Hypothetical OS3.x Improvements?

[fishyfish]

Firstly, feel free to move this to a more appropriate area, I just couldnt see anywhere that this fit, and this appeared the closest match to me.

It's always interesting to watch the "NG" Amiga systems progress, get new features, bug fixes, optimisations, and so on. While I do follow the scene, and do have both MOS and AROS machines my passion is with classics 1st and foremost. There's a few reasons for this, but that's not really the point of the thread, so I'll leave that alone.

As for the point of the thread, Im guess Im just curious exactly what, hypothetically, could be done to speed up/enhance/bug fix/etc. OS3.x?
More than a few times in "vs." type arguements on other amiga websites I see people who are more exclusively "NG" followers dismiss os3.x because "it's a dead end/not developed anymore". Hardware wise this is partially true. There's a healthy amount of new stuff developed, but specs for classics dont improve outside of emulation.
That said, new software is ported along with dependencies and libs are updated, system libraries are still getting patched, compilers are updated and so on and so forth.
In short it'd be a bit of a stretch to call OS3.x a dead end. It progresses and evolves still in its own way.

While I consider myself fairly knowledgable when it comes to the amiga I still have a bit to learn with really low level stuff, so thought Id ask people in the know

I understand its not as "easy" to keep developing core stuff for a closed source OS as it is for the OS4.x/MOS/AROS developers, but Im curious just what can and cant be done. Also, what would be the best stuff to target? Any core stuff known to be less efficient than it could be? What about apis? Anything that could be implemented/added/patched/etc. by a person with an incling that'd make a handy addition/speedup/whatever?

While this is all obviously hypothetical Im more curious about what could realistically be done. Obviously with enough time, effort, manpower and money pretty much anything can be done

[gulliver]

I personally believe a lot can be done to AmigaOS 3.x to improve it.

A good aproach that has been taken by some assembler coders/hackers is taking the challenge to resource an AmigaOS component, and hand optimise its C routines for faster and smaller asm code. It even has the advantage, besides size and speed, that some bugs get discovered and fixed.

Besides optimisations and bugfixes, it would be interesting to have some new features implemented:

-Fixed cdrom filesystem now supporting (HFS/HFS+, UDF, and BlueRay)
-PFS3 as default filesystem (FFS only kept for compatibility)
-Second channel IDE support (a sort of idefix97 built-in)
-A better printer system. It could be either a licensed TurboPrint, or some sort of Linux CUPS with GUI.
-Reworked preferences much better integrated. For eg.: We dont need 3 printer prefs (printer, PrinterGFX, and PrinterPS), they coul be integrated into one. The same goes with Screenmode and Overscan. Just to name two examples.
-A prefs program that should be a cross between wbstartup+ and extmanager
-Updated datatypes for more actual file formats
-A filemanager (something like Dirwork or Dopus 4.x)
-Some sort of ScalOS functinality on workbench (If you have used a bare OS install, and another one with ScalOS, you would understand what I mean)
-A low resource TCP-IP stack with GUI that supports IP v6 and DHCP
-A low resource text based web browser (maybe a port from an existing one, like Elinks)
-Some sort of Partition Magic for Amiga (That woul be indeed better than the aging HDToolBox)
-NTFS and FAT 32 built-in support
-A sort of System installer that will choose when installing the OS for the first time, the best components to install according to your hardware (An OCS 68000 with 512KB is definately not the same as a 68060 128MB AGA system).

Other kickstart enhancements (bootmenu):
-Possibility to softkick devs:Amiga.kick (toggle option) it can be a custom 2MB kickstart rom image
-built-in pcmcia cf device for multi card reader (built-in fat32 support)
-code built-in allows cd/dvd/bray booting
-disk drive boot delay is user configurable no matter if you are using scsi or ide drives
-Serial radboot toggle (loads an adf from serial port and radreboots it). Useful for getting an Amiga to boot workbench and generate floppies out of the blue
-Anaiis usb stack support for (hubs, mouse, keyboard and storage media)
-booting with a 68060 and no 060 lib no longer crashes, it loads a generic 68060.library, and gives you a warning, so that you can afterwards install a proper 68060.library.

And finally, Integration is a very important aspect of this endeavour. Think about why have many workbench packs succeeded: because all the basic stuff has been already put together and, is easily installable. Unlike a bare 3.x system, that requires a lot of manual work to become usable by todays standarts.

There is a lot in mind, but then those were the first few things to come to my mind

[commodorejohn]

Yeah, an Elinks port would definitely be nice. What would be nicer is a low-resource graphical browser with decent CSS support and a Javascript whitelist feature, but that'd be quite the project.

[mfilos]

Gulliver mate... you said it all!!!
/respect

[Amiga1992]

Here's some work already done by awesome people:
Opensourced dos library: http://eab.abime.net/showthread.php?t=60181
icon library: http://eab.abime.net/showthread.php?t=64079
workbench library: http://eab.abime.net/showthread.php?t=65735

[NovaCoder]

There's two streams already happening as far as 'improving' 3.x.

The first is the updated (patches/rewritten) original code and then there is AROS 68k.

The interesting thing is that the original libraries are written in 68k assembler whereas AROS 68k is written in C (as far as I can tell).

This means that only the skilled old-school Amiga coders can update the original libraries whereas any old hack could work on AROS 68k. For that reason, I believe that AROS 68k has the potential to (one day) be the most advanced 68k OS to run on real classic hardware. It's a shame that most of the Amiga developers have moved on to OS4/MorphOS and AROS x86.

AROS 68k ROM
AROS 68k Vision