Hacker News new | past | comments | ask | show | jobs | submit login
Symbolics Lisp: Using the DEC Alpha as a Programmable Micro-Engine (1993) [pdf] (withington.org)
78 points by kristianp 23 days ago | hide | past | web | favorite | 27 comments



OpenGenera on Alpha runs amazingly faster than any of the real iron Symbolics hardware, including the NuBus MacIvories that got embedded in old Macs like the Quadra. You can also put plenty of memory and disk into one of these often server-grade machines. The negative? You don’t get to use the amazing Symbolics keyboards. You have to deal with Tru64 UNIX. You have to run everything through X.

AFAIK, the unofficial x86 port of OpenGenera is an emulator for Alpha which emulates the Lisp machine, and all together it’s still much faster.

I had the pleasure of figuring out how to get X keyboard codes from a Linux box into the emulated Genera. What a pain! As far as I know, the “world building” code has been lost so it’s not even possible to bootstrap a Genera system from scratch. (You can dump worlds from memory of an existing Genera system though.) I think this limitation made it difficult to change some hard coded constants, like the size of the communication buffer which was essentially shared memory between the running Genera system and the host operating system.

I wish this software could be preserved properly, instead of sitting under a rain cloud of uncertain legal standing. I don’t remember exactly how the story goes, but Symbolics software fell into the hands of Andrew Topping, who died, and whose estate went to John Mallery (at one time famous for writing a Lisp HTTP web server than ran the White House website), and he had held onto the IP rights since, with no indication of selling or releasing it.

(Common Lisp historians / Symbolics enthusiasts, please correct any errors I’ve made. :))


> the unofficial x86 port of OpenGenera is an emulator for Alpha which emulates the Lisp machine

You can run the official Open Genera emulator in an Alpha/Tru64 UNIX emulator on top of an Intel CPU.

But the 'unofficial' port is actually a direct port of the Open Genera emulator to 64bit x86.


Which Alpha/Tru64 UNIX emulator do you have in mind?



The 'basic' version of this costs 400 euro a year.

alphavm_free which was removed by the creator - probably because he wanted to increase sales of his basic offering - can still be found on the Internet however.


I can confirm that John Mallery has the IP rights to Genera and has no interest in selling those rights.

I also have a 1U rackmount DEC Alpha that I will give free to anybody in the Bay Area. I intended to use it to run Open Genera but ended up purchasing a Mac Ivory instead.


> The negative? You don’t get to use the amazing Symbolics keyboards.

Can't that be solved by ordering PCBs, your favorite switches, and custom key caps? BTW, has nobody 3D scanned those keycaps on their newer keyboards? The older model seems pretty standard sculpted spherical top ones.

> AFAIK, the unofficial x86 port of OpenGenera is an emulator for Alpha which emulates the Lisp machine, and all together it’s still much faster.

Is there a place this unofficial port can be seen?


The originals use rather exotic hall effect switches. There are decent (if expensive) replica keycaps around, but they're not in the same profile etc.


The switches, which is what drives most of the "feel" of the keyboard, must have been standard parts (the older ones were, AFAIK, Honeywell) and I would be surprised there aren't any manufacturers of similar parts making them now.

We can try at least to characterize the action (force x displacement curve, contact point, audible feedback) of the switches to find something similar. As for the keycaps, 3D printing will only get better. We may need to wait a couple years, but we'll have them.


There are a few modern hall effect keyboards, but I don't know if the feel is similar to the old Honeywells.

For the keycaps, there is probably enough interest (given the amount of replica sets) to just cast a run of something extremely close to the originals with the usual mold injection techniques.


The keyboard fanatics over at https://reddit.com/r/mechanicalkeyboards can probably hook you up. The "space cadet" keyboard is revered there.


This is more Deskthority https://deskthority.net/ country I'd think. Thomas Ran has YouTube videos on the Space Cadet keyboard https://www.youtube.com/watch?v=oDozftThFMw and several other Hall-effect specimens.


Thomas's videos are awesome. Anyone who touches a keyboard regularly should watch.


Man they really don't make them like they used to. I thought my cherry keyboards were 'leet.


The only thing I’d add is that the emulator has since been ported to Linux and it’s perfectly possible to emulate a lisp machine on a modern Linux distribution (you can find instructions online for Ubuntu, but it should be fairly easy to figure out how to adapt them for an arbitrary distribution). However, the IP issues make running such a system a somewhat grey area.


The emulator you have in mind (Brad Parker's) has a lot of bugs and is not exactly stable.


> The negative? You don’t get to use the amazing Symbolics keyboards.

With a special USB-Adapter one can use them on computers with USB.


I have one such PS/2 adapter and for me it works quite awfully. The key timing is all wrong, key repeat is especially prone, etc. Maybe I was just quite unlucky.


This link is to an extended abstract. Is the full paper available?


About 15 years ago in comp.lang.lisp Scott McKay (a coauthor) was asked and says that PLDI didn't accept it, so they didn't finish the paper. https://comp.lang.lisp.narkive.com/9qf4g9bV/looking-for-full...


> This link is to an extended abstract. Is the full paper available?

The website of one of the authors is

> http://pt.withington.org/

Just ask him. :-)


The website contains the same abstract.

I find it strange that the author would not publish the paper on his website but would submit it to random people that ask for it on email.


> I find it strange that the author would not publish the paper on his website but would submit it to random people that ask for it on email.

I don't know whether he will submit it to random people who ask for it. But he will probably be able to answer whether a non-abstract version exists. :-)


I've been looking for the full paper everywhere in the last years. It doesn't seem to exist online. Maybe in someone's private computer.


The most pleasant thing about this is how the lisp machine's instruction set turned out to be a good bytecode for interpretation. Calling it "microcode" seems risible today, but I don't know if it was as ridiculous in 1993.


Not at all, it was pretty common even during the 60 and 70's, e.g. Burroughs and Xerox PARC systems.


The "microcode" is the bytecode interpreter on the physical machines.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: