Print

[rayarachelian]

Apprently back in 2016 there was a Lisa demo scene that we didn't know about.https://bytecellar.com/2016/03/29/an-interview-with-cmucc-members-behind-the-apple-lisas-first-scenedemo-le-requiem-de-lisa/
And an interesting hack that displays color at the end!https://www.youtube.com/watch?v=ViroWYBvNRE
Neat!

[stepleton]

Oh I knew about it

But regrettably I didn't know about it in 2011 when it was first developed, at my graduate school alma mater no less, and within walking distance of my house at the time. I found out in 2015, just before moving from Pittsburgh to London. I wish I could have paid them a visit...

Don't miss their award-winning second demo if you don't mind a rickroll. (It's also mentioned at the bottom of the linked article.) This one uses a custom soundcard that they made for the machine.

I have to wonder: have their demos pushed the Lisa to the limit, or could you something more impressive with a bit of persistence? I have to imagine that if they could have got the beeper to play something more interesting than a beepy soundtrack (like those programs that play lo-fi audio on PC speakers), then they would have done it.

[rayarachelian]

Don't miss their award-winning second demo if you don't mind a rickroll. (It's also mentioned at the bottom of the linked article.) This one uses a custom soundcard that they made for the machine.

Very neat! Did they release schematics for the sound card they've built? Wonder how it compares to the LisaDAC.

[Lisa2]

Here is some more information on this:
https://coda.s3m.us/2012/08/06/creation-of-the-ilisa-demo-soundtrack/

This is a quote from the developer  in 2012 about the custom sound card that I had in my archives:

"Currently, it acts as a valid bus-accessible device, but we've only used it from programs running in kernel mode from Lisa Workshop that are able to directly access the add-in slot addresses.  (Needing nearly all the RAM, requiring deterministic execution time, and needing to slot in our own interrupt handlers for these demos, everything we've written has a loader that exploits a "vulnerability" in the kernel to take over and sweep it aside.)  It should conceivably be quite possible to write drivers for either Workshop or Lisa Office that provides proper user-mode access. Interface-wise, it's a bunch of memory-mapped registers that activate operations or allow the uploading of audio samples.  Given such a driver, it wouldn't be very hard to write a program that plays raw WAV or OPL (e.g. Adlib files from old DOS games) files.  Brief experimentation indicates that compressed audio (e.g. MP3) is laughably beyond the machine's abilities.

However... our current feeling is "it's been fun" and that we've shown off some ridiculous stuff that can be done with this machine but that it's both so limited in capability and rare that it's probably not worth our while to keep going.

At this point I'm inclined to just release the gerber CAD files, bill of materials, the CPLD programming binary, and programming interface to the world so that anyone that really wanted one could build one.  (We're going to hold onto schematics and CPLD source for now.)  The BOM probably runs $150 each for small orders of parts, and the assembly is easily a day or two task for a highly skilled surface-mount solderer if done by hand (some chips with very small pitch pins and 100-some bypass capacitors you mostly don't see on the back).  Automated assembly would make that a lot more sane but isn't going to be economical for small batches."

To answer Ray's question: this is a completely different animal than the Lisa DAC card that allows Macintosh sound support under MacWorks.

Rick

[rayarachelian]

everything we've written has a loader that exploits a "vulnerability" in the kernel to take over and sweep it aside.)  It should conceivably be quite possible to write drivers for either Workshop or Lisa Office that provides proper user-mode access.

I think I've seen this before in tracelogs of LOS, it's more of a trampoline. Off the top of my head if you invoke an opcode that starts with 0xa1234568 it will execute a trap that will then ultimately JSR to address 0x00234568 or something like that. The Mac instead, has real A-Line traps with a look up table instead of this. The Lisa wasn't very security savvy, even though it did have an enforcing user/supervisor mode enabled and MMU memory protection.

I'm surprised they bothered building this for LOS, they could have just directly written their own boot loader (or used Tom's or even reused LOS's boot loader) that brings up their app even without any OS, but hey, sure, why not!

Who knows, maybe they needed the LisaGraf routines.

That sound card sounds pretty neat!