- Thanks for creating this fantastic resource. Much appreciated!!!
- Wow! This being released is wonderful and unexpected. I first heard about these tapes being found six weeks ago yet the NSA being unable to release them due to not having a suitable working 1-inch VTR machine (via this article: https://www.muckrock.com/news/archives/2024/jul/10/grace-hop...)
That article was re-posted here on Two Stop Bits (Link: https://twostopbits.com/item?id=3788) and elsewhere but didn't seem to get much attention and I feared the worst, since 1-inch magnetic video tape degrades with time. Very frustrating since such vintage VTRs do exist in working order in the hands of museums, video preservationists and collectors. Now six weeks later we get the best possible news! Hopefully, that article and the re-postings helped spread the word and someone in control of access to the tape got connected to someone with the gear.
And what an amazing piece of history to have preserved. I'm only ten minutes into the first tape but she's obviously a treasure - clear thinking, great communication and a sharp wit. Even captured here later in life you can clearly see why she was so successful and highly regarded by her peers (including some the most notable people in early computing history).
- bmonkey325 98 days agoThe living history of early computing is being lost more and more. Having this available is amazing.
- I really appreciate it when tech companies make an effort to preserve history and share it.
However, it's important to also share Nintendo has recently launched a major campaign of unprecedented legal attacks on the historical preservation, retro emulation and fan creation communities. While Nintendo has always been protective of its IP rights, this was largely limited to stopping piracy of current titles and protecting their trademarks from commercial infringement, both of which are appropriate and understandable.
However, the greatly expanded and all-encompassing scope of their recent legal actions now threatens aspects of non-profit historical preservation and adjacent fan activities unrelated to Nintendo's present day commercial interests. Previously, Nintendo's lawyers made at least some effort to distinguish between non-profit or fan hobby activities and piracy or commercial trademark abuse. Their senior management also seemed to appreciate the retro community is some their most active current customers and brand fans. Over the last year that restraint has vanished, making it all the more ironic Nintendo is now funding a new museum to celebrate its past while funding lawyers to threaten the non-profit historical preservation and non-commercial fan communities passionately keeping Nintendo's storied past alive for future generations.
While I'll always love Nintendo's rich historical legacy, until they again exercise reasonable restraint in their now-abusive legal tactics, I can no longer in good conscience support the company or its products.
- Very much looking forward to this series. I'm especially interested in where IGS sits technically in relation to other early approaches to visual representation languages like NAPLPS, ReGIS, VT36, and various ANSI expansions, etc. Back when data transmission was glacially slow and memory astronomically expensive, transmitting visual information as the instructions to re-draw it locally was so clearly the right idea, I think almost every major 8-bit and 16-bit platform had disparate hobbyist efforts to implement some form of this. Of course, due to platform fragmentation virtually none gained enough adoption to leave any lasting legacy.
While largely forgotten today, these were the primordial ancestors of the more advanced implementations that would go on to change the world, from Postscript, X11 and Flash to OpenGL and, of course, HTML and CSS.
- The 2nd part of the series will take a brief look at IGS' command structure. If you made a spectrum of protocols based on their commands' "terseness", I imagine NAPLPS would be among the most compact on the left side, RIPscrip might be in the middle, and IGS would be at the right among the least compact.
Code for some of its commands is fairly readable, but others (particularly the `&` Loop command) can be nearly inscrutable.
- Somehow I don't think I ever saw RIP (or IGS) in action on a BBS, I might have always said no not understanding what it was (or whatever clients I used didn't support it).
Though presumably v.44 protocols would compress ANSI graphics text reasonably well?
- This was a thing with the Amiga 500 computer too. I worked at a company which had a corporate field deployment of over a thousand Amiga 500 systems and some of these computers would start crashing intermittently due to the A500's "Fat Agnus" becoming unseated in its square socket, especially during shipping and thermal expansion/contraction. The chip got its nickname because it was a square packaged upgrade version of the earlier rectangular Agnus chip in the Amiga 1000.
These 1,000+ Amiga 500s were leased as complete systems including a 13" monochrome composite video monitor and an Okidata 182 dot matrix printer. All these Amigas only ever ran a single, highly complex and extremely valuable custom software application which auto-booted from the internal 880k floppy, was never publicly released and has remained unknown to this day.
This program was protected from being pirated with hardware in the form of a custom-designed and manufactured 1MB RAM + Real-Time Clock expansion board pre-installed in the A500's "trap door" slot on the bottom. This unique board added the same functionality as the standard Commodore A501 board but was customized by swapping the real-time clock data lines around, which the software checked to confirm it was running on an officially leased A500. The software application was written in C by myself and two other devs and we also doubled as the support, service, training, installation and shipping staff for this fleet.
Since we shipped them to each field location and they tended to be in locations which weren't well thermally controlled after hours, we got a fair number of these errors. Since the vast majority of our users were older female office workers who'd never touched a computer before, any error would trigger a support call. It wasn't easy to convince these nice ladies on the phone to lift the computer 6 inches and then drop it! But it did always fix the problem. We called it a "technical drop" :-)
As a very early Amiga 1000 owner and enthusiast, I'm the one who primarily convinced this small company that the newly announced A500 was the best low-cost option to run field deployments of their niche application and, aside from this occasional error problem, the huge fleet of Amigas performed admirably for many years. Especially considering the machines were individually shipped to each field location and these tended to be pretty hostile environments like temporary mobile offices, sometimes powered only by gasoline-powered generators subject to constant surges and brown-outs (think construction trailers).
I'm proud my unorthodox recommendation ended up working out so well since, despite my obvious bias, the A500 really was the perfect fit for this unique application in early 1987. The software app itself was related to real estate financing, so it had to have a 10-key and the entire system had to fit on a small side desk used by constantly rotating staff in a crowded temporary office. It had to be self-booting, self-maintaining and dirt simple, since almost none of the users had ever touched any computer before. It also had to be reliable because they frequently got moved around in the field between desks and buildings entirely by these novice users. So being all-in-one really helped since every extra wire and component was another thing to lose and/or break. I learned the hard way that talking to a nice lady on the phone who'd never seen a computer in person through reconnecting this "newfangled contraption" which someone else had disconnected, moved and left on the floor in a random pile the day before, was a rapid education on how to communicate physical instructions clearly! I always just pictured them as my sweet elderly aunt which helped keep my cool :-). Finally, the system had to be super cheap because they got stolen fairly often (no security after hours in these isolated remote offices).
Although entirely unknown at the time or after (since there was no reason for this small company serving a narrow niche market to do any PR) this was, as far as I know, the largest ever corporate deployment of Amiga computers.
One interesting note: the application was written to only use the keyboard, mostly just the arrow and Enter keys plus a few function keys we labeled with custom color-coded stickers for navigation such as Go, Back, Menu and Print. Usually we never even supplied the Amiga mouse with the systems, however I was optimistic we could show the users how to operate the mouse so they could run other apps if they wanted to. I conceded defeat after trying to teach the first half dozen users I did installs for how to use the "Bonus Mouse". I thought some would eventually get it but the necessary eye-hand coordination completely eluded these nice ladies circa 1987.
- That’s a great story, thanks for sharing. Where in the world did you deploy 1000 Amiga 500s?
- > Where in the world did you deploy 1000 Amiga 500s?
Good question! It was a small niche business called "Computer For Tracts" which leased a complete turnkey solution to new home construction sales offices. You'd find these in new growth areas where home builders have bought a large land parcel to create a new development by building a couple hundred suburban homes (as well as the local streets, parks, etc). Basically, they'd first clear the land and complete four or five homes to show as models of each floor plan. They'd usually turn the garage of one of the models into an onsite sales office which would be staffed by an employee realtor who could discuss all the various options (flooring, paint), lot selection and write up a sales contract. These little makeshift sales offices would operate for 3 or 4 years until the development was sold out, at which time the models would be converted into the last homes sold and the builder would start a new development elsewhere.
The company was started in the early 80s in Southern California by a realtor who'd spent his career in new construction residential sales and knew the biz inside out. New construction residential sales has distinct needs different from residential resales. So he wrote a software program in the ROM BASIC of a Radio Shack Color Computer to do exactly what a new construction sales person needs to do. This involves calculating the financing options including the down payment, monthly payments, taxes and insurance for all the various loan options available, like 30-yr fixed, 15-yr adjustable rate and then sending prospective buyers home with a customized printout showing their options based on their particulars including down payment, rate, credit score, etc.
It was really quite complex under the hood, yet it made all this easy for a sales agent to present clearly to even novice home buyers. It turns out that doing this is well is key to selling the builder's growing inventory of homes and, obviously, there's a lot of money tied up in such a development. His early versions of the program loaded from an external audio cassette tape player by having the user manually press Play, Stop and Rewind. But this was a major problem because new home construction companies aren't IT savvy and the vast majority of realtors (especially in the early 1980s) had never even seen a computer in person before. So, he learned how to burn his custom BASIC program on EPROMs he installed into blank game cartridges inserted into the Color Computer's game cartridge port.
He was then able to offer a self-booting system that his novice users could learn to operate. He marketed it as a complete solution on a monthly lease (including computer, software cartridge, monitor, printer, cables, installation, training, service and support). Since these Radio Shack computers where just a few hundred dollars and his system was extremely valuable to such construction companies, his little garage business grew like wildfire. I think I was the third non-family employee and we were still working out of the garage and back bedrooms of his house before later moving to an office. Eventually, the software became so complex that it wouldn't fit in a ROM cartridge any more, even after I wrote a pretty neat assembly language hack that lived in the game cartridge ROM and took control on every boot to add new commands to the computer's built-in BASIC interpreter before returning control back to the BASIC program. These new BASIC commands allowed programs to swap between four and eventually eight banks each containing an 8K segment of BASIC code and even pass control to a specific BASIC line number in another bank of EPROM-based BASIC code. It was quite a deep hack into the undocumented system but it worked perfectly and is a testament to the benefits of having an all-hardware system where nothing ever persists between boots. When the capabilities of the program eventually expanded so it couldn't fit in 64K or even 128K of dedicated EPROM (in addition to the computer's 64K of RAM), even with lots of very clever byte-counting optimization and compression tricks, we started looking to move off the 8-bit Radio Shack Color Computer to a 16-bit platform with more memory and disk storage - and I started suggesting the just announced (but not yet shipping) Amiga 500 would be ideal (which it was)!
Over the early years of constant iteration and daily customer feedback, the program had evolved to become really perfect for doing this one, extremely valuable yet highly specialized thing and the little company grew fast and was highly profitable. He could charge quite a bit for the monthly lease because no salesperson who'd used it ever wanted to sell without it. It was really that good. In fact, the lease required the first two and last two months payments up front, and this was enough to basically pay for the computer, monitor, printer, power strip, etc. So after that, the rest of the lease contract was all profit (although, service, support and ongoing training had to be available 10 hours a day, seven days a week, including weekend onsite system replacement if necessary (which could involve driving over four or five hours round-trip on a Sunday if you were on-call that weekend)) so the customers definitely got value for their money. And supporting this highly specialized, extremely deep application took at least a full year for a new person to learn (if they were sharp) which is why in the early days we three software coders were also the support, training and field service staff). While strange in today's world, that's probably why this software became so damn polished to perfection for this use case.
For example, after selecting a home model, to update the main financing screen showing all the different loan options, all the agent had to do was punch a different interest rate, monthly payment, or down payment on the 10-key and it would recalculate all the loans with the new parameter. They didn't even have to select an input field because the program dynamically figured out based on the value typed and the sales price if the entered value should be interpreted as a rate, monthly payment or down payment.
In an era before spreadsheets, without this program, it took these agents at least three minutes to change one parameter for one loan on a manual TI paper-tape adding machine. But our program recalced a dozen loans in two seconds (including every possible option from full amortization tables to rate buy-downs). This made providing full, clear and detailed answers to every customer question trivial (like "How much more do I need to put down to reduce the monthly payment by $500?"). To these sales people this seemed like literal voodoo magic ("How does it know that what I typed in was the desired monthly payment?") :-)
Last I heard the company had transitioned the program to Windows PCs sometime in the mid-90s and was still around in some form post-2000.
- BASIC on the early 8, and more generally 16 bit machines was quite a bit more empowering than pop media tends to speak to.
My uncle Bob (seriously, I have the generic uncle "Bob"), developed real estate contracts using a combination of C64 BASIC and some word processor that allowed for conditional and parametric document assembly, almost Word Perfect style!
He built up quite a business with those efforts!
A bit later a friend wrote an entire trucking business on the PC running GWBASIC.
I myself started out on a beat up Atari 400 with the Atari BASIC cartridge and the cassette storage peripheral I struggle to recall the name of right now.... 410! That was it.
I wrote TV test and alignment programs. Learned all that working at a TV repair shop as a kid. The Atari had just a couple capabilities that made a huge difference too!
One of those was at least 8 grey shades. I know GTIA could deliver 16 and I ended up using them once I made enough to get a newer 800 XL machine.
Another feature was full overscan graphics. 48 bytes per line instead of 40. That made it possible to draw the full frame patterns and properly identify the safe area for viewers wanting the factory setup, and expand viewing for others without showing blank non raster regions on their screen.
Side bar:
Older sets would often under scan by quite a bit! Correcting that often meant a lot to those viewers.
End Side bar
Another feature was enough colors to calibrate a TV for good color more than close enough. I could get purity tests, set color delay phase and some other items pretty well!
Last feature was 320 pixels in the safe area NTSC. That is two pixels per color clock cycle. When set to monochrome, those pixels were just right for focus, convergence, linearity and the whole test pattern.
All this was some percent off the pro gear, but I found out most people do not care. And I mostly didn't either.
As a famous YouTube I love says, "Good enough for the girls I go out with" (AvE)
BASIC with a few PEEK and POKE commands and the occasional bit of machine language was enough to do a lot!
COMPUTE! Published a nice assembler and disassembler too. For some work, a guy could get setup well enough to produce good programs.
Getting back to XP...
I wrote the above for perspective. Of course XP can make sense. So can DOS, an Amiga, and Windows 3.11, just ask Southwest airlines.
Fact is many of us here can probably work magic with whatever gets put into our hands. I can.
And all these skills couple with microcontrollers too.
Perhaps that warrants discussion here too one day. The skills are a great match and when one can build hardware feature matched to the use case?
Boom goes the Dynamite!
- I regret giving away my OG XBox, controllers and half-dozen games about ten years ago. Although, it was to a family with several kids who weren't in a position to purchase any console, so hopefully it was loved in its second act.
- As a software guy vs EE, I was confused about how this works and skimming the paper didn't yield the simpler, high-level TLDR I needed to enlighten myself. I searched the paper's URL on Hacker News and found this paragraph in a comment that helped me.
> Programmmers usually think ROMs as data storage devices, but they are also the most rudimentary form of programmable logic, as they transform x-bit of address inputs into arbitrary y-bit data outputs, so they can implement arbitrary combinational logic. In fact, lookup tables are the heart of modern FPGAs. As a result, you may argue that this means any ROM-based design has ad-hoc FPGAs (especially when EEPROMs are so large after the 1980s, 64 K for 16-bit chips). But the use of Mask ROMs and PLAs in Control Units has always been a legitimate and standard way to design CPUs even back in the 70s, so I won't call it "cheating" (and using ROMs for ALUs or Control Unit wouldn't really be much different from using a pre-made 74181 or AMD Am2900 anyway).
- Edit Six Weeks Later: The tapes are preserved and released online! https://www.nsa.gov/helpful-links/nsa-foia/declassification-.../
---------------------
In addition to pushing forward on the FOIA approach, I'd suggest trying to interest whatever group the NSA has assigned with historical preservation. I'm sure they are understaffed and underfunded but at least they have an interest in preserving things like this, whereas the FOIA people's goal is to find grounds to deny requests as cheaply and quickly as possible.
I imagine there are NSA history buffs outside the agency and contacting them first to get their input on who and how to approach (or maybe even an introduction) would probably be better than a cold approach. Two things that may help:
* Make it easy for them by doing the legwork up front. Find a preservationist or museum with a working 1-inch VTR, ideally one of the "portable" ones (actually like a large suitcase vs a washing machine). It may be that removing the tape from the NSA facility makes it a much bigger ask (keeping in mind they don't know for sure what's on it yet). The closer you can make the initial proposition sound to "it just takes one of your people to get the tape for a couple hours to digitize it. We'll set it up but NOT be in the room for the playback, then NSA can decide what to do with it" the more likely success will be. (obviously, it'll probably be more complicated but once they're in...)
* Highlight that this is ONLY about preservation not release. Old magnetic tapes degrade with time. Accessing playback gear may not be possible in another decade. It's probably still savable this year but will soon be lost forever (create urgency with FOMO).
Don't even talk about release or FOIA, make it about some NSA history initiative that already has support. Ideally, they're working on a "Women in NSA" or "NSA and Early Computing" exhibit for next year. Just between us though, once you know that the video exists digitally, the grounds for FOIA denial is gone :-)
Good luck!
- These are excellent points. We know there's nothing spicy in her talk, but as soon as you mention FOIA all the sirens go off and the walls go up and everyone shuts down and gets grumpy.
And if you can find some NSA folks they would be able to make calls inside the system to get the wheels greased much faster than some FOIA intern will.
- It's been shipping for 3 years with over 1,000 sold so far. The current pre-orders are for the third production run.
- Rodrigo Copetti has just released another of his amazing hardware architecture deep dives, this time into SNK's legendary Neo Geo. This covers both the popular arcade machine platform on which 150 different titles were released 1991 to 2004 and the (very) high-end home game console which was based on the same design. Well worth a read and also available as an e-book download.
- More