Wang 3300 Recollections

This page contains various stories and memories of people who have contacted me about the 3300. Each has given me permission to post these excerpts.

• Ken Sullivan (former Wang district sales manager), and the guy who developed the car dealer market for Wang, which has been documented in the two significant books about the history of Wang.

  I sold the first one [3300] in the Chicago region in Feb '70, to a real estate company in Oak Brook, IL, to Jay Fitts; he was a land use consultant. I don't know if the system ever got delivered. It was so far behind schedule that they may have cancelled the order. I was the Dist Mgr in Kansas by then.

  Then I sold another to Canisius College in Leavenworth, Kansas, beating DG (I was very proud). In this case the 3300 was delivered, but it had problems. Extended BASIC was running but the hardware was unreliable and the customer threw in the towel.

  Neither one worked long enough for the customers to keep them. I bet I loaded the bootstrap program 1,000 times (before the ROM).

  …

  One of the practical problems was the bootstrap program which had to be keyed from the front panel in HEX. It was about 80 lines I recall. Without NV memory it had to be entered every the computer was powered up. One had to get all 80 lines right the first time or restart. This made the 3300 very difficult to sell, even though the enthusiasm from customers was good. The first BASIC compiler fit in 4K I think and was pretty good -- especially compared to the mnemonical calculators. Every time something happened -- and once extended BASIC in 12K was working lots seemed to happen - one had to go back to load the bootstrap. Eventually, they provided a ROM-based bootstrap, but it was too late. I feel this difficult booting, especially when combined with paper tape for loading compiler and programs, was a significant reason why the 3300 never got traction. I notice this is not mentioned in either archive on the 3300.

• John Geremin is the Collections Officer for the Australian Computer Museum Society Inc, but in 1972 he was at the University of New South Wales (NSW) in Sydney when they received delivery of a 3300. There he acted as the system manager.

  I was responsible for reloading the system from time to time, especially after I reconfigured the partitions.

  It was used for Civil Engineering Structural Analysis programs (as well as some simple demo stuff), and from memory I used to set up one large partition (my guess 12K-16K), a couple of medium sized ones (my guess 8K) and a a small one with whatever was left over (probably less than 4k).

  I don't think that we had the biggest configuration - my guess is that it was 48K for a max of 4 users.

  I also suspect that it was touted as a 16bit machine - at least as far as memory addressing goes. We had ASR-33 teletypes and later a dual cassette drive (using audio tapes - modified to remove the blank leader sections).

  The system load tapes were certainly not small - I reckon about half a roll of paper tape - hence maybe 300 to 600 feet.

  It certainly took a significant time to reload. Quick calc:- 300 ft x 12 inches x 10 chars = 36K bytes, but I cannot imagine that all of this stayed in memory (too big). For load time 300 x 12 x 10 / 10 x 60 = 60 minutes (a bit long) - so my guess at the tape length must be a bit high - it was at least half an hour and maybe closer to an hour (I used to go for morning tea and let it run). It was a real nuisance if there was a bad load.

  One of my jobs was to program it for virtual memory work - i.e. doing segmented matrix inversions of problems stored on a cassette that would not fit into the core memory.

  I recall that the system that the University had was purchased through 'Anderson Digital' in Australia and when it was delivered it was still wired up for 110 volts (not 240 volts) and a lot of smoke escaped on the first day. A good lesson for my future!

  I also recall that most of the maintenance sessions consisted of using a soft eraser to clean the contact fingers on the circuit boards.

  I also suspect that there must have been larger memory modules in later times - other than the 4K units initially advertised.

  The minimum configuration that you listed as a 'basic system' would probably have run single-user BASIC without any normal operating system in order to fit in 4K or 8K. The multi-user systems certainly had all the necessary facilities of an O/S in order to handle multiple users.

• Steve Witham said:

  Our school system had one. Housed in the high school where I was going, two online teletypes there, & I think two 110 baud modems, one for each junior high. This was about 1972-1975.

  I remember that you keyed in the bootstrap from the front panel, then loaded the loader from the paper tape (on a teletype), and that loaded Basic from the cassette.

  I wrote a Star Trek game, broken into five parts, so that each part was small enough to fit in one user partition. The five parts were stored on cassette and would load each other automatically. Besides this seeming high-tech in its own right, it was cool because when you engaged the warp engines, the tape drive started accelerating.

  Somehow I got enough information to program the thing in machine language, including i/o to the teletype. At one point I convinced some guy at Wang to send me an assembler listing of the I/O part of Basic, but I found it too complicated (it used interrupts).

  I remember rigging up a chat program between one of the high school teletypes and whoever dialed in on the modem. Some junior high student and I spent an hour talking over the teletype to each other and it was cool. I don't think I ever ran it again.

  I had one machine language program that did a memory dump of the BASIC interpreter, which I cut up and put onto 3x5 cards, one 256-byte page per card. I spent some time trying to disassemble it by hand, yerg. I don't think I got much of anywhere.

  I remember finding strings inside the interpreter and changing them. For instance "3300 BASIC READY" became "3300 READY OR NOT".

  We had a great program to crash the interpreter. You load it into the partition, then use the local mode of the teletype to put the prompt,

  DON'T RUN THIS PROGRAM FOR A REAL GOOD TIME!

  On the paper for the next victim. I think it went like this (& I don't remember whether it worked on the 2200):

      10 GOTO 30
      20 RETURN
      30 GOSUB 20

  1975 is the year of the Altair & the year I left high school. I think I went back later and found that they had upgraded with a cassette-loader ROM.

• Tom Lake:

  I was in high school in Schenectady, NY at the time (1971) and our math teacher, Dr. Burdick, wrote his own manual for the 3300 BASIC which included courseware he developed. I still have the manual he wrote. The computer itself was located in our school with two ASR-33 Teletypes and each of the other two Schenectady public high schools had one Teletype connected to our 3300 via a dedicated phone line. I was responsible for keeping BASIC loaded so I got out of quite a few classes to reload it from the paper tape after a system crash. That's how I know how long it took to reload BASIC. The system ran quite nicely until we tried to add extra RAM and a cassette unit. After that it was down more than up. Dr. Burdick finally got so fed up that he sent the Wang back and rented time on SUNY Albany's Univac 1130.

• Mark Crispin:

  I'd like to offer some of my recollections.

  I was a high school student from the fall of 1970 until the spring of 1973 (I punted my senior year to go to college). Along with a Wang 360 calculator, Middlesex High School (Middlesex, NJ) had a model 33 Teletype that dialed into the Wang 3300 that was located at Rutgers Prep.

  I am pretty sure that I used it during my freshman (1970-71) year. I know that in my sophomore year, the BASIC was extended; I think that it was called something like BASIC PLUS. The extensions added alphanumeric variables (A$, etc.) and matrix operations.

  In my junior year, the Wang was gone, replaced with a system called ALICE ("Alpha Interactive Computing Environment") which had logon (this was also at Rutgers Prep; my high school just had the Teletype). I didn't find anything about it on Google; all I know that it existed in 1972-73. I don't remember much about ALICE; by that time I had discovered CALL-OS and APL\360 on the IBM S/360-67 at Rutgers University and so wasn't particularly interested in minicomputer BASIC any more. It might have been based on a DG Nova. ALICE did have files, which meant that it must have had some sort of primitive disk or perhaps a cassette tape.

  The first version of 3300 BASIC didn't have alphanumeric variables either. You couldn't do:

      10 PRINT "DO YOU WANT TO PLAY";
      20 INPUT A$
      30 IF A$(1) = "Y" THEN 50

  Instead, you had to do something like

      10 PRINT "DO YOU WANT TO PLAY (1 FOR YES, 0 FOR NO)";
      20 INPUT A
      30 IF A = 1 THEN 50

  I was so happy when the compiler was updated and we got alphanumeric variables!

  I forget the full details of the matrix operations, but I remember that matrix printing and matrix arithmetic (including inversion and determinant).

  The instruction

      MAT PRINT A

  for an matrix with dimensions 10 and 15, did more or less the same thing as:

      FOR I = 1 to 10:FOR J = 1 TO 15:PRINT A(I,J);" ";:NEXT J:PRINT:NEXT I

  I don't think that I have ever encountered another basic that had that, although maybe the 2200 did (I never used a 2200).

  "DET" was a pseudo-variable that you could reference after doing an inversion. I forget the syntax for matrix inversion, but it would have been something like:

      10 MAT A=INV(B)
      20 D = DET

  It also had a primitive macro capability with DEF. The syntax was something like:

      DEF FNA(X) = X*X

  and then you could use FNA(10) to get the square of 10, etc. I forget the precise details.

  I forget what RERUN and RESTART did. START was the command to erase the program. I'm thinking that maybe RERUN may have run the program without erasing variables first.

  Another quirk of that basic is that the END statement did not have to be at the end of the program. It was equivalent to a STOP statement, only it would output "END PROGRAM" first.

  The SAVE command just started the paper tape punch on the Teletype and punched a header and a trailer along with the listing. Otherwise, it was the same as LIST command. Similarly, LOAD started the paper tape reader.

  I recall that when doing a LOAD, the teletype would pause briefly at each line, so the 3300 must have been capable of sending XOFF for flow control.

  I remember that if the machine crashed, it would be at least an hour before it came back. So that recollection matches Dr. Wang's.

• Slightly edited comments from Rob:

  Hi Jim, I came across your Wang 3300 site and wallowed in memory - my high school was fortunate to have one, which I used as a student in the mid-1970s.

  I hadn't known that there was a Selectric terminal option - darn! Our configuration had 4 ungainly Teletypes, and their sound is embedded in my brain.

  There definitely was a Fortran compiler; I think it was written by some third party, likely a university. We had the manual and tape for it, but to my recollection we loaded and experimented with it only briefly because we needed the BASIC system for classes.

  I have to disagree with Tom Lake's view of the cassette tape storage - perhaps he had a bad unit or a design change improved it, but ours was a reliable godsend. The interminable delay of loading the system software paper tape roll whenever there was a power outage or static electricity glitch (the 3300 was in firm "DO NOT TOUCH" mode all winter) was relieved when we eventually bought that cassette unit.

  About the BASIC interpreter's terminal time slicing: I realized at some point that the unit of time slicing was one statement, and wrote a trivial looping program with a very slow statement composed of many embedded trig functions, something like FOO=SIN(SIN(SIN(SIN(… - you get the idea. Instantly everyone's programs slowed down a couple orders of magnitude, plainly evident from the rhythmically pausing teletype sounds. I did this only as an amusing demonstration, not as a vandal.

  I had an incredibly favorable impression of Wang as a kid because after I wrote them asking if they had any extra documentation, they sent a large box containing what must have been one of every relevant publication, including a huge printed library of programs. Alas, I recycled all that during a move many years ago.

If you have any links about the 3300 family, please let me know so I can add them to the list.