GEM in 2018 ?

Atari talk, or the life and the universe and things. Just keep it clean!
User avatar
IngoQ
Site Admin
Site Admin
Posts: 653
Joined: Tue Aug 22, 2017 8:38 am
Location: Germany

Re: GEM in 2018 ?

Post by IngoQ » Sun Aug 12, 2018 11:50 am

arf wrote:
Sun Aug 12, 2018 11:45 am
Adimens (relational database, available for Atari, PC, FlexOS, all with GEM as GUI).
That's an impressive one... could be a nice platform for a information retrieval system...
Ingo :geek:

“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.” - Antoine de Saint-Exupéry

stephen_usher
Posts: 148
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: GEM in 2018 ?

Post by stephen_usher » Sun Aug 12, 2018 7:21 pm

The biggest question is. "How good and easy to program is the GEM API?"

If it's simple to create programs and yet advanced enough to give a large enough set of features for a modern application then replicating the API for modern systems might be useful.

Unless the original code base abstracts the graphical drawing to a large degree I doubt that the original code will be much use and you'd have to do a full reimplementation anyway.
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.

keli
Posts: 62
Joined: Tue Aug 22, 2017 1:34 pm

Re: GEM in 2018 ?

Post by keli » Mon Aug 13, 2018 12:50 am

I'm just going to leave this here:
https://github.com/kelihlodversson/pTOS

Still early stages, but it currently can boot up to this:
IMG_20180723_140008.jpg
IMG_20180723_140008.jpg (3.83 MiB) Viewed 194 times

Steve
Posts: 153
Joined: Fri Sep 15, 2017 11:49 am

Re: GEM in 2018 ?

Post by Steve » Mon Aug 13, 2018 2:12 am

keli wrote:
Mon Aug 13, 2018 12:50 am
I'm just going to leave this here:
https://github.com/kelihlodversson/pTOS
WOW this is amazing!! I wonder why I haven't seen this elsewhere. Would it be able to run Atari compatible apps eventually? Or would it need it's own specifically ARM compiled apps?

Petari
Software Moderator
Software Moderator
Posts: 538
Joined: Tue Nov 28, 2017 1:32 pm

Re: GEM in 2018 ?

Post by Petari » Mon Aug 13, 2018 8:24 am

I don't think that EmuTOS is right way - well, unless they change their approach and make it more compatible with existing SW.

I have some ideas how to make some hybrid concept:
It is not only GEM - what is mostly UI, graphic part of TOS. GEMDOS is terribly outdated, or better said, it was never really good with hard disks. Today, it must support some advanced filesystem what can handle very large files, over 4 GB, very large partitions ... So, some of Linux file systems (FAT32 is limited too) . Instead doing all it again, reasonable, I would say only reasonable way is to get some existing Linux distro with sources and make it close to GEM.

Running existing Atari ST SW: 1. GEM SW - should use Gemulator way: executing all VDI, AES calls not via emulation of 68K code and Atari display, but straight in X . That would give it extreme speed. And it will be needed for high resolutions, which are supported by GEM self, just need CPU power.
2. Other SW, like games, utilities with direct HW access and like.
So, basically there should be 2 emulators: Gemulator type, for GEM (AES) SW and full emulator (Steem, Hatari) for other Atari SW .

Since original GEM is limited in many things - like only 2 colors for icons, it may need some extensions. What is already done by some people, teams. I'm not familiar with them, so question is: to select some of it, or making new ones ?

So, to summarize: Linux base, with some GEM like Desktop. At least 2 type of Atari emulation. Which should be well integrated in Desktop, OS .
There is 2 kind of people: one thinking about moving to Mars after here becomes too bad, the others thinking about how to keep this planet habitable.

keli
Posts: 62
Joined: Tue Aug 22, 2017 1:34 pm

Re: GEM in 2018 ?

Post by keli » Mon Aug 13, 2018 8:24 am

Steve wrote:
Mon Aug 13, 2018 2:12 am
WOW this is amazing!! I wonder why I haven't seen this elsewhere. Would it be able to run Atari compatible apps eventually? Or would it need it's own specifically ARM compiled apps?
Save mentioning it on the EmuTOS mailing list, I haven't really announced it widely yet.

Right now it doesn't even respond to mouse events (except for updating the mouse cursor), but my idea was at some point to add a userspace emulator for m68k so it could run well-behaved Atari programs.

keli
Posts: 62
Joined: Tue Aug 22, 2017 1:34 pm

Re: GEM in 2018 ?

Post by keli » Mon Aug 13, 2018 8:35 am

Petari wrote:
Mon Aug 13, 2018 8:24 am
I don't think that EmuTOS is right way - well, unless they change their approach and make it more compatible with existing SW.
I think the EmuTOS team has done an excellent job of creating a legal open source clone of TOS. That's the only choice for me if I want something I can legally build on top of. Besides ARM is not exactly compatible with the 68k, so I don't have to worry too much about badly written software that talk directly to the hardware or rely on a specific version of TOS - that's what hardwarr emulators like Hatari are for.

mikro
Posts: 125
Joined: Mon Aug 28, 2017 11:22 pm

Re: GEM in 2018 ?

Post by mikro » Mon Aug 13, 2018 8:42 am

Petari wrote:
Mon Aug 13, 2018 8:24 am
Running existing Atari ST SW: 1. GEM SW - should use Gemulator way: executing all VDI, AES calls not via emulation of 68K code and Atari display, but straight in X . That would give it extreme speed. And it will be needed for high resolutions, which are supported by GEM self, just need CPU power.
Good news: it has been (being) done: https://github.com/e8johan/tosemu ... maybe the old pioneers would remember the oTOSis project for m68k (with all its counter parts like oAESis, oVDIs etc), it seems those guys have come back (hard to say for how long) with the extension of being not only m68k anymore.

keli
Posts: 62
Joined: Tue Aug 22, 2017 1:34 pm

Re: GEM in 2018 ?

Post by keli » Mon Aug 13, 2018 9:13 am

mikro wrote:
Mon Aug 13, 2018 8:42 am

Good news: it has been (being) done: https://github.com/e8johan/tosemu ... maybe the old pioneers would remember the oTOSis project for m68k (with all its counter parts like oAESis, oVDIs etc), it seems those guys have come back (hard to say for how long) with the extension of being not only m68k anymore.
I did a similar proof on concept and called it ParaTos:
https://github.com/kelihlodversson/ParaTos

Never made it to AES or VDI emulation, but it does support most of Mint's system calls.

Petari
Software Moderator
Software Moderator
Posts: 538
Joined: Tue Nov 28, 2017 1:32 pm

Re: GEM in 2018 ?

Post by Petari » Mon Aug 13, 2018 9:25 am

keli wrote:
Mon Aug 13, 2018 8:35 am
...
I think the EmuTOS team has done an excellent job of creating a legal open source clone of TOS. That's the only choice for me if I want something I can legally build on top of. Besides ARM is not exactly compatible with the 68k, so I don't have to worry too much about badly written software that talk directly to the hardware or rely on a specific version of TOS - that's what hardwarr emulators like Hatari are for.
Interesting, because Hatari is what pushes EmuTOS, for instance. The problem in case of EmuTOS is not badly written SW what accessing HW directly. Actually, that's absurd what you say, right because OS has nothing with direct HW access. Or, if you meant that direct HW access self is bad, that's again bad opinion :D (sorry for my overusage of word bad) . In many cases speed matters, so direct HW access is better because is faster.
EmuTOS just does not execute many TOS, AES, VDI calls faithful - like original TOS versions. I know it because I tested, and I wrote here some concrete examples. They claim that all it must be according to DOCs. And that's the problem, because DOCs are incomplete, with some errors. Atari SW authors had eough headache because of it, and there are some solutions which work not on all TOS versions right because of poor documentation from Atari. But most found the right way, if not by DOCs, then by testing how TOS acts. And EmuTOS simply says that such SW is not well written, and they will not support it. Like Spectrum 512 . Well, that's lack of respect on top of all. In my tests, about 60% of tested SW had problems under EmuTOS. That's just way too bad statistic.
Then, it's totally irrelevant what you said about ARM not compatible with 68K. Problem with EmuTOS is that it executes not properly many TOS calls.
Properly in this case means: like regular TOS versions. Now, if you care only for making new GEM SW and test it under EmuTOS - that's your right. And you can be happy how legal it is. How many people will use it ?
The goal is to create something what will run well old SW, and there will be never some larger quantity of quality new SW for GEM, Atari .
There is 2 kind of people: one thinking about moving to Mars after here becomes too bad, the others thinking about how to keep this planet habitable.

Locked