LisaList2

Advanced search  

News:

2019.06.07 fixed NChat for the "Curve" theme, will eventually move it to its own page and add it to the default theme as well. Other plugins are next. see post in the Meta board for details

Pages: [1] 2 3 ... 10
 1 
 on: March 31, 2020, 08:23:36 pm 
Started by stepleton - Last post by stepleton
The thought had crossed my mind, but my style is pretty basic :-)
Interrupts shminterrupts...

 2 
 on: March 31, 2020, 08:12:40 pm 
Started by stepleton - Last post by rayarachelian
Yeay! Glad it's published!

You know this is how operating systems get written, they start out as small tiny things like bootloaders, or in the case of Linux, a terminal program. :)
(hint hint)

 3 
 on: March 31, 2020, 07:25:17 pm 
Started by stepleton - Last post by stepleton
Hi list, I hope you're all doing well...

I've made two new low-level software items for Lisa. First, there's a hard drive bootloader:
https://github.com/stepleton/bootloader_hd
It's not too sophisticated: it doesn't know about filesystems, and it works a lot like my floppy bootloader --- it just loads blocks into RAM, starting from block $000002 until it sees it should stop.

The bootloader uses the second project: a low-level I/O library in 68k assembly that I've started here:
https://github.com/stepleton/lisa_io
Right now it has components for keyboard and mouse input, fixed-width text display on the screen, and reads/writes to parallel port hard drives. No floppy or serial port support yet---maybe someday later.

You can see a boring demonstration of both on YouTube here:
https://www.youtube.com/watch?v=4WP4kkj1mRM

I hope these will be useful. Please stay safe!

 4 
 on: March 16, 2020, 09:59:34 pm 
Started by rayarachelian - Last post by rayarachelian
Hmmm, looking at this again, it has 64 pins on the connector, the Lisa has only 56. I did see the Lisa 2/10 motherboard have extra pins upto 64 but the ZIF connectors do not. So perhaps this was something considered before the 2/10 was released, or it was for some totally different machine, or something that was supposed to be backwards compatible with the Lisa's expansion slots.

Or they happened to have slightly longer ZIF connectors and tooling for Lisa PCB boards and they built this thing from the same stuff they had available.

There's one tell there, there's a hole right above pin 64, just like on normal Lisa cards, there's a hole right above pin 56 to hold a brake pin.

Whatever this thing was, the case it wanted to be in was at an angle, just like the Lisa's on the side facing the outside.

 5 
 on: March 16, 2020, 07:41:32 pm 
Started by rayarachelian - Last post by rayarachelian
Sure, very possible, but:

https://www.folklore.org/StoryView.py?project=Macintosh&story=Good_Earth.txt&sortOrder=Sort+by+D. says the 6809 Mac was built in 1980 and then Burrel convinced Jef to move to the 68000, and in https://www.folklore.org/StoryView.py?project=Macintosh&story=Texaco_Towers.txt says the Lisa team had moved to Texaco Towers in 1980, and the Mac team was left in the Good Earth building ~ Dec 1980/early 1981.
By Feb 81 Andy Hertzfeld had joined the team, taking over Jef's desk, and that "(in) Bud's office, on one of the empty desks, was Burrell's 68000 based Macintosh prototype, wired-wrapped by Burrell himself".
(also https://www.folklore.org/StoryView.py?project=Macintosh&story=Black_Wednesday.txt&sortOrder=Sort+by+D. )

So they already had a protoype 68k Mac in 1981.

Edit 2020/03/17: I've found this photo of a 1981 prototype Mac board on bitsavers: http://bitsavers.org/pdf/apple/mac/prototypes/1981_ACIA/MC1001-01_PCB.jpg - it's pretty much a complete PCB with its own VIA6522 chip used for the mouse(?), and a pair SYN6551 ACIAs (possibly used instead of the z8530 SCC, which were likely replaced by the z8530 later) - the other big difference is that the Mac, at least this one, did not use parity memory - there's exactly 16 RAM chips (the gold 4164s on the bottom right), where as the GLM used parity since it had 18.
So, point is, by 1981 they already had a protype PCB, so they certainly would have made a bunch of them and wouldn't need something like the GLM to do prototyping on. I'm pretty convinced the GLM had very little to do with the Mac. It may or may not have been related to the Lisa 2/10, but it wasn't anything to do with the Mac project.

https://www.folklore.org/StoryView.py?project=Macintosh&story=Nybbles.txt&sortOrder=Sort+by+DateThis is also set in Feb 81 and describes the IWM controller for the Mac (the Lisa was still using the 6504)

https://www.folklore.org/StoryView.py?project=Macintosh&story=More_Like_A_Porsche.txt&sortOrder=Sort+by+Date set in March 81 mentions a serial link between the Lisa and the Mac - this implies two separate machines, not a Mac on a Lisa card and describes that they had already signed off on the case and that the Mac case signing party happened in Feb 1982 ( https://www.folklore.org/StoryView.py?project=Macintosh&story=Signing_Party.txt )

https://www.folklore.org/StoryView.py?project=Macintosh&story=Square_Dots.txt&sortOrder=Sort+by+Date is set in April 81 and talks about square vs rectangular pixels.

Also in April 81 there's a prototype finder: https://www.folklore.org/StoryView.py?project=Macintosh&story=Early_Demos.txt&sortOrder=Sort+by+D. (possibly twiggy mac from the screenshot, but it's a 5.25" floppy on that screen.):

Quote
In June 1981, we realized that it would be worthwhile to create a stand-alone demo environment, where the Macintosh booted and ran programs from its own disk, even though we'd only use it temporarily. Our own operating system wasn't close to usable yet, but Rich Page had written a simple operating system called the "Lisa Monitor" which was based on UCSD Pascal, that was pretty easy to port - all we had to do was integrate our I/O drivers. Soon, using the Monitor, a Mac could boot up and run demos without help from a Lisa.

By May 81, roundrects were added to the Mac, and here's a photo of the square display: https://www.folklore.org/StoryView.py?project=Macintosh&story=Round_Rects_Are_Everywhere.txt&sortOrder=Sort+by+Title%27A=0

https://www.folklore.org/StoryView.py?project=Macintosh&story=Pineapple_Pizza.txt&sortOrder=Sort+by+D. set in May 81 says:
Quote

When I began working on the Mac project in February 1981, there was still only one single 68000-based Macintosh prototype in existence, the initial digital board that was wire-wrapped by Burrell himself. It was now sitting in the corner of Bud Tribble's office, on one of the empty desks, attached to a small, seven-inch monitor. When powered up, the code in the boot ROM filled the screen with the word 'hello', in lower case letters and a tiny font, rendered crisply on the distinctive black-on-white display.

Dan Kottke and Brian Howard were already busy wire-wrapping more prototype boards, carefully following Burrell's drawings. In a week or so, I received the second prototype for my office, so I could work on the low level I/O routines, interfacing the disk and keyboard, while Bud worked on the mouse driver and porting Bill's graphics routines.

So this says that they already had prototype macs before that board was built. And in fact it refers to PC boards being delivered, and how they attached them to an Apple II power supply and how when first powered on they saw a checkerboard pattern. So for sure, by June 81 they would have had enough prototype Macs made to not need something like a GLM board.

The GLM card has a copyright of 1982 - so by then they already had Macs or at least standalone Mac prototypes, they wouldn't have needed them as addon cards to a Lisa.

https://www.folklore.org/StoryView.py?project=Macintosh&story=Macintosh_Prototypes.txt&sortOrder=Sort+by+Date mentions the prototype boards and it's set in June 81, and it mentioned that board " #3 MC1001-02 Feb1982 build of 50 " Nothing there mentions the GLM, and the GLM is labeled "6XXX-XXX-XXX" at the top.

I agree that the GLM has two serial ports and they say that they debugged the Mac off the serial port, but why would they need to when the GLM was directly attached to the expansion slot bus? And that GLM has no other external ports, and we have photos of square screens from Mac prototypes before 1982. We know it had an IWM and a pair of VIAs by then from these and other folklore.org stories.

So, no, I don't think that's what this board was.

Agreed, if it had the ROMs we could dump them and look for strings and also figure out what code it had and what it was trying to do.

 6 
 on: March 16, 2020, 04:37:16 pm 
Started by rayarachelian - Last post by Lisa2
To me this sure looks like a Macintosh: 68000, 128K RAM, SCC, etc.  Possibly intended for software macintosh development or Mac hardware emulation. (like a hardware version of MacWorks).

I think this was just hosted in a Lisa (type) card slot (as we know it will not fit in a Lisa) for power, and may use the Lisa's FDC, Keyboard, Mouse.

Kind of a boat anchor without the ROM's...

Rick

 7 
 on: March 16, 2020, 04:00:14 pm 
Started by rayarachelian - Last post by rayarachelian
Yeah, not much more info on those links either. Just, kind of "here is this thing, and some photos of it, it's cool, you know."

This is interesting: http://www.digibarn.com/collections/systems/apple-lisa1/lisa1-prototype/apple_glm_boards.jpg as it shows it in use with a much larger board underneath it, but can't tell what it is, and certainly it could be the original Lisa's design, but it wouldn't fit in the chassis.

Oh well. I can guess about the serial ports, but they're not necessary from the point of view of the Lisa as internal communication on the expansion slots would be sufficient. So those are meant to be external communications ports.

Since the Lisa has 3 slots, you could add 3 of these to one and you'd have a 4 core system. It's likely IMO the serial ports are meant to turn the Lisa into a unix server. They're also fast enough for LocalTalk (nee Appletalk back then), but possibly that wasn't what they're for because the board says 1982 and we didn't have AppleTalk until later.

The diagonal slant on the outside of the card shows it was intended to live inside a Lisa chassis as that matches other cards, but it's actually too deep to fit in a Lisa 1 or Lisa 2. So that means a different Lisa chassis would be needed and we never saw much larger Lisa prototypes.

I don't know how much RAM is on these boards, hard to read the chips, I think it says HM4864, which are 64x1bit DRAM? but there's 18 of them, which means 16 bits plus two parity. (https://i.ebayimg.com/images/g/CaMAAOSwvTpaAeEk/s-l1600.jpg form https://www.ebay.com/itm/DRAM-64kx1-Dynamic-RAM-64k-x-1-200ns-HM4864-3-MB8264-20-CERAMIC-DIP-8-pieces-/261614573015 and https://www.datasheetarchive.com/HM4864-2-datasheet.html )

There's a row of empty holes on the board labeled U1B-U14B, and extras at U1D-U4D - possibly for expansion RAM as that's another 18.

So that would give you 128K since there's 16 bits, and 256K if both were populated or just 64K as is in the photo.

Likely there's some shared memory that's accessible to the Lisa expansion slot bus to allow the CPU on the CPU board to communicate with these.

The ROM sockets say 2764, those are 8Kx8 and two of them, so those aren't boot ROMs, but rather for the 68000 on the card, possibly some portion of the space there was shared over the expansion slot to server as an expansion card ID, or perhaps another ROM was used elsewhere on the board for that.

So that's a lot of RAM for this card. Lisa RAM cards were 512K, so that's half of one of the Lisa's RAM cards!  Possibly if they planned for higher density RAM they could add a lot more.

More than enough to spin off another process to one of these CPUs, and memory page the process from main memory to this board, from say a Unix process and also act as a pair of TTYs. Possibly more than one process could be affinitized (if that's a word) to one of these depending on how much RAM it needed. But likely there'd be no MMU at all - and perhaps it would work even better as, since each process would run on its own card and page/swap to main memory, and then to ProFile/Widget, it would provide a higher level of security (unless more than one process was sent simultaneously to share memory in one card.)

So if they could have shrunk these boards to fit the chassis, or made the chassis bigger, this would have effectively made the Lisa a very nice Xenix/UniPlus server. Not quite a mini, but certainly a server for upto 8 terminals. Or perhaps they may have planned a much larger chassis with many expansion slots instead of just 3 and you could plug in other cards such the as two port or AppleNet cards as well, and they wanted to keep that compatible with existing expansion slot cards.

Anyway, just all conjecture on my part, but still, plausible, so my crazy guess is they were planning for a large Lisa server, most likely for some kind of Unix.


Well either that, or this was the original CPU board for the Lisa and then RAM, Z8530 were moved to RAM boards and an I/O board respectively.

 8 
 on: March 15, 2020, 09:45:24 pm 
Started by rayarachelian - Last post by jamesdenton
I read about this once here:

http://www.digibarn.com/collections/systems/apple-lisa1/lisa1-prototype/index.html

And in fact, the archives have at least one thread:

https://lisalist2.com/lisalist1/2513.html

A little more detail:

https://www.computerhistory.org/collections/catalog/X2435.2002A

 9 
 on: March 15, 2020, 06:02:40 pm 
Started by rayarachelian - Last post by rayarachelian
https://www.ebay.com/itm/Apple-Lisa-Great-Little-Machine-Dual-Processor-Prototype-Card-Vintage-Apple/223947491552
Seems to be an Expansion port card that adds a second 68000, ROM slots, some RAM chips, and a Z8530 SIO providing two more serial ports.
 But why? What was its intended use?
The card also appears to be too large to fit in the Lisa's chassis.

 10 
 on: March 15, 2020, 11:06:56 am 
Started by rayarachelian - Last post by rayarachelian
I renamed this thread to Beta to match the current release state.

Just cut a beta release 2020.03.14 and pushed source to github last night, and uploaded to the downloads page:Most of the non-display bugs are fixed, but lots of display bugs are still there. Once those are fixed, there's one or two more features to add, and then will move to release. I can't say when that will be as the display bugs are many and fugly.

On macos x the display refreshing is really badly broken, so I need to do a lot more work there. The good news is that I was able to create a Mountain Lion ( 10.8 ) VirtualBox VM with XCode 4.6.3 - I had to use macports as brew.sh (aka homebrew) does not support it.
In theory I used the 10.7 SDK, so it might work on macos x 10.7, but as I don't have a Lion machine to test it on, I can't yet tell. This uses wxWidgets 3.0.4 as 3.1.x will not build there.
I think for the final version I'm going to do some shell scripting tricks on MacOS X to glue several binaries together - without using lipo. Instead of launching a Mach-O binary for LisaEm, I'll run a shell script which will detect the OS and what binaries are available and run the latest one that the current OS will support. So if you're on PowerPC it will launch the G4/G5 version. If you're on an older intel it will try to run one that works. If you're on catalina it will launch one with wxWidgets 3.12 or 3.13, etc.
That is assuming the next macos x doesn't remove the shell altogether.  :P
I probably need to write some automation, scripts as well, to cut releases like these, (creating the DMG and populating it, pushing source to gitlab and binaries to the download page) as doing them by hand is getting annoying.

Pages: [1] 2 3 ... 10