanted to have something more. True color graphics, host filesystem access, networking support (which wasn't present even in MagiCPC that time). As the development went on my focus were those pieces. I must say that I admire Joy, Milan, Andreas and others that have worked on all those low level things like CPU and HW emulation parts.
I must say that during those ARAnyM years my knowledge of our OS internals went somehow up (of course not only due to ARAnyM). Today I find hacking VDI, AES and FreeMiNT kernel to be my daily routine.
Because of all that progress I/we made I now know for sure that in 2001 it was not the time to go oTOSis way simply because an absolute lack of open-source GEM compatible OS (here I mean all the parts... kernel, GEMDOS, VDI, AES). We have chosen a way that made us to contribute to the scene effort instead of making people to essentially move to new platform. We are trying hard to keep the effort to contribute and improve the software that works on our good old Atari HW and this is intended to be the spirit of the whole project. Any other way would make more people go for their Macs or PCs instead. My thanks would go mainly to the people around the operating system development. Namely Frank Naumann, Johan Klockars, Henk Robbers and many others who jointly and continuously improve our environments.
Who knows what would bring us the future... You know, when you have VDI and AES implementations as well as the support libraries for applications it is not that difficult to use another kernel if you put some effort into it. I would for sure like a MacOS X like system for my GEM apps. But again, if I am about to go this way I would ask and think a lot to aviod people moving away off the platform.
STanda