• Root (271)
  • digest (139)
    • debate (1)
    • economy (12)
    • education (5)
    • freedom (7)
    • freespeech (18)
    • idiots (9)
    • media (26)
    • music (1)
    • politics (29)
    • religion (1)
    • society (2)
    • stumbleupon (6)
    • tech-sci (3)
  • fun (22)
    • videos (7)
  • ideas (6)
    • epistemology (5)
  • meta (6)
  • misc (11)
  • news (36)
    • media (5)
    • politics (25)
  • personal (12)
    • florin (5)
    • lupo (6)
  • series (4)
    • complexity (1)
    • ponerology (3)
  • tech-sci (35)
    • comp (4)
    • medicine (28)
    • psychology (1)
    • quantum (1)

RSS feed

GPG/PGP pubkeys

IgorPro bug when saving "unpacked experiments"

author: www-data

I'm (still) using WaveMetric's Igor Pro v5 inside a virtual machine for data analysis. And, like a good computer geek I am, I'm managing my data analysis files using a decent versioning system: git.

Problem: Igor Pro .pxp files (Packed Experiment Files) are simply too big to be handled by git. And, as any decently "packed" files, one minor change ripples across the whole file, so I end up with 500 MB large patches in my versioning system. Ugly.

Sollution: use unpacked experiments files of Igor (.uxp). The data is then saved in a directory, with many subdirectories and files, resembling the organisational structure inside my 'experiment document'. The best part: changes to single data sets remain purely local, and as such they can be easily managed by a versioning system.

Problem to the sollution: Igor sometimes refuses saving .uxp files with the error: "The system cannot find the specified path." Saving as .pxp would work, but that, again, would cause problems with git.

Sollution to the problem to the sollution: (...is more of a workaround...) don't use 1-letter data folder names. Igor will let you create them, but will not be able to save them as .uxp. Use two or more lettes instead, and your problems are gone :-)

(For the record: No, I don't like Igor particularly much. It is powerful and pretty efficient indeed, but very clumsy at times. And, like all Windows software, it suffers from the "we have anything you need, why in the world would you *ever* want to use anything else?"-syndrome, making interoperability with other tools a pain in the ass. And it's an ugly monolith, that has been growing over the years carrying lots of historical baggage. And one gets to feel that at every move. I find myself working around Igor's brain-damaged historical mistakes 70% of the time I should be actually spending analysing data.)

2009-09-08 15:02 | www-datarootshell.ro | [/digest/tech-sci] | permanent link