Daynote ,Svenson

MM-xii     Wednesday

2000-01-12

The mist has all been frozen out of the air tonight, not that -4°C is real cold. Definitively not when there is no...
No mistake, the weather is an exact copy even though there is less frost with all the mist being used up yesterday and all that.

I didn't start on the last HCC program but I solved a problem with prompting program. The unfortunate fact is that I only make about 3 or 4 brand new programs per year . The rest of the time I am slaving at old bugs programs, adding features, fixing bugs, amending inaccuracies and such, of coughing up quick and dirty, behind the back recovery programs or hole covering patches . That is why I don't know the company standard for adding shining new stuff. The result was that the prompt program worked absolutely as I wanted but none of the text and labels on screen were usable for normal users. Correcting that brought some string handling code to my attention.
I had to add a code to the end of a parameter string. In RPG3 you do that with the CAT (concatenate) opcode. (field1 CAT field2:0 result). That of course still works in RPG4 but I used the alternative method off the EVAL statement (EVAL result = field1 + field2 ). The only problem is that I forgot the stripping of blanks. RPG3 does automatically replace all trailing blanks from the first field with a single blank (adding the :0 removes even that last blank). Of course the RPG4 coding, using EVAL doesn't do that. The simplest solution would be to use the CAT statement but some days my motto is "Why do something simple if you can do it otherwise."
So I thought up a rather convoluted EVAL statement with some new built-in functions. Rather nonstandard but it works. <G>

The discussion on the problem I mentioned yesterday (about the ease with which business rules can be chnaged) was resumed today. The greatest problem with changing business rules is that it has the potential to corrupt your database. Actually such a corruption was what cause the Americans to call for a change. Well I lost out in this argument. My proposal was to add a validation and correction procedure to the rules changing program. Easy to do and, depending on the rules that are changed not really bothersome for the local manager. Peter joined in and canceled my proposal with the simple statement : "database consistency is the responsibility of the local manager, it is not our business". So what will happen now is that any record that is fetched for maintenance will be (invisibly) changed and passed trough the database write trigger even if the user doesn't change a bit. If a record is not fetched for maintenance no tests will be done and the corruption remains. In my opinion that is ......

.....recipe for problems that will haunt us later.

I took yet another route home to, successfully, avoid the border pile up. That is three days of "thoroughly checking the border against illegal immigrants. And two days in a row that I cross over without even seeing someone from customs. Now who are they fooling?
This is fun. I'll see if I can keep up taking a different route each day for three weeks.

PS I try to change the tagline and the next-day content daily but I don't succeed each day. It's called time constarints. And the comments were not intended to be displayed. That is called messing up code :)

Adios