Quote:
Originally Posted by saw7988
I was doing scientific computation stuff with numpy and pandas. Numpy arrays, panda series, and panda dataframes can all kinda semi-interoperate to some degree but not completely. Maybe not the most generalizable case, but I do think static typing helps a ton in general with writing code that runs correctly the first time, greatly shortening development time.
Sent from my Pixel using Tapatalk
Yeah, I can get this. I think that a lot of that is with csv in particular. I've done quite a bit with csv and pandas, and really the initial domino is at csv's flimsy half-assed format combined with every spreadsheet program having it's own interpretation of what is correct, assuming "correct" means "does nothing consistently."
I'm a firm believer in keeping data strongly typed. In my experience, malformed data with improper typing is a layer of hell that sticks its tendrils throughout the rest of the program. A proper database makes an immense difference here.
When I talk about about static -vs- dynamic, I'm not really talking about raw data, I'm talking about the general system of programming. Of course, trashy data makes things a whole lot worse, and it's possible when people mean type, they really mean dealing with ****ed up data, which isn't the same thing at all.