new os3.x software ?

[fishy_fiz]

Firstly, Im mostly just bored and thinking aloud (or in text) here, I have no specific particular ideas or plans, Im mostly just interested in peoples thoughts.  I guess the crux of it all is me wondering just how many 680x0/classic/amithlon/heavy uae, etc users there are still ? Personally my main used "amiga" system is amithlon, but I do still have a strong interest in actual classics too (of which I have no working machine atm unfortunately). The main reason I ask is out of curiousity towards how much money there is in this portion of the amiga pie ? Not that I expect anythone to get rich from it, but, for example are bounties for os3.x/68k software/features/etc. feasible? Bounties arent the be all/end all solution, but they have done some good for the other amiga systems. I know there's a few things Id like for os3.x. Im also curious as to what sort of sales a (hypothetical) good, commercial/commercial quality new 68k amiga game might generate? Again I know no-one is going to get rich, but would, say, 100 copies be feasible ?
I know other routes are considered "the future of amiga", but I still enjoy os3.x/68k amiga systems and it would be nice to get some nice new "big" software/features from time to time.

[fishy_fiz]

It's good to see some interest in stimulating development for os3.x/68k, but as some of you have touched on, where to start ? As for what to include, for now I think nothing should be off the table. New drivers, new apps, new patches, ports, whatever. Being that os3.x is these days a community driven system it's up to the users to shape what happens. Best plan of attack in my opinion is just for people to say what they'd like to see and hopefully the most commonly wanted things will become clearer.
Here's my thoughts:

Software should be completely system friendly. This makes bounties of use to OS4.x/mos/amithlon/draco/etc users as well. Theoretically it also opens up potential for people doing bounties for other amiga based systems to simultaneously do a 3.x version when/if applicable.
Software should consider users of classic users where applicable. (for example, a 68k open office might run ok on amithlon/winuae/etc., but if there's other roads to more modern features that will be usable on "real" amigas maybe they should be considered 1st). For me it's not about big names, but new features/functionality/software for os3.x.

As for what I'm interested in supporting:

Either a new browser, on updating existing browser (I guess Aweb is the most likely candidate for the latter). Alternatively there's a few interesting 68k browsers for other OSes. Cant hurt to email those guys. (ICab for 68k mac os and an atari st browser whos name I cant think of at the moment). Netsurf and OWB 68k versions are useful, but not really something I'll use (yet) for my main browser, and unfortunately not really useable on real (non emulated) 68k hardware.
Support for ms word .doc files. Unfortunately this is essential for me as my uni assignments are in this format and often with images, etc.
Video editting software. Ideally with some sort of compatibility with Adobe Premier (again for uni purposes), but I have no idea what sort of open source options there are out there, let alone ones that are realistic. Id be happy with something simpler though, so long as it exported to things like cdxl and iff anim.
I dont know how feasible this one is, but Id dearly love some openpci gfx card drivers with 3d acceleration. Being that Im predominantely an amithlon user I'd love to get 3d acceleration. I dont really know the technicalities involved, but if it was possible to develop a 3d subsystem and a driver or 2 that would work on both amithlon and classic amigas I'd be a happy man.

There's plenty of other stuff I'd like to see, but theyre probably the things that are of most interest to me at the moment.

So, feel free to say what you'd like to see or any thought or ideas, and hopefully eventually a clear path as how to progress will start to show itself.

[fishy_fiz]

Yeah, a more "amiga-fied" Netsurf probably seems the most realistic option at the moment as far as a browser goes. There's also Aweb which is open source, but there's probably more work involved in bringing that up to date. Maybe Awebs javascript could be borrowed for Netsurf ?
As for games, Im actually pleasantly surprised by people saying they'd like to see games that utilize upgraded hardware. Apart from doom clones we saw very few games aimed at an upgraded amigas strengths, and it's always something I've want to see. Also being that my main interest in computers is making games it definately doesnt hurt the motivation to pull my thumb out and make a "serious" amiga game.
Oh, not important really, but my dream game to make for 68k amiga is a Resident Evil type clone. With a bit of imagination from the programming side I'd imagine something similar to Resident Evil on the the ps1 or saturn should be doable on an '030+aga machine.

[fishy_fiz]

It seems one of the common thoughts is a more modernised browser, and an amigafied netsurf really does seem to be the most realistic option. My only concern with that is so far it does look a little heavy. As much as I'd prefer efforts to focus on upgraded machines I think targetting something like an '030 with 16 or 32 meg fast as a minimum spec for software to be useful is a better idea than only having something usable on '060. Ive never really looked into netsurf sources, but how much can it be speeded up ? current os3.x uses sdl if Im not mistaken, both for input and gfx ? While I appreciate the efforts by the os3.x maintainer, it's still not really an option for the majority of classic users, and has a little work to be done yet. Personally Im all for supporting a bounty based on netsurf, but in my opinion there would have to be a fairly big speed up for it to be the right choice. Does anyone else have any broswers they know of that may be viable ? There's always OWB, but that's heavier again. And Im I right in assuming a browser might be one of the first 68k bounties to try ?
Please keep any thoughts and ideas coming. Im eager to get something started, but arent 100% sure where to start yet.

[fishy_fiz]

OS3.x native devtools are maybe a little outdated (maybe not, but it's hard to find definitie answers about where it stands, let alone binaries) for people wanting to port *nix software, but Im under the impression cross compilers are more up to date. OS4.x GCC if Im not mistaken is based on 4.1.x, so it's not too out of date. Having said though though, being that it's still an actively developed commercial product I'd be surprised if it didnt receive at least semi regular updates still.

[fishy_fiz]

What type of impact does this lack of optimization cause? I've read a little about it, but haven't read its impact. Impacting graphics performance, I/O performance, data processing performance, etc? For example, a game--what is significantlly impacted such as blitting, game logic, sound, network and at what processor level is it seen? What processor level, if any, would no longer notice the performance hit?

Do you mean in regads to the level of support "compiler x" has ?  The code generated by a heavily optimised compiler  vs. a more generically compiled compiler can be pretty big actually. Depending on the software it's probably not impossible to achieve something around %50 faster (and more in certain cases) on a heavily optimised compiler. As for using compiler optimisations vs none using the same compiler, there's probably not a massive differerence, unless there's supporting code for "cpu x".  At the end of the day though, worse case scenario for a binary generated can be a lot worse than best case scenario depending on compilers and optimisations used.
Hope this answers what you was asking ?

[fishy_fiz]

Probably one of the most useful tools for any Amiga user who enjoys programming in 68k is ReSource...

Thanks to this little gem of a program I've managed over the years to disassemble and then optimize, bugfix or add new features to many old original pieces of software that I use frequently.

Currently I'm in the process of disassembling AWEB, which I think was written in C originally. Im not knocking the authors of AWEB but when you see the code disassembled then it's obvious that there can be many optimizations done to this program that should help speed it up quite a bit.

I've got a long way to go yet before it's fully disassembled but that's one of my pet peeves about code written in C, it's bloated and not efficient.

It's the same with the roms, there are several hundred optimizations that can be safely made, ok the speed increase in nothing great but as we all know every wasted cycle on the Amiga that you can get rid of can only be a good thing...

AWeb is actually open source, so might be easier to just download the sources than disassemble :-)