Wednesday, August 14, 2013

Amateur software is the bane of my life


an example: I have to parse some CSVs from another tool I didn't write. The folks who wrote it didn't use a standard CSV output generator, and thus failed to generate correct output a disturbing fraction of the time: Specifically: un-escaped embedded quoting characters. And this is in VERSION 2.0 of their software--seems like they never really examined their own output, and that no one else did either, meaning we have already parsed/ingested an unknown amount of data of now questionable validity.

I basically told them "you have to use a standard tool for this--one that YOU didn't write" so as to avoid this kind of thing, one that follows the conventions that pass for a standard for CSVs (which does not have a formal std, but does have well-known syntax).

I recommended three possible choices: Apache Commons CSV (not my fave, it's a little cumbersome), JavaCSV (my fave), and OpenCSV (fewer control options). I think they're going with Apache; I am using JavaCSV.

That is now resolved, but this is a beginner kind of error. Yeesh!