Without sounding argumentative (and I may fail at that) ...
Please note that I've been installing & using unix, dos, win, and os/2 since the early '80s, and linux since the early '00s, as a sysadmin and as I moved from company to company, and as each of these OS's came about. At Sematech, I had IBM'ers all over me, over 5 years, as we all sat around and worked out all the offerings of each version of OS/2 and how to use it; then I went on to being an SE for OneUP Corp, again, going around and helping folks with IBM offerings, including OS/2.
I know my way around an OS, and over the decades, I've learned how to get them to do what I want, vs what the vendor of the OS wants me to use them for. I end up with new or creative solutions to the current problems of the day.
You could call me a "newbie" WRT ArcaOS, which I've recently "discovered", and which solves some problems for me. That's exciting!
I appreciate all the efforts that folks are making in various attempts to rework a kernel underneath OS/2 that I read about on these forums ... licensing issues aside, I think it's great when folks push the boundaries of the current box. To avoid licensing issues, and 32-bit limits, we may have to think outside of the box. And, virtualization helps with that ...
I appreciate all the efforts of those folks who are working hard & tirelessly to get things "ported" to OS/2. I contributed devbucks to BWW, & made a license purchase to ArcaOS, both of which further increase my options to get working solutions assembled.
One may not like virtualization, but it seems harder to argue with all those that are doing it, either at the enterprise level, or now, at the home level. One can always choose not to use it, when assembling a solution ...
If an issue is strictly performance, there's all kinds of arguments with speeds and feeds, choices in hardware/software, configurations, etc., and with folks that love to split hairs. All are constraints and choices that hold back solutions. If an issue is, rather, "how can I solve a particular problem", or "how do I do this right now", given what's all around us, then things should open up.
If one is running OS/2 on a physical box, yes, one would be limited by constraints/choices and OS/2 at 32-bits, and to something that is possibly augmented by X-windows, RDP/VNC. If running Win7 under vbox under os/2, resources would, understandably, be limited. I don't say this *for* anybody, of course ... these are others' constraints/choices, not mine. I don't tell anybody how to do anything, but there are *other* possibilities, and other *choices*. If one is running OS/2 virtually, then options open up ...
Again, it is possible to use virtualization to run x64 apps from within OS/2, under your control, against OS/2 datasets, and not via X/RDP/VNC ... I'm doing this right now (beefy PC, Host-OS, v-layer, OS/2 in vm, other vm's) ... all I did was re-assemble the parts and pieces of today's technology, to solve a problem that others' posted about. It's fast ... all the vm's are fast. The solution is fast. I didn't program (not really, although programming might make the solution even better), and I didn't violate anybody's current licensing schemes (as I've looked into them). x64 CLI, x64 GUI ... yep, against OS/2 datasets, under OS/2 ... simple plumbing.
I *might have* thought outside of the box, and amazingly, it just works ... initially, this post was to see who else was using virtualization, and if the solution was already apparent to all. Perhaps it isn't ... I'm not sure yet what that tells me.
I've learned over all these years never to say I *can* do it, until I *have* done it. And if I've learned anything, it's to not abandon the enthusiasm, for a new day, and a new way to do things, with OS/2 or with any new technology ...