Team

I have been "quiet" for some time now and have sadly watched many "well meaning" discussions degrade into a "Pissing Contest" where reasonable suggestions for Amiga(oid) Community unification are drowned out in parochial "noise".

I tender this thread as an example: Yet another noise thread

Now from where I'm sitting there are four vibrant and healthy Amiga or Amiga like (Amiga-oid) Operating systems. Each has their own strengths, weaknesses, OS and Application developer communities(not listed in any order):
1. Classic Amiga OS
2. Amiga OS4
3. Morphos
4. AROS

I think it is silly and counter productive to try to force only one of these OS' to be the "One and Only" as each is developing clever and unique solutions for their respective communities.

They all currently share a common "Application Programing Interface (API)" set from the Classic Amiga OS as they have all endeavorer to have some degree of "Backwards Compatibility". In this way "older" applications are (mostly) able to be run on all of these platforms which has been great.

Unfortunately (In my opinion) they have all diverged from that standard into different (and to some extent competing) new API's as they have attempted to modernize their persecutive OS'.

This is inefficient as it requires any Application Developer to "work harder" to support these seperate OS' with any new or upgraded applications that they are working on.

My Suggestion:

I would like to get some kind of meeting between the key developers of each of these OS' and see if some common agreement can be reached on "Modern API's".

If some agreement could be reached, it would allow each OS to evolve and develop at it's own pace, and in it's own way... But still allow a common "Application development framework" to be created. This would allow support for all of the OS' can be done with less effort for individual Application programmers.

This would not be an easy process as the various "modern API's" used by each have good reasons for being there and it's hard to change your own projects to support other OS developers BUT I beleive the rewards for doing so would well outway the effort involved.

Even if it takes years to get results, working towards a common "Modern API Framework" would benefit all of the Amiga(oid) OS' and may help heal some of the rifts in this community.

We can then choose whatever one or more OS' we want for our own reasons and still work together as one community for Application development.

Anyway, That's my thougts on the topic. Please share your's as to why this could/could not work and PLEASE PLEASE PLEASE do not let this thread descend into another "pissing contest"

*** Forum moderators, I'm not sure if "Pissing Contest" is appropriate language for this forum, please feel free to edit/replace with a more "politically correct" phrase if you believe it is required.

Regards
Darren

@newbee

You complain about "noise threads" and than you start one yourself ? Way to go ......

On topic:

Start with something simple and basic like the GUI-library:

Here you will have the AROS-guys you only have Zune (a somewhat incomplete MUI-clone) and bunch of really obscure and underdeveloped systems noone really uses.

Then you have the MorphOS-crowd who simple expect their SW to utilize all the nice MUI4-features.

Next is the OS4-gang, prefering ReAction preferably in the last version, unavailable for any other OS.

On top of that you have the 68k-fanboys who can't aggree on anything even half as controversial as GUIs.


Or in short, there is no point in trying to set a standard, those developers willing to support all OSes wil restrain themselves to MUI.3.8 other will use whatever is available on the OS off their choice.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@newbee

I think it could only work if work is restricted to new, non-existing APIs only. And even then it will be very difficult because APIs are such moving target where new features are constantly implemented.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@itix

I re-read my post and think I gave the wrong impression.

I suggest "starting the journey", not necessarily "reaching the destination".

The various OS' will need to create and grow their own API's and this is healthy. I am just wanting to get "some forum" where the person developing the API can present it to the other OS developers and then they can give feedback and choose to implement it themselves (or not) as they see the need.

Any commonality achieved is a bonus.

It should also get easier over time if all OS developers at least make a "best endeavor" to participate as the communication channels and common ground will improve for everyone. There is an old saying that comes to mind "the rising tide floats all boats" which means (I think) that anything that improves the situation for everyone is a good thing to put your efforts into.

(edit: changes last few sentences as was unhappy with them on re-reading).

Last edited by newbee on 28-Feb-2010 at 11:55 PM.

@newbee
Somehow I actually agree with Kronos - simply not going to work. Can't speak for MOS/AROS, but when some OS4 devs have been approached about this before, they weren't interested (not least because it would be impractical).

There's nothing to stop you implementing a new library/API, which is implemented for all Amiga flavours. But trying to restrict all existing Amiga flavours to some rigid pre-agreed API... far too late for that, even assuming the relevant developer egos could agree to it.

_________________

@Kronos

I was complaining about "signal to noise ratio", not "tread length".

If a health and vibrant discussion emerges... that is a "win".

As to your other points, yes you are correct... There is almost no common "modern" standard and it will be very hard to get one.... That is actually the point I was trying to raise...

It will only get harder to find common ground, the longer the communities drift apart...

Do you honestly believe there is no benefit to even attempting to get a set on modern and compatible API' across all of the Amiga(oid) Operating systems????

Maybe if each camp can export their GUI structure to a common XML format we may have a chance.

Having a common Amigoid Framework would help create/port applications at a faster rate. Unfortunately, we don't have the motivation or developers to take on such a huge task, unless we break up the project as small stage bounties.

The Java developers have automated tools to maintain compatibility and generate visual class/function hierarchy diagrams for docs, etc.

At one time Amiga programs 'just worked' - it was like magic. The very name 'Amiga' made peoples eyes light up and a 'wow' always followed.

_________________

@ChrisH

Yes, developers egos are a very complicated thing to manage as each of us suffers some degree of "Not Invented Here {NIH)" syndrome.

It's inevitable that "my way" is going to be correct/better than "your way".

There are some presidency with what you are suggesting, like the "QT Framework" from the Linux world.

I think an enforced "rigid pre-agreed" API' would be counter-productive, but there must be some common ground that could be explored and developed.

I am suggesting an "ideal" and recognize the naivety of my idea. That does not make it a bad idea.

@ChrisH

Quote: "Somehow I actually agree with Kronos"

- See, it can be done.... A positive result already... even if it's agreeing on why we cannot agree... That is better than the position you had before...

@newbee

But if the developers of the OS's have their own ideas and directions and goals.. who will write the common API's? what happens if the newer aros with the MP comes out, and you cant implement memory protection API's in the others? Would you want AROS MOS and OS4 to be the exact same OS but on different hardware anyway? how far do you go?

Anyway, the problem isn't the different platforms, developers and user bases, its that the users have gone insane from posting on forums for the past 10 years. :)

@jahc

I already answered this:
quote - "who will write the common API's? what happens if the newer aros with the MP comes out, and you cant implement memory protection API's in the others? Would you want AROS MOS and OS4 to be the exact same OS but on different hardware anyway? how far do you go?"

My Quote - "I am just wanting to get "some forum" where the person developing the API can present it to the other OS developers and then they can give feedback and choose to implement it themselves (or not) as they see the need."

ALSO:

quote - "Anyway, the problem isn't the different platforms, developers and user bases, its that the users have gone insane from posting on forums for the past 10 years. :)"

I think you have the cause and effect backwards... I believe you have to be insane to be posting on these forums for the past 10 years

It's a funny old situation really (a regular soap-opera in fact). As I understand it, Morphos and AROS were created mainly because official development had stalled which is not really the case these days.

Ideally you would have an official branch and then have different streams [projects] (open or closed source) off that branch (eg Morphos, AROS, etc). You would then be able to have a common API (to a large extent) and have each sub-project benefit each-other and the main project.

I don't see this happening of course, primarily because there's still a lot of bad feeling between the various flavours but also because it would be a massive amount of work to go back and do it properly this late in the game. This is of course a great shame as so much more could be achieved if developers worked together.

Last edited by NovaCoder on 01-Mar-2010 at 10:13 PM.

The current common API is the 3.1 API. It is the lowest common denominator. This can be both a good or a bad thing depending on your point of view and your API preferences.

You will never get the Dev teams of OS4 and MorphOS to agree on anything regarding an API. What you are suggesting would not be unlike a Linux fanboy insisting that MacOSX and Windows APIs be consistent. Not going to happen in a million years and Linux fanboys know better than to waste their breath.

Only application developers outside of the commercial OS dev team could possibly create and adhere to a common API. But they have to have both the compulsion and the will to do so. Developers creating apps for OS4 may not want to support MorphOS and vice versa and will lack that compulsion. Maybe there are features, or an approach, used on that OS that makes a port untenable or maybe they just plain don't want to for their own reasons(which I don't want to get into).

Basically, if you wanted a NEW common API that is either more advanced or simpler to use, you would have to create one from scratch and promise support for all platforms. An argument against this is that it would require an abstraction layer, which itself would become the lowest common denominator. An abstraction layer usually incurs a slight performance penalty, which a lot of Amiga aficionados might frown upon, but if your app is event driven may not matter too much. If you open sourced the API development you might get there quickly but it may be difficult to get any community surrounding it to agree on anything.

There seems to be a few projects to create new APIs, and perhaps some of them are cross-platform.

Last edited by BigBentheAussie on 01-Mar-2010 at 06:35 AM.

_________________
Leo Nigro, CTO Commodore USA, LLC
Opinions expressed are my own and not those of C= USA.
Commodore/AMIGA "Beautiful, High-Performance, Home Computers for Creativity and Entertainment."

If your reason for asking for a common API is just to get applications to have cross-platform compatibility, there are various ways that this can be achieved.

There is the promise of high-level languages providing some common ground. Java may be ported some time in the distant future, and who knows if something like PortableE or MathiasBasic(?) can fill some of the cross-platform slack if they get UI APIs. Then there's Hollywood that is already cross-platform, but is interpreted Lua, I believe.

SDL is one approach to cross platform ability. It is commonly used for games, but I heard SDL is used by Cinnamon Writer and Hollywood. So it is not impossible to create cross-platform applications right now, they'll just be clumsy in terms of UI integration with the larger OS.

I remember the days when it seemed a lot of apps had different UIs. As long as they were slick I really didn't mind. Others might not agree..... and that is fine.

Last edited by BigBentheAussie on 01-Mar-2010 at 06:37 AM.

_________________
Leo Nigro, CTO Commodore USA, LLC
Opinions expressed are my own and not those of C= USA.
Commodore/AMIGA "Beautiful, High-Performance, Home Computers for Creativity and Entertainment."

"Common API"

That is what we need. That is something the community is able to start doing, when OS developers are not able to co-operate.

After/if it gets enough momentum, perhaps also OS core developers start helping by building wrappers/etc..

(logical, initial testbed might be AROS??, that is most sorely lacking modernized/NG API, IMHO)


Update/ideas:
- the API should handle SMP and MP
- to me it seems that Qbox is abandoned by MOS team, was it bad?
- could intent give some interesting design points for the new API (surely, binary compatibility would be too far fetched, eventhough interesting)
- any learnings to be dig out from Apple multi-CPU-architecture binaries

(I would also like to see a linux-apps-sandbox for Amigalike OSs. So that we could run almost all PPC linux apps in a window on any AMigalike OS. I think Cygnix? is not there yet. in the long run ... it might also enable wine on AROS ??? Co-operation with Anubis guys???)

Last edited by KimmoK on 01-Mar-2010 at 08:01 AM.
Last edited by KimmoK on 01-Mar-2010 at 07:56 AM.
Last edited by KimmoK on 01-Mar-2010 at 07:55 AM.
Last edited by KimmoK on 01-Mar-2010 at 07:52 AM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@newbee


This sounds really interesting to me. I think i've created a project years ago about this, a slide show with some content about common api, if i recall well.

_________________
retired

With SDL, the common argument is that it "doesn't look native anywhere" and indeed, (at least in my opinion) it's "mostly for games", which often are better with their own custom GUI in fullscreen window. And generally it's a bit faster if you use native API's directly instead of SDL.

But other than that, there is common ground with OpenGL, for example. Let's not start discussing which platform has best implementation, they are all rather limited (no shader support etc.). And unlike SDL, there isn't really a "better way" to do the same thing. Also, all platforms have AHI support for sound. TCP/IP should be very close aswell from developers' point-of-view.

Seems like the only common APIs (in addition to the 3.1 API of course, which is the base of the three OS), are the open(source) ones: AHI, SDL, OpenGL...

Amiga is and has always been closed. People/developers driving it also are closed. Not to mention if you followed the history they have gone way too far, and too personal to even think about having a meeting.

I think the only common API can be done by external developers indeed.

As for the GUI: there is a nice one (although not finished) API: Feelin. That's too bad nobody uses it since it has really neat features (XML, CSS,...). All other APIs have a lot of legacy, be it MUI or Reaction. It is too bad no one seen benefit in switching to a clean/fresh standard...

_________________
http://www.warpdesign.fr/

@Leo

Damn,for the first time,i actually agree with you

_________________

@ChrisH

Quote:

There's nothing to stop you implementing a new library/API, which is implemented for all Amiga flavours.

I think this is the best suggestion. Rather than trying to get agreement, put code out there, whether it's new functionality, or compatibility layers for existing functionality (i.e. looking at MUI vs. Reaction, a lot of "basic stuff" would be easy-ish to hide so developers would at least have an easier time writing GUIs that work on both).

_________________
Wiki for new/returning Amiga users - Projects: ACE basic compiler / FrexxEd / Git