Software Wishlist for 2023

[Martin Iturbide]

Hi, Happy New Year

I always like to discuss our wish list for each year and get some discussion. Why? because software is a thing that changes with time, and what was good some time ago changes too.

My wishlist has:

- Otter - We still require a solid webbrowser.
- Qt6 - I have my hopes with Paul here
- OpenJDK - Java Runtime enviroment
- Eclipse - Development IDE

This is the open source Wishlist we discussed on 2022-09.
- The GIMP.  Photo edition tool
   - To port this it is required to port GTK+ first.
- Inkscape. illustration tool
- Audacity. Sound recording editing tool
- QEMU. Machine emulator and virtualizer.
- Openscad . The Programmers Solid 3D CAD Modeller.
- QCAD. 2D CAD
- R - Statistics, 3-D and dynamic graphics
- TeX live. Recent version of the TeX typesetting program, successor to teTeX.
- Kate. Text editor.
- Notepad++. Text Editor
- Peazip. Archiver tool
- OBS. Video streaming, desktop recording.
- KDevelop. Development IDE.
- Lazarus. Development IDE.
- wxWidgets.  C++ library GUI API (Update it)
- LibreOffice - Since OpenOffice seems to be evolving to slow

Regards

[Remy]

I would like add Upplay QT5
(High need here and for those enjoying high end music / streaming)

[Eugene Tucker]

Definitely Audacity. I would also like to see Falkon browser as it passed Otter up in working features.

[Dave Yeo]

Falkon is a KDE browser, so would need at least parts of KDE ported I assume.
OTOH, if it was possible to port KDE, that would result in lots of other software being portable. Whether KDE is fully portable to OS/2, I have no idea. One gotcha I see is everything uses 3D graphics to be pretty and take advantage of modern graphic cards and their accelerated 3D drawing.
Anyways while wishing, we could wish for KDE. Then there's GTK, which would also be hard to port (like QT but likely harder) but would also give many programs such as the Gimp

[David McKenna]

Hi Martin,

  I like your list, and while we are dreaming, would like to add updated software OpenGL (which is an actual component of OS/2).

Regards,

[MPJ]

Would be nice if ArcaOS 5.1 would finally be finished.

[Dave Yeo]

Eventually, we're likely to need ipv6 support. Probably mean a new TCPIP stack, a major job.

[Mentore]

Eventually, we're likely to need ipv6 support. Probably mean a new TCPIP stack, a major job.

Indeed a major rewrite, since the TCP/IP stack available does know nothing about IPv6. Maybe many low level drivers to be rewritten from scratch: I'm afraid of possible problems related to 64-bit needs...

Mentore

[ivan]

It sounds as if the real answer would be a 64 bit kernel and support programs.  Something I assume will never happen.

[Roderick Klein]

Eventually, we're likely to need ipv6 support. Probably mean a new TCPIP stack, a major job.

Indeed a major rewrite, since the TCP/IP stack available does know nothing about IPv6. Maybe many low level drivers to be rewritten from scratch: I'm afraid of possible problems related to 64-bit needs...

Mentore

About 98% of the code in the TCP/IP stack drivers such as socketsk.sys is recompiled FreeBSD code, It would possibly be easier to dump the old stack on OS/2 that we have and if somebody has the time port combined IPv4 and IPv6 stack. There are light wheight combined stacks for embedded systems that should be able to run on OS/2. Not needing 64 bit support.

How much work this would be. I had a URL to an open source IPV4/IPv6 stack but I lost the link.

Roderick

[roberto]

It would be interesting to add to that list a program that I do not remember the name, but that was a copy of the Dbexpert, for Linux. If someone remembers it.
saludos

[Andi B.]

For what do you really need IPv6?

My own LAN is running IPv4 as it always did. If I've ever have a need for outside IPv6 connection, I'm pretty sure my LTE and my xDSL modems will do the conversion for me. For me there's no sense in making a direct connection from the outside world to one of my private systems except VPN. What are your use cases for IPv6?

[SoundChaser]

Frankly (and I have all of Stevens' TCPIP books here) IPv6 couldn't be lower on my list. We're still on IP-beta (aka v4) in 2023 and the migration was supposed to be completed by 2010. Like FM radio, it's here to stay.

As a community, we need developers. People to back the work of Paul, David, the folks at BWW, etc. The very first thing new developers want is a friendly development environment.

On ArcaOS it is sort-of there. At least there's ANPM/yum for dependencies management. There should be a single 'awesome ArcaOS' git(Hub,lab) with everything you need to get up to speed with a tested stable development kit. That means OpenWatcom 1.9 and gccand all the environment scripts sorted out and standardized, a modern editor... and I know someone can take a few hours to set up by themselves. Bullshit. It's just yack shaving and non-productive. I want Suntan Special for developers.

Once that is in place, a developer community can actually get started to support ArcaNoae, bww and Paul on Qt6, blink/chromium, KDE apps (thanks qt6) and only then - maybe - gtk apps.

But I tell you the developer experience is currently very opaque, cumbersome and error-prone.

Since the most urgent need is on C/C++ side, I obviously wasn't talking about REXX development.

[Martin Iturbide]

Hi SoundChaser

There should be a single 'awesome ArcaOS' git(Hub,lab) with everything you need to get up to speed with a tested stable development kit. That means OpenWatcom 1.9 and gccand all the environment scripts sorted out and standardized, a modern editor... and I know someone can take a few hours to set up by themselves. Bullshit. It's just yack shaving and non-productive. I want Suntan Special for developers.

I take these words like some kind of "ArcaOS Developer Kit" should exist. Not matter if it is on github, or any other place.
An idea to combine development tools, compilers,  libraries and source code samples in one single package. Like some kind of OS/2 Toolkit and IBM Device Driver Development Kit, but trying to bring to 2023 for the ArcaOS platform.

Regards

[SoundChaser]

Most desperately so.

Should I use GCC (in absence of clang, it's to only sane choice) or OW? well it depends to who you talk. Many developers only swear by OpenWatcom which is an utterly confusing mess since a developer took on him the burden to revamp the stack with variable-to-absent consideration for testing on ArcaOS or OS/2 4.5 .

If one is to port an app... EMX? Klibc? BWW libc? It appears that bww is the only viable choice, but with all the references out there on the other previous iterations, it makes it still very confusing.

Once you decided the target libc, how to have the auto-tools and cmake recognize the ArcaOS SDK requirements (linker?!)

Each step in the way is a mix of forum or mailing list archives digging, archeaology and sweat. There's some fun in that, but it's not productive. And most hobbyist developers won't know where to look and they'll go back tinkering with their RPi instead. Much more developer friendly.