|
| In article <dunuj2lsrlmfl2562ajued527airpmi4db@4ax.com>,
Clark Morris <cfmtech@istar.ca> wrote:
[snip]
>Coding is the easiest part. It is the analysis, research, control and
>testing that is difficult.
It is obvious that Space Aliens are stealing my brainwaves and implanting
them in the heads of others yet again; I was saying almost this Very Same
Thing in a meeting the other day.
A user in the field noticed an error in a report that has been running,
more-or-less unchanged, since some time around Magellan's
circumnavigation. This hot, steaming load gets dumped into my lap; I
ascertain that the last time the programs involved were compiled was for
Y2K mods (anyone remember those?) and dig through the data.
I find out that this subsystem is built around an inherent flaw, a naieve
way of dealing with the complexities of 'how many two-w pay periods are
in a given quarter of a year?'... this can be discussed in another thread,
if anyone is interested. Anyhow, in the meeting mentioned above someone
was complaining about the bad results, I explained what was going on -
*always* a Fun Thing when dealing with folks who announce proudly 'Well,
I'm not Technical, you know' - and concluded with:
'The problem lies, in essence, with how people are to be counted by office
for a pay period. Given that this report does not count 'fractional
people'... when I run across Person A who makes $A for Office A in pay
period 1 and then I run across Person A making $B - which may or may not
be equal to $A - for Office B in pay period 2 of the same quarter... how
do you want that person and those dollars counted and totalled for a given
month in a given quarter? I don't know what you want - if I were a mind
reader then I might not be doing this for a living - so you have to tell
me how you want to see this; once you tell me the code will write
it'sself.'
(room goes silent)
The fellow then replies, in a small, weaselly voice, 'I don't see why you
just can't turn it back and make it the way it was before.'
'This *is* the way it was before, either the data have changed or someone
is noticing the error for the first time. If I change the data then the
results are inaccurate; I cannot change the fact that someone notices
errors. So... when Person A makes $A for Office A in pay period 1
(etc)... how do you want that counted?'
'I *still* don't see why *you* just can't turn if back and make it the way
it was before.'
DD
|
|