I profoundly admire people who keep maintaining systems in old or arcane architectures, they are the real unsung heroes. However, at some point, these resources could probably be better used elsewhere.
To me, it seems NetBSD isn’t willing to let go if it isn’t working, and just keeps architectures around to improve credibility without proper care and feeding. OpenBSD made sure everything was working, and if push came to shove, could let go easily.
That did bother me it was still included as “supported” despite how it was decaying. Isnt helping their credibility. Now Im thinking I would have to do thorough functional or stress testing on about any other odd architecture on list. So, let’s check on it.
I like the three categories. Fairly accurate way to divide much software. VAX as described in main post should probably be considered Life Support. Or about to be terminated by apathetic parent. Not quite growing organically. Now, wonder how many more on this list are on Life Support.
Apparently just not much interest in FOSS for it as they don’t have dependence on legacy apps for VAX. ;) The bigger loss for me was Alpha ISA given it had a nice balance of simplicity, performance, and PALcode. PALcode, a microcode alternative, could more easily handle things like database transactions or encryption resistant to timing analysis that took way more work on current CPU’s. Itanium has some kind of PALcode per a powerpoint I saw but I’m not sure if it’s open to developers. Alpha I.P. got acquired by Intel, jointly licensed with Samsung, had at least one buyer, got cloned by Chinese in ShinWei, and then disappeared far as I can tell. If I get a chance, I’m going to try to convince them to dig it up and dual-license it openly a la OpenSPARC. Might be more benefits to I.P. holders and OSS projects given new projects are still interested in it (eg SAFE CPU is Alpha ISA).
Likewise, VAX software started to get non-maintained, I.P. for CPU bought by company that EOL’d it, they killed its badass OS (VMS) as well, Itanium successor to Alpha + PA-RISC is on life support, and HP grudgingly brings VMS back to life support for Xeon port by smaller 3rd party. This is all collectively a terrible end to the VAXen that let people do so many great things. Often the case for great technologies it seems. Especially if their management started to be total idiots. (cough) DEC (cough) Compaq (cough) HP (cough)
PALcode is just emulating instructions in firmware - that is, it’s executing Alpha instructions from unimplemented-in-silicon instructions. You can do it on x86 by trapping in the SMM.
I profoundly admire people who keep maintaining systems in old or arcane architectures, they are the real unsung heroes. However, at some point, these resources could probably be better used elsewhere.
This brought back good memories from college :)
Disclaimer: I’m not a NetBSD user
VAX was one of those systems I thought people would keep maintain until the end of times.
Interestingly, OpenBSD also recently dropped support for their VAX port (http://www.openbsd.org/vax.html).
[Comment removed by author]
To me, it seems NetBSD isn’t willing to let go if it isn’t working, and just keeps architectures around to improve credibility without proper care and feeding. OpenBSD made sure everything was working, and if push came to shove, could let go easily.
That did bother me it was still included as “supported” despite how it was decaying. Isnt helping their credibility. Now Im thinking I would have to do thorough functional or stress testing on about any other odd architecture on list. So, let’s check on it.
https://www.netbsd.org/ports/
I like the three categories. Fairly accurate way to divide much software. VAX as described in main post should probably be considered Life Support. Or about to be terminated by apathetic parent. Not quite growing organically. Now, wonder how many more on this list are on Life Support.
They are. Just in a smarter way:
http://logical-co.com/nuvax/
Apparently just not much interest in FOSS for it as they don’t have dependence on legacy apps for VAX. ;) The bigger loss for me was Alpha ISA given it had a nice balance of simplicity, performance, and PALcode. PALcode, a microcode alternative, could more easily handle things like database transactions or encryption resistant to timing analysis that took way more work on current CPU’s. Itanium has some kind of PALcode per a powerpoint I saw but I’m not sure if it’s open to developers. Alpha I.P. got acquired by Intel, jointly licensed with Samsung, had at least one buyer, got cloned by Chinese in ShinWei, and then disappeared far as I can tell. If I get a chance, I’m going to try to convince them to dig it up and dual-license it openly a la OpenSPARC. Might be more benefits to I.P. holders and OSS projects given new projects are still interested in it (eg SAFE CPU is Alpha ISA).
Likewise, VAX software started to get non-maintained, I.P. for CPU bought by company that EOL’d it, they killed its badass OS (VMS) as well, Itanium successor to Alpha + PA-RISC is on life support, and HP grudgingly brings VMS back to life support for Xeon port by smaller 3rd party. This is all collectively a terrible end to the VAXen that let people do so many great things. Often the case for great technologies it seems. Especially if their management started to be total idiots. (cough) DEC (cough) Compaq (cough) HP (cough)
PALcode is just emulating instructions in firmware - that is, it’s executing Alpha instructions from unimplemented-in-silicon instructions. You can do it on x86 by trapping in the SMM.
Appreciate the tip. But is it as efficient and easy? I never tried it on x86.
Why use a vax when a Raspberry Pi is faster?
[Comment removed by author]