Filthy weather day.

Another one to add to the recent cache.

We've had enough now. OK?????

Best viewed Large to get the best effect.

Spent all day asking myself - Why do I get all the rubbish jobs? Who'd be an analyst?

Work pressures are taking an awful lot of brain time today. I need to do a brain dump….

On top the awful weather, it's been a day of trying to work out what went wrong over recent weeks and especially Wednesday and yesterday. Having updated 635 equipment records (each taking several steps) covering ten new TV transmitter systems (Anyone want a masterclass on how your terrestrial tv gets broadcast?) I discovered last thing last night that the inventory system hadn't been updating related items as it should as I have been amending the data.

No idea why. The prod (live production) system did it correctly when I tested it a couple of weeks ago. The test system did things properly today, but, the prod system hadn't been doing it - I hadn't noticed, nor do I know how long its not been working. I didn't know how much of the recent work could be incorrect or how much poor input data had polluted the work. Grrr.

Now, this left me with a problem because while the source data I'd been given was flaky, the method I was using would have prevented data problems from propagating through if things were working properly. The problem meant this self-checking was no longer the case and I have to go through every changed record again and check it, and, where it hadn't updated as expected, correct it. The thought of doing it all manually is soul destroying. At least a weeks work lost (Don't tell the boss - it will upset him).

A nightmare. It's taken all day to work out how to do it and produce a test plan to check it was going to work. This is still better than the five days it taken to get to here.Just got the changes to errored records to do now.

The upside is I have also hopefully identified a way of automating the update process to just a couple of steps for each of the outstanding systems yet to be done - Having tried before I thought this was feasible. Even if I can sort the process, it will still need some work to put together and test. The conundrum is however, do I keep going the current way (and meet the targets) or do I spend days writing the code and database applications to that will cut down the update for each system from about three hours to fifteen minutes? With about 65 more systems to do over coming months, its a bit of a difficult question to ponder. The immediate cost/payoff against the time pressures doesn't work, but over a longer period, it would be helpful. But do the updates manually for a while and then automate later means it won't pay back time wise at all. Working alone doesn't help - no-one to use as a sounding board. Rock and a hard place.

Brain dump complete. Now….

If you've got this far though this incomprehensible drivel, you deserve to have a great weekend.