Icaros 1.1.2 disponible

Hi, first of all sorry if I write in english. I can understand français (I’ve studied it a little at school, but it was sincerely too long ago), but neither I can speak, nor I can write at it.

Then, thank you for your nice comments about Icaros Desktop: when I initially started this little project, I couldn’t believe it would go so far. It has already surpassed my initial expectations, and now I can reason about it in another level of “magnitude”.

This evolution reflects (and I’m sorry the article on obligement doesn’t seem to have noticed or completely understood it, but maybe is only a language barrer here) the one that whole AROS has in the last 24 months.

AROS has started in 1995 and slowly progressed in the first 10 years of its life with a big idea in mind, but a pathologic lack of manpower. Aaron had hard times trying to motivate people and the only things that allowed the project to survive were the fact that AmigaOS APIs were well documented, and the few people involved strongly believed, al least for the whole life of their involvment, in the AROS goals.

AROS started being ‘promising’ in 2001-02 when it got the initial shape that led to what it currently is. The problems that still remained were the lack of support from the Amiga community (which was still bound to what ‘the Bills’ promised) and the fact that, without a clear vision of the project goals and potential, the code has always been written for fun, in a way it “should” work. lack of real applications that would stress the AROS reimplementation of the API kept in largerly untested: yes, you could run little test code programs to test its functionality, but the real world is just another thing.

After 2005, the project started to grow in a exponenial way, at least in the AROS perspective, with more hardware supported and many apps ported. It started getting a ‘real OS’, facing real problems. Now we are in the middle of this trip: the more apps and pieces are coming to AROS, the more its flaws and mistakes emerge. Final users would perceive this as a plague, but in reality it’s the exact opposite: addressing more bugs allow ironing them out.

And, best of all, having a barely usable OS – which is better than aving a not usable at all OS – let our userbase to grow. Yes, people is still using AROS mainly for testing it, but at least Icaros proves there are the basis for a daily usage of AROS (well, I have a Icaros machine and I often use it). Users and real apps are a good source for motivation, and I can see that on Aros-Exec and the devs mailing list. The spirit that aims AROS developers has slightly and positively changed: now we have real expectations to fullfit and a OS we can be proud of. And I am more optimist about the whole project.

My best regards,

p.bes

[code]Auteur: Daff Posté le: 9/7/2009 14:00:26

Paolo : I’m the author of the article on Obligement. Say me if you noticed some errors, I can correct them.[/code]

Hi Daff. There are no mistakes in your article, everything you said is plain right. I could only see some weak comprehension about the emulation side of AmigaOS classic applications, that in my opinion should be explained a little better:

1. AmigaOS 4’s and MorphOS’ direct integration of 68K code simply can’t be done on AROS, due to different endianess of processor. This difference is a big issue, because you can theoretically run the app alone in a sandbox and provide swapping of bytes inside of it, but then the app would “talk” with other tasks (being them x86 or m68k ones) and at that point it would be impossible to understand if the contents of two bytes should be either swapped or not. This problem does not arise, obviously, on PPC architectures, but if AmigaOS 4 or MorphOS will be ever ported to x86, they would face exactly the same issue.

2. AmiBridge doesn’t enclose the m68K in a box much more like than Windows 7 does for older Win application in its XP-mode (well, I should sue Microsoft for having stolen my idea, but maybe Apple would do the same with both of us… 😆 ). It’s not true the app will stay just there. You can run a m68K app with AmiBridge and then save your data on the AROS filesystem. I have also a beta here with clipboard sharing enabled, but I didn’t include it in Icaros 1.1.2 because I had some speed problems with this newer version of Janus-UAE, and Oliver Brunner is still working on it. On the other side, I can also add that currently our integration is the one that works better with classic applications: while on AOS4 and MOS the integrated emulator some times fails because of the lacking of real Amiga hardware, our one provides exactly the same environment and can dramatically speed up oputput, thanks to the JIT compiler. I don’t think that running an app on a 1 GHz PPC machine with the other Amiga flavours would be faster than running them on Amibridge on a 3 GHz x86 processor.

If you run an Amiga App “à la Amibridge” on a PPC version of UAE, you wouldn’t get the jit compiler, thus slowing down everything. So, as you can see, with AROS you loose some “natural” feeling, but you gain in reliability and speed. And this is only the beginning. Soon you will be able to treat the AmigaOS windows inside AROS exactly in the same way of the native ones: you will still have UAE running in the background, but you’ll just not see it. Look at o1i’s blog for better informations, or maybe this screenshot. You’ll see amiga apps integrated with clipboard sharing in action. 🙂