i'm guessing you don't want to hear about my fun, fun time mired in the bureaucracy of The Nation's Capital (TM) - suffice it to say that took the greater part of last week.
not that you want to know about the other parts of my internship either
but you know what YOU HAVE NO CHOICE
or i guess you could press back or close your browser or something
ANYWAY ask any programmer what the worst programming language ever created is and half of them will say COBOL. the other half will say Visual Basic. (the snarky ones will say INTERCAL though). i am spending the next month of my life in Visual Basic. and not a nice modern version, either - Visual Basic 6.0, released by Microsoft in 1998.
so, what is it that i'm doing exactly?
well, NASA's going to launch, in 2011, a satellite whose mission is simply to measure (and create pretty charts of) the salinity of the world's oceans. now, how the hell can a satellite know how salty the oceans are - it's not like saltwater looks any different from freshwater to a human eye or even a spiffy multispectral camera? from the mission description paper:
Salinity remote sensing is accomplished by measuring microwave emission from the sea surface in terms of a parameter called brightness temperature, in Kelvins (K), and correcting for other natural emission sources and sinks. Ocean brightness temperatures are related to the dielectric properties of seawater, and at lower microwave frequencies, these are modulated by salinity. The frequency 1.413 GHz (L-band) is sufficiently sensitive to salinity to make viable measurements (Klein and Swift, 1977; Swift and McIntosh, 1983) and is legally protected for scientific purposes (radio astronomy and Earth remote sensing) from radio interference. The dielectric model function, commonly known as the “Klein and Swift” was derived more than 30 years ago from rather antiquated seawater dielectric measurements, yet has proven to be quite accurate relative to recent experiments (Wilson et al., 2004). A comprehensive new set of seawater dielectric measurements at 1.413 GHz are now being done to update the model (Lang et al., 2007).
"Lang" refers to a certain Professor Roger Lang at George Washington University, and "et al." refers to a couple of grad students and, for this summer, me. my job is to streamline the process of taking these measurements (the key word here is "comprehensive;" this has taken four years and will take four more; anything we can do to speed up the process helps). unfortunately, the current software being used to automate the measurements was written four years ago by a GW undergrad as his senior project (something like that would never happen at TJ now would it). he wrote it in VB6. we now have to live with the consequences, since i for one am not going to rewrite the whole bloated thing (and nor will any of the grad students, for that matter).
but it's worse than that - what this program does consists of two main functions: it controls the measurement apparatus (a network analyzer to measure the actual dielectric properties and a multimeter to record temperature readings from thermistors) using an even more godawful language (think assembly but controlling a network analyzer whose manual is a thousand pages long) and it combines all the measurements into nice little excel spreadsheets with a bunch of pretty graphs using macros (for an example of the idiocy of the existing program, look no further than the fact that it produces spreadsheets which have time along the horizontal axis (skipping three-fourths of the columns for good measure), necessitating that the data for a single experiment be split among three unwieldy spreadsheets)
doesn't this just make all of you want to go into academia
right now
