One of my Amiga obsessions is what-if-ing over what AmigaOS would be in it's next iteration after the original 3.1.
Now of course we have a real 3.2, which is wonderful, and it certainly does a great job of 'finishing' the things left hanging like adding Help and BOOPSI (with Reaction), as well as rounding out things like modern filesystem support.
Then of course we have defacto standard extensions such as P96 and AHI.
So what's left? If there was an Amiga 1400 in 1995 for example, what features would it have to help it compete with Mac OS 7.6 and 8, and Windows 3.11 and 95?
I don't mean memory protection and things that the hardware just can't do, and I also exclude virtual memory since memory is cheap now even if it wasn't then.
Some obvious candidates - TCP/IP and LAN working were filled then by AmiTCP and Envoy, and now by Roadshow, and SMBFS.
For me the obvious missing piece is a multimedia framework, equivalent to Quicktime, or gstreamer. I'd like to think there's a parallel universe where Amiga had its own Quicktime like API with all that that entails.
What AmigaOS missed in 1995 was an owner with a solid vision, and competent business leaders and marketing team. Lacking these things was already a problem back when Commodore was still alive.
Well the topic is what features should have come next in the roadmap, not what leadership was missing. The latter is well understood and not particularly interesting to debate IMHO. The former is I think still a fun intellectual exercise at least for me.
The delta between AmigaOS and the MacOS of the day, and even windows 95 is not that big, Amiga was even ahead in some areas of course, and behind in a lot of others. But the gap is not huge when factoring in the 'third party' stuff like AmiTCP, Reaction (albeit MFC was a lot more substantial than Reaction), Envoy, P96, AHI, and your own efforts among others. Again, in my mind the biggest miss is a multimedia framework like Mac had with Quicktime, and to a lesser extent windows media. I like what the EntwhistlerX guys have done with Emotion and FFMpeg. I feel that comes close to filling the gap for OS4.
For me the obvious missing piece is a multimedia framework, equivalent to Quicktime, or gstreamer. I'd like to think there's a parallel universe where Amiga had its own Quicktime like API with all that that entails.
AmigaOS has datatypes for multimedia playback, loading and saving (saving in native format is not implemented in most datatypes and falls back to some basic IFF format instead), not only with MultiView, any software can embed it with BOOPSI in it's GUI. For videos there was even an Amiga specific format usable on very slow hardware (https://en.wikipedia.org/wiki/CDXL ), something like MPEG decoding in real time was impossible with the Amiga hardware of that time (except with the https://bigbookofamigahardware.com/bboah/product.aspx?id=488 FMV hardware MPEG decoder for the CD³²), but it was mostly used in CDTV and CD³² games only.
Yes of course I'm aware of datatypes and the IFF system.. Quicktime even in those days was an order of magnitude more sophisticated, it was a proper pipelining multimedia framework with pluggable codecs, containers, source and sinks, timers etc.
Amiga had datatypes, realtime library and, uh, nothing else. A natural addition therefore would be a multimedia framework IMHO.
Emotion uses the ‘VA library’ / VAAPI - so you will need an SDK for it. I asked once in the forum on this topic. Apparently the SDK is related to ‘Enhancer Software’ - I don't know. I don't have the latest version.
There was a request for help related to vaapi and mplayer https://www.amigans.net/modules/newbb/viewtopic.php?topic_id=8600 Nobody answered. Already with FFmpeg 6 you had to undo patches to get something VAAPI working ( under Linux it works without problems) There is now a version of FFmpeg 7 where the old vaapi (AOS4) no longer works for sure. The code was already obsolete from FFmpeg 5 and I don't think anyone would go that far back to make it work with FFmpeg 7.
The Emotion developers are working (information from this forum) on an Emotion Update. They haven't written what version of FFmpeg it will be based on. They may have managed to get it to run on FFmpeg 6.
There is now a version of FFmpeg 7 where the old vaapi (AOS4) no longer works for sure. The code was already obsolete from FFmpeg 5 and I don't think anyone would go that far back to make it work with FFmpeg 7.
This is news to me (and I wrote the va.library). So what's the minimum VA-API version that ffmpeg 7 needs in order to work? I did a quick search for VA_CHECK_VERSION() in the ffmpeg source code, and couldn't find any indication that v1.5 is too old (va.library is at v1.5).
Bear in mind that the functions that interact with the OS' bitmaps and windowing system are different on AmigaOS, so you can't expect the ffmpeg code to work out-of-the-box
In the Amiga world, I went in one giant leap from AmigaOS 3.1 running on my Amiga A2000HD computer to AmigaOS 4.0 running on my MicroA1 computer. I missed alot of Amiga software and hardware enhancements during the time between those two systems.
I won't speculate on what could have been or should have been.
I am still amazed that back in the day I was surfing the net using AmiTCP and Aweb-II over a dialup modem connected to my Amiga A2000HD. Likewise, I am amazed with what I can still do on my MicroA1 using AmigaOS 4.1 Final Edition Update 2 and IBrowse, Odyssey, NetSurf and other programs.
Sure, we are far behind modern Windows and Mac systems. Even my Samsung tablet is more powerful than my MicroA1.
I see a few issues that even the latest OS3.2 doesn't solve. But back then they were more urgent to solve. If they wanted AmigaOS to be taken seriously.
Icons were a mess and easily became sloppy. There was too much work organising it and saving it, usually needing 3 menus to do it, which didn't always look right. The "E310" format was designed for updating but lacked any major version updates. It remained stuck at 640x256 screen size icons in 4 colours. After 1.3 the colours didn't match either. But a related issue was icons didn't scale. By this point they needed to become vector images so they could scale. Being natively planar meant that the hardware was suitable for layered vector images.
It needed TCP built in. Even the attempts from OS3.5 and OS3.9 were third party. So integrated TCP is another on my list.
An obvious one is RTG. And not with an external CyberGfx or P96. Again, integrated into the system. And easier to setup. RTG was and still is too technical to setup. Amiga OS3 RTG lost it to Windows 95 but Amiga OS4 RTG gained some back. Embracing DDC standards would help but mostly being able to easily setup a monitor with standards resolutions.
System management was too awkward. While Storage was a nice idea there was no system utility to manage it. Instead you had to manually open up windows and drag icons around. The windows themselves opening up for each drawer you opened made it become cluttered. So some kind of paged or browser based windows would be next to alleviate that.
The system became modularised by introducing user scripts and moving mounts into separate files. But the idea wasn't developed enough. For example, User-Startup became a mess, with a mix of random scripts. There was no script drawer to split them into separate files and organise them. Nor an assign manager. So assigns remained complicated to manage. When it became obvious that moving assigns into their own drawer was another next step. The Installer helped but it had to modify one file to add or remove scripts, which didn't always work out, so was forced to install or remove the hard way. The Installer also wasn't installed to HDD as standard so needed to be included with every program package using it. And on that note a package manager also became an apparent need.
There would be others but there's 5(+) main ones off the top of my head.
Yes, Commodore's mismanagement was the root of all problems, a revolving door of ineffective but highly-paid CEO's while engineering starved. AAA in 1992 instead of AGA would have put them in a better place in 1995, building on 16-bit sound, a built-in HD floppy controller, and the improved graphics. Concerning the OS, I think of it as a lack of "polish" or "fit and finish". Windows didn't get a built-in browser until Windows 98 and MacOS didn't get built-in networking until OS8 (Safari wasn't until OSX), so I don't fault Commodore for that, but...
1) They created the datatype system but IIRC barely included any datatypes. If you're striving for a multimedia OS, I would expect maybe two dozen picture, sound, animation, text, and video datatypes. Maybe they could have sold a separate expansion pack?
2) Ditto for ClassAct/Reaction. The framework was there but only a handful of SOB's existed.
3) There was no simple GUI text editor to compete with Notepad on Windows and SimpleText on the Mac.
4) I don't recall any disk recovery utilities included after DiskDoctor. DiskSalv Lite would have been nice.
5) Workbench was under-powered. The ARexx port in OS 3.9 helped, but that was years later. There was no dedicated right-click menu, no icon cache, no way of hiding unused icons, DefIcons didn't appear until 1996-ish as freeware, and "Trash" seemed like an afterthought.
This is news to me (and I wrote the va.library). So what's the minimum VA-API version that ffmpeg 7 needs in order to work? I did a quick search for VA_CHECK_VERSION() in the ffmpeg source code, and couldn't find any indication that v1.5 is too old (va.library is at v1.5).
What about DVPlayer, is it still being developed further?
I would like to see streaming support and some updated codecs. It would be great if they could do that.
This is news to me (and I wrote the va.library). So what's the minimum VA-API version that ffmpeg 7 needs in order to work? I did a quick search for VA_CHECK_VERSION() in the ffmpeg source code, and couldn't find any indication that v1.5 is too old (va.library is at v1.5).
You are right, the minimal version checks different/newer/older versions of VA_CHECK_VERSION. Unfortunately no one is checking this for AOS4. The API has changed and what I've read and checked doesn't work. I took the time to understand how the VA-API for AOS4 works a year ago by asking people who have access to the SDK. Too bad the SDK (without proprietary libraries) is not generally available. There was a time for mplayer to work with VA-API, but at the moment no one except Emotion developers can do anything about it.