MIM-cccxxxxviii

 

Svenson

 

-- Yesterday -- Previous week -- This week -- Next week -- Tomorrow --

1999-12-14     Tuesday

 

Some days ago I lamented the lack of change in the weather. Oh dear. It was about 2°C when I got up at 5h45 with just the beginning of frost on the car and the remnants of mist, quickly settling down. By the time I got to Venlo the wind had picked up some speed and clouds were racing by. From then until I left, at about 18h00, we had rain in all imaginable forms, from drizzle to torrential downpour, some dry spells, some sunny spells and a bit of sleet. None of these conditions lasted longer than say ten minutes. Great. Especially if you can watch it from inside the office.

There were some times I'd rather have been outside today.

I have a nagging problem in the hierarchical code program. I want to clear the file before filling. In RPG III that was not a problem, you could call any system command. Now in RPG IV the same method just doesn't work. I have a program, in RPG IV, where the method does work but this is another type of command. Not good.

Then I had to explain the OFI interface, from sales to accounting, to Theo. Because Iwan, from Norway, had asked to explain the difference between v7.8 and v7.9. He sent his question both to me and to Theo so I explained it twice before I noticed that Theo wanted to know how to explain it to Iwan. Losing direction? You'renot alone.

And then I had to rescue Germany. They are releasing orders twice. This is only possible if they do that explicitly. Normally you release an order and if you check sometime later you notice the order status is changed and it cannot be released again. If however you release it a second time before you update the screen the action is started. On a normal AS/400 the first release job is finished before the second one is actually started. The German system is however so overloaded and badly managed that both release job become active at the same time. This causes the logistics module to produce strange results (they don't prevent duplicate records).
I found a solution but, because our system works normal without congestion it is quite hard to test. The only way it can go wrong, is if the second job becomes active when the first one is at a specific point. There is a window of opportunity of about 25 instructions where an overlap can cause problems. Try to hit that for testing!

I have been banging away at problems from 7h50 till just after 18h00 today. Without lunch pause so I only checked out Tom's daynote and threw in a few quick replies to mails.

I am cooked for the day.


Adios

 

-- yesterday -- tomorrow --
-- Home -- Links -- Sjon -- Calendar -- Jobs to do --

Mail Me
Svenson © 1999

Every day is too short.