New Years
This seems to most efficient way to send notice of my not attending any New Years parties this year. Living off the beaten path in a town that does a horrible job of keeping the roads clear has it's drawbacks.... and on a night when I would hope to laugh excessively and drink moderately the added prospect of driving late at night on bad roads seems to be a bad idea. So I'll miss most of you again this year as I hang out in my jammies, drinking champagne on my couch while getting beaten bloody by my wife's merciless scrabble-fu. It's tons more fun than it sounds. Hope the rest of you have a great new year!
Since I'm online... and one of a small handfull at the office let me take this time to add a bonus grumble to my rare posting. I hate working with scientists. "People think in floating point!" I keep hearing. Apparently they also think in at least 3rd order curve fits... however I do my best to think like the 8bit micro that has to run the f'ing show and it doesn't like wasting time on 3rd order equations on data sets that are not only very linear, but completely useless! I need to write one value out the spi to a dac on power up, and after that we simply step the dac value up or down to servo lock in a particular setpoint... taking care of course to not exceed some limits. There is no need for a 3rd order curve fit except to make an asshat feel better when using his labview based pc interface to control something that should be entirely in firmware already. I love my job, except when I get to work with Asshat 'n CO. Since they pay me, I made it work. I wish I had the raw data to work with... I bet I could do a nice linear fit and be accurate enough that they wouldn't even notice... 3 entry table lookup worst case.
Since I'm online... and one of a small handfull at the office let me take this time to add a bonus grumble to my rare posting. I hate working with scientists. "People think in floating point!" I keep hearing. Apparently they also think in at least 3rd order curve fits... however I do my best to think like the 8bit micro that has to run the f'ing show and it doesn't like wasting time on 3rd order equations on data sets that are not only very linear, but completely useless! I need to write one value out the spi to a dac on power up, and after that we simply step the dac value up or down to servo lock in a particular setpoint... taking care of course to not exceed some limits. There is no need for a 3rd order curve fit except to make an asshat feel better when using his labview based pc interface to control something that should be entirely in firmware already. I love my job, except when I get to work with Asshat 'n CO. Since they pay me, I made it work. I wish I had the raw data to work with... I bet I could do a nice linear fit and be accurate enough that they wouldn't even notice... 3 entry table lookup worst case.