They started up the Vaha. It was just a nice coincidence that I'd get my first 32-bit machine on my 32nd birthday. I'm still surprised they did this on a sunday, but then with all the oours working pretty much live on the system, installing the new box on weekend seemed necessary.
It being my birthday, some treat was due. No booze but rather just so, snacks, soft drinks and whatever I got for them hot meal chits. Then it turned out that the two girls from stambena were there, and not being sure whether I had enough for them too, we decided to call them a bit later if there's anything left. The trouble was Joži, who sat with them in that other room, across the hall from Radoja's office. One of operator girls solved the problem by pretending to go tell him to have, like, a call on the phone waiting. She just peeked into that room, and said „Joži, you have wire“. He came over a minute later, saw the spread on the table, I said it my birthday my treat, exchanged congrats, he sat, took some of the layout, and then a couple of minutes later he smarted and dived for the phone, lifted the headset and said „hello?“.
In the next two days I restored all the source code and tables for all of our members (which would be the hardware stores of Čelik, the lumberyards, the kiosks and bookstores of Presprom, and only salaries for the textile and supermarkets of NuProm). I rewrote all the batch files that served as menus (on pdp and Vaha these files could be stuffed with some GUI, using lots of VT-100 escape codes, and particularly on the Vaha they could call lots of system API functions and use the results in various macro expansions... a true language by itself).
Then there was a bug found only when the code actually started running. The cobol compiler on the pdp was virtually identical to that on the VMS, but we found one significant difference: to exit the wait state of the form (and thus exit the program), one needed to press ctrl+Z, i.e. to simulate EOF. This caused an error in the Vaha version, which needed to be trapped... in about 50 cases. Every form was a program to itself. It took me some time to fix in all those places, and we were under serious critique for a while - the Presprom's (tobbaconists) data (one of the larger datasets in stour's system, because they had all those kiosks with lots of small stuff) weren't accurate to the day! They were almost three days behind!
Few days later, I went to the bank with Vera. I wanted to lift some cash, and she had her mom's savings booklet to update (with the transactions that happened in her absence). The bank had the only other Vaha in town; their current chief programmer graduated same as I and probably didn't spend five years teaching and one unemployed, what with having a higher (up in the mountains) pedigree. And they were playing with their Vaha for a whole year now, if not more. The result: I got my money alright, but the savings booklet couldn't be updated until the paperwork arrives. Because in their system the savings accounts were up to date to some date in april.
So... I migrate three oours and all their data and code in three days, with a total of one bug, and they almost vote to shave my beard. The guys in the bank scratch their balls for months, their production tables are five months behind, and they get off scat free, and are even praised.
Location, location, location.
This will keep happening to me, to be criticized for minor faults as if they were something big, while others did ten times worse and got away with it. Specially the authors of software that we were using.
This was probably the time when I finished the basement walls, and poured a collar of concrete on top of it, with the slight grade towards the 60:40 point of the edge away from the house, which was about 8cm lower than the ends and 4cm lower than the house side edge, to have that perfect 1% slope for the rain. Then I arranged, through stambena to have a small crane come and lay the ceiling to the basement.
Then we poured the slab - dad was working the mixer, I was gunning the wheelbarrow, she was flattening the poured concrete, and mom was sitting at the little camp table waiting for us to finish. At some point she concluded that we won't be finishing it anytime soon and that the lunch will get cold, so she called for a break. We were only about 40% done, almost to that low point, so yeah, why not a break. This was the first closed volume of space that we have built.
Several years later there was a crack at the place where we took the break. The concrete poured afterwards didn't quite set with the first batch, and some water entered the tiny cracks. Eventually there was a drip, never too big (the slope!) but visible. We'll try several ways to plug this, but nothing would stay for long enough - the temperature differences, from -26 in worst nights of february, to +45 (or more on the surface) in august, anything cracks. This will be resolved only in 2007, but the solution created a different problem, still unresolved.
20-IX-2012 - 20-XI-2025