The New Scientist is running a rather interesting story, “Taking a trip down memory-chip lane,” which is linked through a Slashdot story that is full of great user commentary. The article looks at the recent surge in retro computing interest, both as a hobby and as a serious means to get at old data that lies fading away on floppy disks and data tapes. The article identifies retro computing buffs as the ones with the resources and know-how to get at this old data. “Computer archaeology” they’re calling it. A term that fits, really.
In a rather obscure tangent, I was introduced, in the /. comments thread, to the phenomenon of John Titor, an alleged time traveler who posted his story to various Internet bulletin boards during 2000/2001. He claims to have been sent back to 1975 to recover an IBM 5100, needed by the government to help get past the Unix timestamp bug that (as expected) surfaced in the year 2038. A crazy story, but with a nice vintage computing slant.
I’ve got a good number of old machines in my computer room here. A prime concern is always how to get data (disk images, etc.) from the Internet to these old machines. Some, like my NeXTStation, are friendly LAN citizens. Others, like my Mac Plus, have Ethernet connectivity, but can be challenging to work with over the network. My Amiga 1200 has the network hardware, but I’ve not gotten the drivers squared away to make it all happen. There are but 24 hours in a given day.
My Apple IIgs used to talk fairly happily via LocalTalk to my PowerMac G4 that was 1) equipped with an old-school serial port and could 2) boot into OS 9 natively. Since I handed down the G4 and moved to the G5, things have been more difficult. As a solution, I decided to equip my IIgs with a SuperDrive, which is what Apple used to call its 3.5″ floppy drive that could read/write both variable speed 800K floppies as well as standard 1.44MB media. I found one on eBay, new-in-box, which is always best! More elusive and expensive is the requisite controller needed to interface the drive to an Apple II. Fortunately, no sooner had I posted an inquiry to the Usenet than I received a note from someone in Australia who had a new-in-box (!) controller that he agreed to let go for $75 USD.
Both arrived, installed easily, work great, and now I’m reading PC formatted 1.44MB floppies in the IIgs real horrorshow. Problem solved.
I took some pics of the new units as they emerged for the first time from their boxes, here nearly 15 years after they left Apple. There’s a few general GS hardware shots in there as well. Have a look.
Now, here’s something you don’t see every day. Michael J. Mahon has created the AppleCrate, a parallel computer comprised of eight Apple IIe motherboards running on Mahon’s NadaNet network. From Mahon’s site:
In 1996, I began thinking about how one might network Apple II computers using only its built-in serial I/O: the pushbutton inputs and the annunciator outputs, and wondered what could be done with such a network. I have worked intermittently on this project over the intervening years, more frequently since I have retired. This document describes its current state.
The possibility of creating a useful network using only wire and software was esthetically appealing. Because it initially required no hardware other than the connecting wires, I dubbed it “NadaNet” (“nada” is Spanish for “nothing”). It was an interesting challenge to design an Ethernet-like network from the ground up. The exercise provided an experimental vehicle to illuminate the various issues and tradeoffs in creating and using such a network. It also became a tool for exploring various higher-level applications of networking, such as client-server and parallel computing.
To add more processors and save space, I decided that I would package several Apple //e main boards together, without keyboards or peripheral slot cards. I settled on a wooden cube about one foot on a side which I slotted to hold up to 8 main boards.
A rather impressive piece of work. How many IIe motherboards would it take to outperform VA Tech’s Terascale 1100 node Xserve G5 supercluster?
Most every day I take a look at the first few pages of auction items in eBay’s Vintage Computing area. There’s always interesting relics to be found, but it’s rare that a real treasure jumps off the page. Recently, I found such a treasure. Jason Harper was selling off his author’s copy of an action game he ported to the Apple IIgs known as Airball. This is how the auction item started off:
A long time ago, on a platform far, far away… I wrote an Apple IIgs version of MichTron/Microdeal’s popular Airball adventure game, which had also been ported to the Amiga and PC. It was this publisher’s only IIgs title; they had no idea how to market it, and gave up after selling only 144 copies. All I had to show for a few months’ effort was one pitiful royalty check, and two author’s copies of the game. I gave one of those to a friend long ago, the other one has been sitting on a shelf since then. It’s unopened, in the original shrink-wrap, although the top seam in the wrap has split open (actually, I think it might have been that way when I received it). It’s been years since I had a working IIgs system, and in all that time I’ve never even bothered to set up an emulator to run any of the software I wrote, so there’s obviously no point in holding on to this any longer.
This really jumped out at me for several reasons. First, Jason Harper was somewhat of a legend in the Apple IIgs scene for a variety of applications he wrote in the mid ’80s, most famous of which was SHR Convert (which later evolved into SuperConvert), an image conversion application that had to be the most popular piece of shareware ever written for the machine. Second, I’ve always loved the game Airball. I was an Atari 520ST user just before I got my GS and one of the best games for that machine was Airball from MichTron, a native ST title. Gameplay, graphics, music—it was all there. Several years later we saw a GS version floating around on the BBS’s, but it appeared in no ads, stores, or mailorder catalogs. The assumption was that a full port for the GS had been underway, but the project must have gotten scrapped before it hit market. As it turns out, this wasn’t quite the case.
I bid and won the auction for this game [ see box photos here: front, back ] and Jason was kind enough to answer a few questions I posed with regards to his work on this project:
Me: I can’t believe MichTron (Microdeal) rolled it out and did such a poor job of pushing it. Did they make any formal announcement at all?
JH:Not that I know of. I forgot to mention in the description that they later sold some copies to some IIgs dealer – Big Red Computer Club, maybe? (it’s been too long to remember.) I don’t know if those copies were part of the 144 that I actually got royalties for, by the time I heard of this that company had gone out of business.
Me: Would you mind telling me if it was a C or asm project?
JH:Assembly, both the original and the port (that was probably true of every arcade-style game of the era).
Me: Airball has a really nice audio track. Is the music identical to either the ST or Amiga version?
JH:It’s using the music data directly from the ST version – I have no idea if the Amiga version was the same, I never saw it. However, I ignored some note envelope data, as the music sounded identical to my ears without it (and it would have been a lot of work, and slowed the game down, to interpret that data).
Me: Any other comments you can give on the project, just for general interest?
JH:The only other major detail I remember is that the game ended up working quite differently internally than the ST version, due mostly to the rather different organization of pixels with the graphics memory on the two platforms. I did it their way at first, and the game was far too slow, so I rewrote it using a complete Z-buffer to handle the display of the airball that can be partially obscured by objects in the room. That turned out too slow as well, and the amount of work needed to speed it up would probably have been sufficient to fix the original method, but I was too committed to the Z-buffer approach by the time I realized that. The memory needed for the Z-buffer meant that the game couldn’t run on a 512K machine as the ST version could, but that probably would have been the case anyway since I was running the game under an operating system (the ST version booted directly into the game, and couldn’t be quit other than by a reboot).
I had such high hopes of porting their Time Bandits game as well. However, communication with MichTron was always difficult – they never seemed to grasp the fact that if I asked them a question, there was an implied request that they actually answer that question…
Me: Are you still coding these days?
JH:Yes, although mainly for microcontrollers (embedded systems) these days, along with the circuit design and printed circuit board layout needed to support that work.
It’s great to have had a few words with this GS developer of yore and also to have come into posession of so rare an artifact from that period. This is definitely one of the most prized items in my collection.
Airball exists for the Atari ST, Amiga, Apple IIgs, IBM PC, NES (never officially released), and the Gameboy Advance (unofficial port). The eBay auction item can be seen for the next 90 days or so here. To read the rest of Jason’s auction posting, which describes the game itself, click the link below.
I’ve expressed before that the inspiration for the name Byte Cellar is somewhat derived from the fact that, in my current home, the computer area containing all of my machines is in a (newly finished) basement room. I’ve got much more space in this 15×17′ room than I had in the previous house. Recently the room entered a rare state of neatness which prompted me to generate a 360-degree QuickTime VR “movie” of the room for the interest of other vintage hardware nuts.
I used Apple’s QuickTime VR Studio 1.0 software (circa Mac OS 8.x) and my Apple QuickTake 200 digital camera (640×480 resolution) to snap the shots because of QTVR Studio’s great lense profile for this camera. (I could never get the software to properly patch together images taken with my 3.2MP Canon PowerShot S30.) 30 pictures were taken at 12-degree rotational increments around the room, from a tri-pod mounted close to center.
The neatness has already worn off to a large degree and soon utter chaos will, doubtless, once again take hold. Other shots of the computer area, including a QTVR taken several years ago, can be found here.
It is with sadness that I make this post. Jef Raskin, pioneer of user interface design, Apple employee #31, and the man who started the Macintosh project at Apple in 1979, died this past weekend of pancreatic cancer. He was a visionary and all around swell guy, and he will indeed be missed.
Although the Macintosh is currently Apple’s flagship product and known worldwide, Raskin’s original vision for the machine was very different than what ultimately was released as Macintosh. Raskin Macintosh, which he named after his “favorite eatin’ apple,” was to be a simple, 6809-based machine with an intelligent, intuitive, text-based display which could serve as “a computer for everyone.” When Steve Jobs took the reins of the Macintosh project, it was transformed into a Lisa-like personal computer with more horsepower and a graphical user interface. Raskin’s Canon Cat, released in 1987, was more akin to his original Macintosh vision than what Jobs and the Macintosh team ultimately delivered.
Jef Raskin was diagnosed with pancreatic cancer in December 2004 and died at the age of 61 on Saturday February 26th, 2005 in Pacifica, California. My condolences go out to Jef’s family, friends, and the rest of us who lament his passing.
About two years ago I setup a nice SGI O2 system to finally have a chance to play with IRIX and the 3D video hardware that SGI is so famous for. I’ve had fun with the box, but never had any app or demo that really pushed the hardware. Eventually I became aware of Certain Impact [via Internet Archive], a flight simulation game/demo from Paradigm, known for their work on flight simulators for the US government and also for designing the flight engine in the popular PilotWings and PilotWings 64 games for the Super Nintendo and Nintendo 64, respectively. I searched on and off for a year and a half before finally finding Certain Impact, so you can understand how quickly I fired up my O2 to get it installed once the CDs arrived. That’s when I was greeted with the chilling message.
There are some things that can be seen on a computer screen that truly send a chill down one’s spine. I have a screenshot of a screenfull of horror. Memory error. Thanks to the 256MB of ECC memory in the O2, the error was rather precise in nature. I cracked open the unit and rotated/reseated the DIMMs—twice actually, and sadly the problem remained. It seems that, for the first time in the 23 years I have been using computers, I have experienced a memory failure. For the time being, I should be able to pull two of the eight DIMMs and run with 192MB RAM, hopefully seeing no further memory errors.
A sad moment, indeed. Hopefully this will bring the system back up to speed, allowing me to waste precious hours in front of Certain Impact.
UPDATE: The system has been running fine for days now with the bad DIMM (and its neighbor) pulled, dropping the total memory to 192MB. What’s more, Certain Impact plays just great! I will have to post screenshots at some point.
Updated on 05/22/2020 to fix RAM error screenshot link, removed YouTube video embed (replaced with my own video), and add Internet Archive link for Certain Impact page.
I am pretty much always in the middle of setting up or modifying or reconfiguring one of my old school systems. Invariably when I start overlapping such projects, chaos ensues. I really need to straighten up my workspace. I have lately been careful not to fall down and become either electrocuted or strangled by the cable web.
One weekend last summer, my wife and I drove to a remote shopping center in search of a particular lamp shop. When we pulled into the parking lot, we found a massive “flea market” spread across the asphault with hundreds of people peddling table after table of what was basically junk. I noticed some old game consoles among the junk and thought I had best dig about for a bit in case something interesting or rare was lurking there, waiting for me. While I didn’t find anything good, I did notice a fellow selling old PCs out the back of a large cargo truck. When I got closer I saw that among the stacks of PCs were a few old, large Sun servers and a slab-shaped machine, of which I could only see the rear. The longer I tried to identify the slab-shaped machine, the more I felt I recognized it. After a few moments I realized it was an HP 9000 712 “Gecko,” which I had seen reviewed in an early-90’s NeXTWORLD magazine. (I was impressed with myself at having identified a machine I had never seen in real life, and from its rear interface configuration alone. My wife was simply scared.) I asked the gentleman what he wanted for it and soon I was walking off $20 poorer with a machine that I knew would run NeXTSTEP over twice as fast as my current slab.
The machine I brought home that day was an HP 9000 712/60, also known by HP’s codename, “Gecko.” The Gecko is a workstation based on the PA-RISC processor that normally runs HP/UX, HP’s flavor of UNIX. The Gecko is unique in several ways. First, it is one of the few non-NeXT platforms that can run the NeXTSTEP operating systems (another UNIX). In early ’94 I had purchased a 486 66 that was specially fabricated to run NeXTSTEP for Intel. (It was even black!) Around the same time that NeXT ported its operating system to the Intel architecture, NeXTSTEP was also made to run on certain HP PA-RISC machines as well as several models of Suns. NeXTSTEP saw real performance gains running on these new RISC workstatons above NeXT’s native 68040-based CISC hardware. Beyond performance, however, there is another detail that sets the Gecko apart.
NeXTSTEP had a graphical user interface with a capital ‘G’. Its interface was lush and vibrant. It looked best on true color video hardware with a 32-bit framebuffer which used 24-bit color (16.8 million) + 8-bits transparency info, but back then the required video hardware was extremely expensive. NeXTSTEP also supported more affordable 16-bit framebuffers, using 12-bit color (4096 colors) + 4-bit transparency. In this mode, NeXTSTEP still looked amazing, but dithering had to be employed and was noticeable. The Gecko features a technology known as Color Recovery that employs an extremely unusual 8-bit framebuffer to achieve true color video output, with the help of a DSP, akin to what would be expected with a 32-bit framebuffer. The following is a clip from an HP Journal article by Anthony C. Barkans, publushed in 1995, that describes how Color Recovery works:
The simplest explanation of HP Color Recovery is that it performs the task your eye is asked to do with an ordinary dithered system. In essence, an HP Color Recovery system takes 24-bit true color data generated by an application and dithers it down to eight bits for storage in the frame buffer. Then as the frame buffer data is scanned from the frame buffer to the display, it passes through specialized digital signal processing (DSP) hardware where the work of producing millions of colors is performed. The output of the DSP hardware is sent to the display where millions of colors can be viewed. It is important to recognize that since the data stored in the HP Color Recovery frame buffer is dithered, thousands of applications can work with it. It is also important to recognize that these applications will run at full performance in an interactive windowed environment. In other words, applications do not need to be changed to take advantage of HP Color Recovery.
When I read NeXTWORLD’s brief rundown of Color Recovery, I could not understand how such a system could effectively generate a true color display—I could not picture what the output would look like. It wasn’t until here, 10 years later, when I finished setting installing NeXTSTEP on this Gecko, that I saw it with my own eyes. And I must say…the result is amazing. The best way I can describe it is to say that the display looks like a screenshot of a true color desktop, saved as a JPEG with medium/high compression. The tradeoff of this 8-bit true color is mild artifacting, but despite this, the display is without question preferable to the dithered, 4096-color desktop generated by standard NeXT boxes and other 16-bit framebuffer platforms. A most unique workstation, indeed.
It seems Retro Gamer magazine, which I spoke of a while back, has added a few seconds to my allotted 15 minutes of fame. I’ll explain how. In issue 10 there is a nice, 10-page article about the Apple IIgs [ picsofmyIIgs ] that reveals a wide range of historical facts and trivia. It’s a great read complete with many color photos and it’s most pleasing to see a machine that figured so prominently in my past put back in the spotlight. One area where the article is a bit misleading, however, is in its description of the IIgs as a 16-bit machine akin to the Commodore Amiga and Atari ST.
While the IIgs has graphics capabilities more or less similar to those machines, and audio hardware more powerful than either, strictly speaking the IIgs is an 8-bit machine. The Amiga/ST employ the Motorola 68000 CPU which is a 16/32-bit CPU, in that it uses 32-bit internal registers but sits on a 16-bit data bus through which it communicates with the outside world. At the heart of the IIgs lies a WDC 65C816 processor which is an 8/16-bit CPU, in that it has 16-bit internal registers, but sits on an 8-bit data bus. The only way one could consider the IIgs to be a 16-bit machine is if the internal CPU architecture alone is being referenced, and from that point of view the Amiga/ST become 32-bit machines.
At any rate, feeling the matter needed to be clarified, I sent an e-mail to the editor that described the above…and promptly forgot the matter. Several weeks later I was paging through issue 11 when I came across a letter to the editor which was pointing out this inaccuracy. I began to read with interest, as I recalled that the same inaccuracy had bugged me when I read the original article. (I still did not recall that I had written in about it.) It wasn’t until about half-way through the piece that I was so struck by this reader laying out point after point that I had wanted to make, that I jumped to the end to see who it could be. Shock and surprise—it was me! There was my e-mail, in full, complete with a link to this site.
One weekend last summer, my wife and I drove to a remote shopping center in search of a particular lamp shop. When we pulled into the parking lot, we found a massive “flea market” spread across the asphault with hundreds of people peddling table after table of what was basically junk. I noticed some old game consoles among the junk and thought I had best dig about for a bit in case something interesting or rare was lurking there, waiting for me. While I didn’t find anything good, I did notice a fellow selling old PCs out the back of a large cargo truck. When I got closer I saw that among the stacks of PCs were a few old, large Sun servers and a slab-shaped machine, of which I could only see the rear. The longer I tried to identify the slab-shaped machine, the more I felt I recognized it. After a few moments I realized it was an HP 9000 712 “Gecko,” which I had seen reviewed in an early-90’s NeXTWORLD magazine. (I was impressed with myself at having identified a machine I had never seen in real life, and from its rear interface configuration alone. My wife was simply scared.) I asked the gentleman what he wanted for it and soon I was walking off $20 poorer with a machine that I knew would run 
