System iNetwork

hi roger,
i read the initial blog with much interest, i look forward to continuing
blog issues about the history of the S3x, S4x, and now into the eternally changing i series arena.

having had 37 years in the i/t recruiting industry, one good thing i was fortunate to have was the vision
of this new era of computers in 1971.
i went so far as to throw away every COBOL mainframe resume i ever had, and started recruiting every rpg developer i could find in the country, cause i just knew s3x was the new wave for ibm.
it has been rewarding , challenging, and very frustrating during all those years, and all those changes.

bravo for your new upcoming series

best regards, frank

If you're going to bring S/34 and S/36 into the picture, then our OCL roots go back to the System/3, with it's 2-character Halt code window, a 1403 printer borrowed from the 1401 system and driven in real time by the central CPU, and midnight calls to the IBM SE who would actually open up the cabinet and resolder wires connected to the 128 K of main memory (more or less)...oh, and let us not forget 96-column cards and the MFCM, complete with the horror stories of when I dropped a deck of cards that was a compiled program and how it took me hours to reproduce the object deck so processing could resume.

Has It Really Been That Long? How about 38 years? IT was called DP when I started in 1969. I was hired to operate an NCR 500 while still attending Junior College at the tender age of 20. No, I did not ride a Brontasaurus to work. Actually it was a 65 Corvair and I still curse Ralph Nader for killing a car that was a decade ahead of it's time.

I actually wrote some programs on that NCR 500 in machine code. I think it had 12 registers - maybe 24. The company was growing and by the time I turned 22 we had gone to a WUCU Service Bureau. Now how many out there can decode that? They used a 360/20, the first RPG machine. I was taking 18 hours at BJC that semester including RPG so I helped the WUCU employee (there was only one) develop our ILA Union payroll system to transfer off of the NCR 500. Then after a year with them my boss learned he could lease a System/3 for less than we were paying WUCU so he named me the company's first ever (and only ever) Data Processing Manager at age 23. Who remembers having to have a personal relationship with your CE because he would be there at least 3 times a month to rebuild the card paths in the MFCU? We actually rented our S/3 to the owner of the WUCU franchise at night and they got rid of their 360/20 and their office and we ended up making a profit by bringing in our S/3 to replace them. That would be 1972.

The inevitable transition to Sys/34 came late for us in 1978 and I put in over 120 hours a week for six straight weeks following the cutover. I had stood on my soap box and refused to use any Migration Aids instead re-writing every app to native S/34. I had tested using my IBM test hours but I was the primary coder as well as DP Mgr (btw who remembers DPMA? I was the youngest member in Broward Co FL for several years) and the only one who knew every piece of it. I had two programmers and I alternated letting one go home every other night but I stayed thru it all. Was burned out for two years.

Transition to S/36 was a snap about 1981 then we began re-writing for our S/38 conversion in 1984. But the company started to slide into bad times, cancelled the order after I had already used most of the test time and finally eliminated my position in 1986.

Easy to find another job with the midrange background then so marked time for 15 months at a blah job until I found the one I am in now - and that was 1988. So I am starting my 20th year here after 17 in the first job and the 15 month patch in between!

"Here" is Norwegian Cruise Line and I was in the very first wave of putting integrated systems on cruise ships with two new ships we built in 1988. We used 5362 models of S/36 as footprint was everything on the ships of that time. Went to AS/400 in 1992 after retro-fitting all the older ships the company had in those first three years. I used Type 1 cabling for both the twinax and the 4680 POS. We still have three ships running 9404 models today if you can believe that! The latest ship has a 9406-520 and we only use it for one inventory app these days and they are looking at several replacement options. Most of the apps have moved off to Windows Client/Server platforms (Gak!) and I have morphed into pretty much an Administrative and Operations Mgr for all the Ships Systems of NCL. I have spent 1869 days at sea to date.

I keep hoping that we'll get a chance to return to the stability and complete control of those bygone days and I still have eight more years planned to work. The hell of starting so young is having to work so loonnng!

It's been a long and mostly good ride. I have yet to take a single education course in anything relating to the PC or Networking world although we have three racks of servers and a complex satellite to Cisco network on each ship. I depend on the Network and Server support groups of the larger IT dept for any technical stuff my officers onboard can't handle. I have one Systems Mgr and one or two Asst Sysmgrs on each ship. One lady had a background as an IBM midrange SE and has been on the ships with me 16 years now. She may want to blog here as well.

Thanks for reading if anyone did.

Q38 - the best technical resource we had in our shop in '86.

The early System/38 was so specialized a machine that the only ways to communicate with it besides using 5250-type workstations over twinaxial cable was to use either a Binary Synchronous Communications (BSC) adapter or a Synchronous Data Link Control (SDLC) adapter. But it was such an attractive and useful business platform that there immediately rose a demand for ways to network the System/38 to machines other than an IBM mainframe or another System/38. In response, at least three hardware vendors (Wall Data, Perle and Decision Data) came up with protocol converters that were supposed to convert an EBCDIC 5250 workstation data stream (SNA LU7) into an asynchronous ASCII data stream that would conform, more or less, with display rules compatible with the VT50 or VT100 displays. The challenge was that ascynchronous ASCII communications with other computers, such as DEC, were not always compatible with the protocol converter's idea of how to convert to and from a 5250 workstation session. So, besides having to learn how to code RPG so that it would bypass the 5250 interface restrictions within the System/38, I had to insert a data line monitor into the data communications connection and learn to read the hex code that was frames of SNA SDLC data containing LU7 screen paint instructions. I got so familiar with the hex display that SDLC data frames looked to me like train cars going by on the screen, pretty much like the movie "The Matrix" where technicians could see the green codes dropping down in columns on the display and spot The Woman in Red!

My first day on the job, I was shown a long rack of manuals for the company's System 36 and I was told to start reading. What I signed up for was a part-time data entry position and what I have is a twenty year and counting career. But the job has never been routine. Thanks for the reminder of where we have been and the heads up on where we are heading.

Those 96 column cards were the greatest for jotting a reminder or shopping list on and tucking them into your shirt pocket. And we bought them in myriad colours to help identify which OCL set to grab from the wall rack that had like 36 compartments that could hold about ten cards each.
// LOAD PROGRAM
// FILE NAME-BLAH,LABEL-BLAHBLAH, UNIT-R1
// RUN
data cards here
/*

It comes back so easily and makes more sense to me still than all the time we spend trying to figure out what happened to our Active Directory replications via our satellite connections to the ships all over the world.

Thanks for the great comments to kick us off! I especially relate to Richard's comment about the good ol' days of 96 column cards making more sense than much of the stuff we're using today!

rp

Enough of the last 25 years!!!! Try going into the wayback machine like me, learned to program on a IBM 1401, 4KB machine and worked on an old 406 accounting machine! Graduated to 360/370, then down to system 3, and on up through the systems to the Iseries. WOW what a trip!

Still Kicking

Well it has been 30 years for me since I started on a SYS 3 mod 12. The cards were wonderful until the sorting with the @#$%MFCU anyone remember "F0". From there I went to a SYS/32 and boy I couldn't figure out how to type in my program on to one of those 8 inch floppies then I was shown SEU, how wonderful from cards.
From there we installed a Sys 34 and was on it for the next 5 years.
One company I worked at went from a UNIVAC 9030 to a Sys 34 and a year later I rolled the 5th system 36 to be installed in Canada into the computer room. The installation went great till MAPICS. The version had not been released to convert to the SYS 36 when I wanted to install it. Had to go to IBM Toronto to get the program.

The next company had 10 System 34's and was looking into a new system. They decided on the Sys 38... I was in culture shock. The system was Pansophic RMS 3.0! The plan was to put SYS/36's in the plants as distributed machines and use the Sys/38 as the main processer. Then the AS/400 was announced. We ordered ten of them 1 B20 and 9 B10's.
I have been on the 400 since and I hope to retire still working on it.

An interesting thing happened a few years ago when I took my son to a can recycler in San Diego. I saw, to my dismay, every computer I worked on in a row starting with the System 360 I learned on and progressing from the Sys 32, 36, 38 luckily no AS/400 at that time. All were gutted and ready for metal recycling..

My career in a nut shell.....
Thanks
Martin Collins

WUCU = Western Union Computer Utilities. Franchisees were called Computer Utilities of, for example Boston.

Programming was centrally based in Florida, and the idea was that the franchisees would do the analysis for a new customer's application and send the specs to FL for coding.

Once they'd programmed one A/R system from scratch, all others would be similar, the theory went, and that was a pretty good idea. The existing code could be pulled out of the drawer, a few tweaks made, and presto!, another customer was about to become billable.

The customers never had to pay for the analysis or programming or installation on the Model 20 or the S/3. Those activities would be paid for over time by the processing fees, a percentage of which went to the Florida programming site for their support.

As I recall, Western Union had invested in the idea when it was approached by the people who came up with the idea. Each franchisee went out and got his own local customers, and Florida was to find national customers for each local service bureau to support. I can only recall one or two instances of that, one of which was Western Union itself. The franchise I worked at did TELEX and TWX billing for example.

So that's the WUCU stoy as I recall it, in response to George's challenge to decode the phrase WUCU!

Jim C