OS2 World Community Forum

Public Discussions => General Discussion => Topic started by: Martin Iturbide on March 19, 2025, 06:39:14 pm

Title: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 19, 2025, 06:39:14 pm
Hello

Yes, again, the typical thread for brainstorming about the future of OS/2 and ArcaOS.

On Dec of 2023 I sent the attached wish list document to Lewis based on past discussions.

Today ArcaOS is running on UEFI and finally has some language localizations like German, Spanish, and Russian.

It seems to me that the top requirements from the perspective of the OS are still:
1) Modern web browser
2) Wifi Driver.

What do you think it is missing or what had changed from 2023's wishlist to today?

On the other hand, what else can it be interesting to do as a community?

Regards

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Mentore on March 20, 2025, 07:59:18 am
Hello

Yes, again, the typical thread for brainstorming about the future of OS/2 and ArcaOS.

On Dec of 2023 I sent the attached wish list document to Lewis based on past discussions.

Today ArcaOS is running on UEFI and finally has some language localizations like German, Spanish, and Russian.

It seems to me that the top requirements from the perspective of the OS are still:
1) Modern web browser
2) Wifi Driver.

What do you think it is missing or what had changed from 2023's wishlist to today?

On the other hand, what else can it be interesting to do as a community?

Regards

Being both a developer and an user, I can see from both sides.
Yes, we need drivers and yes, we need a modern web browser.
For these targets to be, eherm, targeted, we need developers.

It seems to me that on OS/2 we have gone back to the '80s era, when there was no software and you had to build it yourself. This is pretty bad-looking from a common user's point of view: we are really used to the plain and simple "I need some software to do XY, I search on the 'net for XY, I find the software, I install it, I'm more or less satisfied, case closed". This is not really the case on OS/2.

So something good to do would be to nurture a little more our shrinking developers' community.
I'm really afraid this would be a HUGE task, but I feel this should be some starting point to try and obtain something more from this old beloved beast.

Mentore
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Andi B. on March 20, 2025, 09:24:28 am
My top priorities would be -
1) stable operation on Ryzen7 either with integrated graphic or external (at least AMD B550) without trap when Virtualaddresslimit is above 1536 and 3140x2160 resolution. If you see the ticket you know what I mean in detail.
2) solving PM (GPI, gradd, Panorama, ??) problems which leads to no window update anymore after some time or when memory gets low (usually seen with VBOX running, but pretty sure not a VBOX problem only, but mem management and graphic)

Regarding browser I fear we never ever will have a browser with the features I need (I'm used to) again. So no priority for me. Dooble works to some extend and I appreciate the efforts made here. But it never will get the features I'm used to from SM or FF.

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Silvan Scherrer on March 20, 2025, 03:39:47 pm
As long as we can't fix the memory issues (4GB border), everything else is more or less useless. And I doubt we ever get this one fixed.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 20, 2025, 04:12:58 pm
Hello.
As long as we can't fix the memory issues (4GB border), everything else is more or less useless. And I doubt we ever get this one fixed.

I guess that just as USB 3 or UEFI was a blocker to run OS/2, I guess this is our new blocker. Not a blocker to run OS/2 on hardware, but to have modern software compiled on it.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: JTA on March 20, 2025, 05:00:53 pm
Always interesting to see how folks solve, or don't solve, technical issues at the OS level. If you can come at the problems with an open mind to all possibilities, the range of solutions also opens up.

If you run OS/2 (ArcaOS) directly on hardware, then yes there will be lots of issues, notwithstanding AN's efforts to work through them. You'll get the 4GB ram limit, no selection of modern browsers or other 64-bit software, and so on.

If you run OS/2 in one (or more!) VM's with a 64-bit HostOS (win, linux) controlling the hardware, then you get all the resources of the 64-bit hardware and 64-bit modern OS delivered to OS/2. You'll get 64-bit apps (via Winflector), NTFS or Linux FS's as the filesystem, no driver issues to speak of, and modern OS/2 (ArcaOS) running at near native speed.

Run two VM's, and you start to move past native OS/2 4GB ram limitations with workarounds (two os/2's in vm's offering up to 8GB ram to native os/2 apps).

Modern hardware is inexpensive, and gives us 64-bit cpu's, 16/32/64gb of ram and ssd's. With Virtualbox or similar, you get at all of that. Once the HostOS is in place, you can spend more time doing things with os/2, versus fighting problems, or worse, giving up on os/2.

Spread this message to new developers, and you'll start to build up that core again of development interest. Think of the possibilities of:
- 1 os/2 vm as your daily driver, serving up any 64-bit app, or even local AI, via Winflector
- 1 or 2 other os/2 vm's dedicated to the os/2 compiler environment
- other vm's hosting specific job engines or other specializations
- HostOS providing security, backup, and more to all vm's
- all running on 1 inexpensive laptop/desktop

How hard is this to do? 1.) Grab a spare machine. 2.) Install Windows Pro (or any Linux). 3.) Install Virtualbox and Winflector. 4.) Install ArcaOS in one or more vm's ...

... and get to work doing whatever you want in OS/2, vs fighting problems in OS/2.

Full details are over in the virtualization sub-forum, but really aren't necessary if you are comfortable with the above 4 steps.


Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Ibrahim Hakeem on March 20, 2025, 06:05:04 pm
With the amount of software we'd need to port to bring OS/2 to a viable level, I'd be interested to see how much we could leverage AI to do some of the heavy-lifting.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 20, 2025, 08:46:51 pm
to JTA:
the idea to put OS/2 into VM host was discussed since 2000, Kim Cheung have spend many efforts to this project:
https://en.os2.guru/projects/reviews/index.php?id=263
(https://ru.os2.guru/projects/reviews/index.php?id=263)

The developers have discussed this in the past.
* yes. this approach is good *for some* users, consumers
* no, this will immediately kill OS/2. no need to develop code for OS/2 if every user can jump to windows/linux and run win/lin application with one click.

So, conlusion: OS/2 is alive while it can run on real hardware, while it's fighting for the life.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: JTA on March 21, 2025, 12:33:48 am
I'm not sure how much it matters if the "idea" of running os/2 in vm's has been in existence since 2000 ... what comes together right now, in just the past few years, is lots of technology ... new/improved:
- hardware: faster cpu's with more cores, lots of ram, super-fast ssd's ... all on one inexpensive machine
- software: virtualization, winflector

If you use os/2, and want to do more with it than running native, more tools in the toolbox gives us options right now. We can do things that weren't available back then:
- virtualbox lets us run multiple vm's on one machine, introducing new ways to use os/2
- winflector lets us run any 64-bit app right now inside of os/2, vs waiting for that browser to come along, or waiting for that port to finish.
- existing os/2 developers can use these new tools in the toolbox to speed up os/2 development right now
- new os/2 developers aren't running constrained on one box, trying to shoehorn everything into 4gb ram on a 32gb machine ... they can expand immediately into more vm's doing more work for them. If we give them a highly tuned dev vm, they can run two or more, alongside their os/2 desktop vm, compiling in some and testing in others.

And, yes ... consumers (all the rest of us) can explore os/2 among many other os's, which shouldn't necessarily kill off os/2. It might expand the pool of folks who find, play with, and come to love os/2 anew:
- os/2 runs perfectly in a vm, yet runs with all kinds of problems natively ... run it in a vm, and you are instantly in business, alongside all your other efforts, on one machine
- explore all the old os/2 (or dos, or win) apps, and perhaps find new ways to bring them back into production ... if it runs in a vm, it's no longer tied to any new hardware/software that forces many os's or apps into retirement.

I respect folks that want to run native ... please do so, if that is your thing.

But adding more tools to the toolbox shouldn't be a bad thing, because who knows what the next generation of os/2 users will do with them. New users might become new developers.

os/2 might be dead (to me) if I hadn't played with it again in the past few years, discovered new tools to resolve all the current issues in this forum, and gotten it to the point where I have time now to explore all the new things I can now do with it, vs fighting it to resolve all its native problems.

progress should be happening on all fronts ...
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 21, 2025, 12:48:05 am
This relates to economics. better tell "housekeeping".

Imagine that OS/2 it's a small village.

you offer import all milk, bread and water from large powerful city located nearby to our small village.
so what will happen? what will happen with the residents of our village?
where the residents will find money to pay for the milk? what they will do without local small factories and farms?

No,
we should generate products by ourselves.
Moreover, OS/2 village is so weak  that it's necessary prohibit relocation to the large city.
(= prohibit usage of vitual machines in OS/2)

Remember, Virtual machines are created by the owners of giant powerful OSes to kill small / legacy / old OSes.

OK, the thread is related to OS/2 Whatsnext.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Mentore on March 21, 2025, 08:29:27 am
Hello.
As long as we can't fix the memory issues (4GB border), everything else is more or less useless. And I doubt we ever get this one fixed.

I guess that just as USB 3 or UEFI was a blocker to run OS/2, I guess this is our new blocker. Not a blocker to run OS/2 on hardware, but to have modern software compiled on it.

Regards

Looking at all the troubles Dave Yeo and Paul Smedley are going through to port Qt 5/6 and Dooble / Otter / other browsers, yes, this is one of the key points we likely will not be able to address.
So two solutions come to mind: open up a niche (not ours) with specific interest, or try to push again for the kernel (at least) source.
Which, likely, is impossible to achieve.

Mentore
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Paul Smedley on March 21, 2025, 09:10:43 am
Not wanting to be negative, but even *compiling* qtwebengine is a PITA due to memory exhaustion - let alone running the thing.....
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 21, 2025, 11:36:29 am
Wifi Driver - what to do?

* the development requires many efforts.
* how many users need wifi?
** all users with os/2 laptops are connected to internet via ethernet already
** yes, wifi will help them if go to public places. yes, it's important in hotels.

* so, how many users really need wifi today, at friday? 10 users or 30 users?
30 users, every one has different model of wifi adapter. manufactured in 2006 year, 2008, 2010, .. 2015, 2022, 2025, ...
* conclusion: every os/2user has different wifi. you will never meet the requirements of our wifi users.


Temporary solution:
there are ethernet to wifi access points.
Blonde Guy mentioned it 100 times:
https://www.blondeguy.com/
https://www.blondeguy.com/TravelRouters.html

I tried to use this adapters during os/2 conferences.
and I failed because hotels interfere into the traffic and show dialogs to verify you, get money, etc. and our web-browser can't process this dialogs, as I understand.
so, this adapters are unusable usually.

So, better spend "virtual 10 000 euro, 20 000 euro"  into web-browser development. instead of driver development (if we see that it requires too many efforts).
So the web-browser will correctly process all this modern javascript, .. dialogs and allow to connect to modern wifi network.

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 21, 2025, 01:14:30 pm
Hello Mentore

So two solutions come to mind: open up a niche (not ours) with specific interest, or try to push again for the kernel (at least) source.


This had been always interesting for me:
1) The Open up a niche, even that I have tried to think about this it gets hard.
2) Pushing for an open kernel. There are dreams of replacing only our kernel and leave the rest (CPI-PM-SOM-WPS) thinking it's running over the OS/2 kernel. I know it may be hard, the most known kernel today is Linux, but I'm tempted to see other kernels (since we are multikernel guys), like Zircon or the Haiku OS kernel. But it required deep knowledge of both kernels and hard work to try to connect CPI and $DRIVERS to a different kernel. I want to brainstorm about this on a different thread.

Regards

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Dave Yeo on March 21, 2025, 04:14:09 pm
Another up coming problem is IPv6, which luckily for us is taking forever to catch on. Eventually we would need a new network stack.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 21, 2025, 04:44:54 pm
1) Seems that web-browser is 100% priority. The users understood this several years ago. Os2.de forum was discussing too .

2) the second main question:
How to use os/2 today?
How to use it to earn money?
So new people will spend money to it, invest into new apps.

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: JTA on March 21, 2025, 05:33:53 pm
All the solutions I listed are tools in the toolbox ... pull them out and use them, or leave them in the toolbox and run os/2 native on a platform. It's everyone's choice; we can't tell anyone what they can or can't do. Unfortunately, there are little to no os/2 tools in that toolbox, for many reasons that are very resistant to change.

If you want to solve existing problems, start pulling out all the other tools and start using them ...

- We can harness the power of other OS's to do work for us under os/2. The HostOS concept does all the heavy lifting in utilizing full resources of a laptop/desktop, but it doesn't have to be seen if you don't want it to ... boot straight into os/2's desktop vm. It feels native, yet you get at all the things that keep many from wanting to use os/2 natively. That is a powerful tool ...

- the dev uses of multiple vm's are compelling ... many of the resources of powerful desktop/laptops sit idle, but with one vm dev'ing, and another hosting the os/2 desktop, you are running "thrusters on full". There are tools to split the dev load among multiple compilers ... this might be applicable to os/2 compiling. Things I haven't even thought of, but might be possible ...

- there are flat out "no device driver" problem, no "browser" problems, no problems of any kind ... because you booted straight into the os/2 desktop vm, and it all just works.

What might be needed, to prove this concept to others, is the ability to produce a USB LiveCD stick of this entire AToF concept. Boot a machine, and show all of these things running. Linux "live cd" versions abound. One problem I'm working to overcome is the IBM/ArcaOS constraint of "IBM doesn't allow ..." demo licenses and such, and ArcaOS custom-stamps ownership into each os distro. This needs to be resolved, but I can't resolve it myself. So, I have some methods to chase down, and hopefully they'll bear fruit ... in the meantime, it's super easy to build the AToF concept, but perhaps a livecd would be even easier for many.

The point of all of this is that there are options now, and these don't require massive funding, waiting for years for the slightest little thing to change, and so on ... use these options if you want, or choose not to use them. It's all good as far as one person chooses. But, if more can see that there are viable options (more tools in the toolbox), why not try them?

Again, progress on all fronts ...
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Klafi on March 21, 2025, 10:13:54 pm
What would need to be changed in ARCAOS to have a modern network?

**ChatGPT Response:**
To implement or improve a modern network in ArcaOS, several changes and enhancements would be necessary, as ArcaOS (a development of OS/2) has a more limited network architecture compared to modern operating systems like Windows or Linux. Here are some key changes that would need to be made:

1. **Updating the Network Stack Software**:
   - **TCP/IP Stack**: ArcaOS uses the "IBM TCP/IP for OS/2" stack, which has not been maintained for a long time. Updating or switching to a more modern network stack (e.g., `inetd` or an open-source project) could improve network performance and compatibility.
   - **IPv6 Support**: The current TCP/IP stack in ArcaOS only supports IPv4. To support modern networks, IPv6 would need to be implemented or at least complemented with a compatible IPv6 stack.

2. **Extending Network Protocols**:
   - **Modern Protocols**: Support for current protocols such as HTTP/2, HTTPS, FTP/FTPS, WebSocket, and other common network protocols should be added to ensure interoperability with modern web services.
   - **SMBv2 and SMBv3**: The current network file system protocol, SMB (Server Message Block), is outdated and uses SMBv1, which is increasingly blocked in modern networks. Implementing SMBv2 or SMBv3 would be necessary to remain compatible with modern Windows and Linux servers.

3. **Driver Updates**:
   - **Network Card Drivers**: Support for modern network cards (e.g., Gigabit or 10-Gigabit Ethernet) would need to be added, or existing drivers would need to be updated. Many modern network cards are not directly compatible with ArcaOS, so specific drivers would need to be developed or updated.
   - **Wi-Fi Support**: ArcaOS has limited support for Wi-Fi. The development and integration of modern Wi-Fi drivers for common chipsets (e.g., Broadcom, Intel, Realtek) and support for Wi-Fi 5 (802.11ac) or Wi-Fi 6 (802.11ax) could expand the network capabilities.

4. **Firewall and Security**:
   - **Firewall**: A modern firewall capable of monitoring and controlling traffic based on current security standards would need to be developed or integrated.
   - **Encryption**: Support for modern encryption protocols such as TLS 1.2 and 1.3 for web communication and VPN services (e.g., OpenVPN) should be enhanced to ensure secure communication.

5. **Integration of Modern Network Services**:
   - **DHCP and DNS**: Support for DHCP (Dynamic Host Configuration Protocol) for automatic IP address assignment and advanced DNS services would need to be integrated or improved.
   - **NAT and VPN**: Support for Network Address Translation (NAT) and VPN protocols like OpenVPN and WireGuard could enable ArcaOS to securely communicate in modern networks.

6. **Updating the User Interface**:
   - **Network Management**: A modernized user interface for network management (e.g., for configuring IP addresses, DNS, Wi-Fi, VPN) could improve the usability of ArcaOS and bring it on par with other operating systems.

7. **Software Compatibility**:
   - **Modern Applications**: Many modern applications and services rely on the latest network standards. Compatibility with these applications requires significant improvements in the network protocols and interfaces in ArcaOS.

**Summary**:
To equip ArcaOS with a modern network, extensive changes would be needed, both at the network stack software level and in terms of drivers and security protocols. While some of these changes could be addressed with new software solutions and drivers, it remains a long-term project, as ArcaOS was not originally designed for use in modern network environments.

**I don't know if all of this will be achieved, but I thank you.**
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 21, 2025, 11:00:02 pm
Let's ask ChatGPT rewrite the code of OS/2 kernel.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Paul Smedley on March 21, 2025, 11:33:26 pm
What would need to be changed in ARCAOS to have a modern network?

**ChatGPT Response:**
   - **SMBv2 and SMBv3**: The current network file system protocol, SMB (Server Message Block), is outdated and uses SMBv1, which is increasingly blocked in modern networks. Implementing SMBv2 or SMBv3 would be necessary to remain compatible with modern Windows and Linux servers.
This part at least is incorrect. ArcaOS has some some time (since 5.0?) included a Samba plugin that supports smb2+
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Dave Yeo on March 22, 2025, 12:18:23 am
What would need to be changed in ARCAOS to have a modern network?

**ChatGPT Response:**
To implement or improve a modern network in ArcaOS, several changes and enhancements would be necessary, as ArcaOS (a development of OS/2) has a more limited network architecture compared to modern operating systems like Windows or Linux. Here are some key changes that would need to be made:

1. **Updating the Network Stack Software**:
   - **TCP/IP Stack**: ArcaOS uses the "IBM TCP/IP for OS/2" stack, which has not been maintained for a long time. Updating or switching to a more modern network stack (e.g., `inetd` or an open-source project) could improve network performance and compatibility.
   - **IPv6 Support**: The current TCP/IP stack in ArcaOS only supports IPv4. To support modern networks, IPv6 would need to be implemented or at least complemented with a compatible IPv6 stack.

This is really the main thing, most of the other points are already covered or things like WiFi, a different discussion.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: David McKenna on March 22, 2025, 12:56:56 am
 I agree the 2 main needs are #1: Browser and #2:WiFi. But like Dave points out there are other currently lesser needs (like IPv6) that may become urgent as time moves on. Also it seems there are other hardware issues becoming apparent - I have a new motherboard that does not have anything for IRQ0, and so has very weird timing issues, as well as only 800MB memory below 4GB. This forced me to go with a VM in Windows - wasn't happy about that.

  Have a possibly dumb question, not being a programmer: is it possible to produce OS/2 executables with GCC inside of Linux? If so, would make compiling large apps easier, no?

Regards,
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Dave Yeo on March 22, 2025, 02:36:46 am
  Have a possibly dumb question, not being a programmer: is it possible to produce OS/2 executables with GCC inside of Linux? If so, would make compiling large apps easier, no?

In theory, it is quite possible. In practice, it would mean recompiling most of our tool chain on Linux as Linux binaries as a start, using our headers and libs and who knows what else. How much porting would be required in the recompilation is unknown, we'd also be porting to 64bit. There may be a chicken and egg problem too, needing cross compiled tools to cross compile.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: ivan on March 22, 2025, 03:26:23 am
Rant !!

I may be a lone voice in this discussion bur what I would like to see is the removal of the RPM/YUM abomination from our Operating System.  We had/have a very good method of installing programs - warpin which does not require OS/2 to pretend to be linux which it is not.  Those that want to play with linux would be better installing  linux and using that.

Rant over.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 22, 2025, 08:41:48 am
To Ivan:  RPM/YUM relates to linux libraries & software only. If you don't use ported software then you never use RPM/YUM.  There was resistance to all this linux stuff in ~ 2011. But it flooded OS/2 and you can't change this. The porters need this infrastructure else they can't put everything to order.

2015 - ArcaNoae have integrated RPM/YUM to ArcaOS.

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 22, 2025, 02:47:19 pm
Hi Ivan
Rant !!

I may be a lone voice in this discussion bur what I would like to see is the removal of the RPM/YUM abomination from our Operating System.  We had/have a very good method of installing programs - warpin which does not require OS/2 to pretend to be linux which it is not.  Those that want to play with linux would be better installing  linux and using that.

Rant over.
You came late to the rant. I have a dislike for the FHS (https://en.wikipedia.org/wiki/Filesystem_Hierarchy_Standard) (what is where the files are located) and not rpm/yum as it is. WarpIn can not replace RPM as it is right now. The good thing about RPM is having a server (or group of server) with the packages and help to manage the packages dependencies. When I started to use RPM/YUM I felt like I lost control of where the files are being located in my HDD, at the end I get used to that because it was not that dramatic for me to be "file control freak", I guess the Linux people got the same anxiety at first.

Sadly there is no man power to avoid FHS. It was explained that porting Linux software to OS/2, plus adding the complexity of relocating all the files in a different way it is too much work.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 22, 2025, 04:25:16 pm
Hi

  Have a possibly dumb question, not being a programmer: is it possible to produce OS/2 executables with GCC inside of Linux? If so, would make compiling large apps easier, no?

I don't have the full knowledge here. But this article is interesting:
- Project: 2ine. You have no idea how much effort went into getting this stupid white square on the screen (Jan 21, 2018) (https://www.patreon.com/posts/project-2ine-16513790)

It is a long story but he used an old gcc to compile a little OS/2 software that will run in Linux (interpretated) and in OS/2. But it was very experimental.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Dave Yeo on March 22, 2025, 07:14:08 pm
I think that David means pure cross compiling, building OS/2 exe's and dll's on Linux.
This is already possible on Linux using OpenWatcom, with a bit of work, it wouldn't be hard to compile Doodles screensaver on Linux as an example. Cmd build scripts would need to be rebuilt using sh scripts, not sure about the Rexx scripts but Linux supports Rexx or at worst ignore them (no bldlevel info), Warpin wouldn't work for packaging, otherwise quite possible.
Same could be done with GCC, at least in theory. People regularly build Windows programs on Linux.
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Andi B. on March 23, 2025, 10:30:00 am
Rant !!

I may be a lone voice in this discussion bur what I would like to see is the removal of the RPM/YUM abomination from our Operating System.  We had/have a very good method of installing programs - warpin which does not require OS/2 to pretend to be linux which it is not.  Those that want to play with linux would be better installing  linux and using that.

Rant over.
If you only read the recent post here about ApacheOpenOffice and understand the problem behind, then you will see very fast how wrong your statement above is. Or do you wanna say we should completely avoid all these ported software?
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 28, 2025, 01:48:44 pm
Hello

For the moment we have in no specific order:
1) Modern web browser
2) Wifi Driver.
3) Memory issues
4) IPv6.

And side discussion:
1) There is no need to worry because we have virtualization.
2) The dislike of RPM/YUM or FHS.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 28, 2025, 01:51:48 pm

I still think that we are falling behind in applications. A modern web browser will help on that, since there are a lot of applications that are on the web without the need to install those, and on the other side compiling and porting applications are now starting to get complicated because of the 32bits/Memory limitation.

A Modern Browser and solving the memory limitations are getting more important for new apps.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Eugene Gorbunoff on March 29, 2025, 10:37:15 am
memory issues:
as I understand, libc uses shared memory.
there were attempts to make it use HighMem.
There were discussions, there were announcements.
there were attempts to patch exe files to use HighMem.

But current situation: you can start Firefox and OpenOffice only 1 time.
if you load this heavy apps then memory is occupied and you should reboot system to use firefox again.
Java VM (I don't remember, does it depend on libc), is consuming low memory too.

Question: what is the current state of libc ? it was modified?

P.S. It's not a claim. I make  technical request.

Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Lars on March 29, 2025, 02:32:38 pm
About libc, read here:

https://github.com/bitwiseworks/libcx

section "Notes on C heap displacement."
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Klafi on March 31, 2025, 11:44:03 am
Dear all,
How about an OS/2 or ARCOS without Windows and DOS? Wouldn't that be easier?

Thanks everyone
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on March 31, 2025, 02:14:31 pm
Hello Klafi
Dear all,
How about an OS/2 or ARCOS without Windows and DOS? Wouldn't that be easier?

Thanks everyone

I personally don't use the DOS VDM and Win-OS/2 as a vital thing of OS/2, but I was amazed to to see some years ago the quantity of people defending it because they run a DOS or Win16 software on it.

If you think that ArcaOS without DOS VDM and Win-OS/2 will make it easier because some patent or fee that is currently part of Microsoft, and will make ArcaOS cheaper, I don't think so. It looks to me that IBM has the right over the MS binaries included in OS/2 and there is no active money exchange anymore. I think of this because the IBM and Microsoft Jointly development agreement that I was able to read, it says that both parties has the right of the source code exchanged there.

I also don't think that DOS VDM and Win-OS/2 is the only limitation for IBM not to open source OS/2. There is also a lot of MS code in PM, Kernel, drivers and also code from third parties (Adobe). Even that I think that Microsoft won't be a problem when open sourcing OS/2, I think IBM does not want to take the risk without a good business case that provides them more benefits then trouble.

But I'm open to other opinion and ideas.

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Digi on April 01, 2025, 04:24:41 am
How about an OS/2 or ARCOS without Windows and DOS? Wouldn't that be easier?

I think that DOS (and Win-OS2) should have been thrown out of the system many years ago. :-)
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Roderick Klein on April 01, 2025, 08:12:08 am
memory issues:
as I understand, libc uses shared memory.
there were attempts to make it use HighMem.
There were discussions, there were announcements.
there were attempts to patch exe files to use HighMem.

But current situation: you can start Firefox and OpenOffice only 1 time.
if you load this heavy apps then memory is occupied and you should reboot system to use firefox again.
Java VM (I don't remember, does it depend on libc), is consuming low memory too.

Question: what is the current state of libc ? it was modified?

P.S. It's not a claim. I make  technical request.

I think its kernel 14.203 from Arca Noae that resolves this issue. High memory with older kernel (if DLL's are loaded high into memory) and then unloaded the memory is not released.
Using the latest ArcaOS kernel this should work.

Roderick
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Roderick Klein on April 01, 2025, 09:49:53 am

I still think that we are falling behind in applications. A modern web browser will help on that, since there are a lot of applications that are on the web without the need to install those, and on the other side compiling and porting applications are now starting to get complicated because of the 32bits/Memory limitation.

A Modern Browser and solving the memory limitations are getting more important for new apps.

Regards

The memory limitation can be solved with magic wand with the current kernel. We might want to wish for this but its not going to happen.
Talk to Steve Levine who helps BWW out with debugging GCC programs. He also helps Paul Smedley to debug programs. But you need to use possibly memory dumps, Thesues and othee tootls to find out WHY a program runs out of memory.
From what I understand from Steve is that for example in the current PHP there is code added to prevent PHP for allocation more memory (Steve if you are reading can you comment please).  While not all limitations of 64 bit apps can be resolved some of them can be resolved by digging in the code for example and add checks to prevent it from consuming to much memory.  But this is of course time consuming process.

Roderick
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Roderick Klein on April 01, 2025, 09:52:42 am
Another up coming problem is IPv6, which luckily for us is taking forever to catch on. Eventually we would need a new network stack.

Its going to be interesting I would need to have more indications about this. What I see happening i that some modems from ISP have IPv4 emulation and tunnel over IPv6 if I get it correctly.
The company I work still has with 800 people IPv4 internally on the company LAN. I agree a multi protocol stack (Ipv4 and IPv6) might be nice to have but we simply lack the human resources.
From what I understand the old socketsk.sys is 98% recompiled FreeBSD code and 2 is OS/2 specific code. The point being we might be able to port a new stack from FreeBSD.
But this will not be done in month and will require a full time developer.

Roderick
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Roderick Klein on April 01, 2025, 09:55:07 am
Hello

For the moment we have in no specific order:
1) Modern web browser
2) Wifi Driver.
3) Memory issues
4) IPv6.

And side discussion:
1) There is no need to worry because we have virtualization.
2) The dislike of RPM/YUM or FHS.

Regards

It would be great if Arca Noae can officialy support QEMU as of version 2.92 or 2.94 QEMU now works with ArcaOS. But we would need mouse and clipboard integration.
I think VMworkstation is less attractive with the unknown commercial goals of Broadcom (read VMware). Sorry there ESXi server prices are a pure rip off. Some company customers I know that have seen in 1 year time
the cost of ESXi go up by a factor of 10. I wonder how price competitive VMware (Broadcom products will remain).

Roderick
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Roderick Klein on April 01, 2025, 09:56:05 am
Dear all,
How about an OS/2 or ARCOS without Windows and DOS? Wouldn't that be easier?

Thanks everyone

Its about 25% of the kernel source is the VDM support. But removing that is not going to make it easier I am afraid to get a new kernel.

Roderick
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Craig Miller on April 10, 2025, 01:02:34 pm
I would like to add two things:

1. Updated Thunderbird or an email program that can connect to email sites like outlook, yahoo, google, etc.
2. Libreoffice

I know they are big projects but I can always hope =)
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Martin Iturbide on April 10, 2025, 02:25:20 pm
Hello

For the moment we have in no specific order:
1) Modern web browser
2) Wifi Driver.
3) Memory issues
4) IPv6.
5) Email app (Modern) compatible with yahoo, gmail, etc
6) QEMU. Not only the software, but guest integration too.
7) LibreOffice

And side discussion:
1) There is no need to worry because we have virtualization.
2) The dislike of RPM/YUM or FHS.
3) A alternative kernel discussion (https://www.os2world.com/forum/index.php/topic,3873.msg47808/topicseen.html#new)

Regards
Title: Re: Brainstorm: What is Next with OS/2 ?
Post by: Dave Yeo on April 10, 2025, 04:49:23 pm
About the email client. We actually have a bug that breaks Oauth2, which causes the login problems. Often, like I do with Gmail you can use an app password. Not sure about Yahoo, but even following the SeaMonkey list, outlook is weird, slightly different way of doing Oauth2 and various servers where only some work.
May be app passwords available for both.
Here's the instructions for Gmail, https://support.google.com/mail/answer/185833?hl=en (https://support.google.com/mail/answer/185833?hl=en)

As for LibreOffice, one of the first things they did when they forked from OpenOffice was rip out all the OS/2 code and since a lot of development has happened there. It would be a huge job and we don't really have anyone working on OpenOffice. Would be nice if OpenOffice was back porting some of the stuff from LibreOffice but there's no volunteers.