@OlafS25

I meant 68k, because I expect AROS x86 to get seamless integration of 68k executables at some point.

That was sort of the whole point, to have a classic fallback for increased market penetration. It's of course up to each and every developer to do this, but all platforms need more software so surely this cannot be a bad idea.

And yes, I know a classic version will not be as advanced as an AOS4/MOS version, but if developers think of it as a way to encourage users to try out their platform of choice to get more features, this cannot be bad.

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@OlafS25

Seriously, does it matter what desktop GUI people have? People have been using Scalos on Amiga for decades. I hope you're not saying Scalos or Magellan cannot run standard Amiga executables?

What is it about "reduced functionality fallback version" that is so difficult to grasp?

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@olegil

it is very advanced (when combining it with 68k code that is easily possible). And it also runs Scalos and Magellan.

@olegil

people look at Wanderer and mix the limitations of the Desktop with AROS. But AROS is the API and not the Desktop. That was what I wanted to explain.

@olegil

No all programs work on all desktops (Wanderer, Scalos, Magellan)

@olegil

im not sure what adavantages of aros you are actually looking for. they might be many, afaik aros tries to stay compatible with os4 and mos extensions where possible and where they make sense, but to be honest i dont know what these extensions in particular are.

maybe you are able to highlite few examples of practical os4 extensions in comparison to 3.1, and then we can look what aros provides in that field.

what interests me most on aros 68k is the aros hidd infrastructure, especially pci infrastructure that would pave a way to broader pci support on classics (i hope). this infrastructure is currently heavily worked on. as effect sam port should soon be usable again as well.

also there is many tiny but handy things in comparison to 3.1. like:
- better overall compatibility to older kickstarts (in fact cross compatibility is the target)
- automatic setpatch, same iso for all systems, no messing with processor specific libs
- modern features included out of the box to be enabled just with a mouse click, like networking, usb is not working yet.
- handy tools like simple but ergonomic text editor
- windows can be moved out of the screen (very handy on low amiga resolutions, very annoying a handicap on original aos)
...

all that keeping maximum backwards compatibility in mind (again).

Last edited by wawa on 01-Mar-2013 at 11:33 AM.
Last edited by wawa on 01-Mar-2013 at 11:32 AM.

@wawa

Quote:

maybe you are able to highlite few examples of practical os4 extensions in comparison to 3.1, and then we can look what aros provides in that field.

User interface: application.library, notifications
2D GFX: compositing (eg. for fast 2D games in HD resolutions even on Classic hardware)
3D GFX: drivers for GFX cards newer than Voodoo
Memory subsystem: virtual memory (eg. A1200 supports only 256 MB Fast RAM with Blizzard603e)
FileSystems: JXFS (for partitions bigger than 128 GB and files larger than 4 GB)

@pavlor

application.library - no, i dont think so, dunno what exactly this is good for, seems like buncha additions thrown together into a lib. same for notifications. i dont think it is priority and to be honest i dont think i need notifications popping up while i do something on an amiga (like) system, im tired enough of it on windows.

2d gfx compositing - no, however some sort of implementation is planed that would i think go into layers, the dependant on gfx hardware available different methods of gui rendering could be utilized. as example, aga display would work all like up till now, but the same machine opening a rtg screen on an accelerated hardware could take advantage of it features.

3d gfx - not yet, in fact even voodoo is not supported yet as pci bridge support has not been enabled yet under 68k, even though the devices are being basically recognized by pci tool. as i said hidd infrastructure escpecially the pci support is being worked on now.

virtual memory - good question, i have not noticed anything like that. on non rtg amigas i would like to have virtual chip ram being possible to be paged out to fast. otherwise the chip is pretty limited as we all know.

file systems - there are several, im not sure if jxfs is particularly supported. im fine with pfs most of the time.

@wawa

Answering "I don't want it" isn't really helpful when discussing features programmers want to use which are making them shy away from supporting your OS

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@olegil

Quote:

Answering "I don't want it" isn't really helpful when discussing features programmers want to use which are making them shy away from supporting your OS

well, thats so far my personal opinion goes. of course design decissions are up to aros team and not being educated enough i hardly ever interfere with the process, so far ive been content with what has been the outcome and try to follow the discussion.

sofar what pavlor mentioned i must say the compositing is clearly a score point for os4, especially on slow zorro machines where i expect it to contribute to responsivity quite a bit. i dont think it is any soon on schedule for aros.

vitual memory will hardly be a priority too, since amiga-related systems either suffer form 1.5gig addressing limit or may have any amout of physical ram they see fit, while on genuine amigas i still need to encounter a real low memory situation with 128mb, even if using owb on aros68k.

the priority should be getting all sorts of pci drivers working on 68k though, especially gallium, and i hope it to happen soon.

Last edited by wawa on 01-Mar-2013 at 02:32 PM.

@pavlor

If such OS4 APIs are desired, someone has to write them. If there isn't an in demand, open source, Amiga app making use of them, then there's little incentive for any non-OS4 equipped AROS developer to work on creating them. The features probably should be examined for future consideration though, along with features commonly implemented in web browsers, gtk,qt, wxwidgets, toolkits etc. Not to directly port them, but to be aware of how the ideas might translate over to AROS application support.

There's some graphics hidd compositing features which is used for sliding screens. There is some ports of cairo, but it hasn't been made shared or made to use low level hardware acceleration features where available. AROS doesn't yet have an Amiga file system capable of >128G partitions, I'm not sure if the MS compatible ones support that or not, or are bootable under AROS.

AROS is still an alpha release work in progress. Its getting a makeover in its ATA disk driver design, and hardware bus HIDD layer. There was just some fix done for double buffering support put into the amiga-m68k-aros graphics driver. Zune incompatibilities with MUI are getting work done on them. Many other areas (like networking, multi-core-cpu-support, and memory management), could use updates... But that's where the "R" is for research comes in, and involves time and work to make them happen.

Last edited by radical on 02-Mar-2013 at 06:31 PM.

@radical

I have generally a problem with phrases like "alpha release work". Alpha compared to what? And what platform? Alpha is generally used for software in early state that is mostly untested and has lots of errors. I do not see that and it is badmouthing the work that is done.

@OlafS25

I personally favor AROS, since its free and open for development at all levels, and for
any hardware or hosted platform the developers choose to support.

That said, AROS is Alpha on all AROS platforms, and that's just a reality check.
Sorry if you take it as an offense. Its not meant to be one. I wouldn't want someone to think
its ready to completely replace their Amiga OS 3.x, MorphOS or OS4 based system, with no
questions asked.

It's when the platform developers say they've completed their work implementing key items
that beta level and release testing occurs. You have a system ABI, Amiga APIs,
AROS addition APIs, and specific hardware (or hosted platform) items as well as
included software to support to determine that goal. For m68k-amiga-aros you have
the further goal of running all the common Amiga applications including ones that do
fancy things with planar screens and copper list manipulation, which hasn't been a
priority on other AROS platforms.

There are many parts of what makes up an Amiga-like OS, and there are Amiga features
that haven't been well examined and tested out to function correctly under AROS yet.
(the original goalposts) Add to that -- features not present in Amiga OS 3.x proper, but
commonly implemented in third party additions, in AROS, and there's a lot to get right to
be release ready.

Its great that alpha testers, and distro makers are finding these areas that need polished,
and help AROS to improve. Its getting better all the time.