3 ms·
I would hate spreadsheets a lot less if they were testable. Or didn’t “help” by silent type conversions that skew analyses (which you’ll never notice because,
by hprotagonist 7y ago
I would hate spreadsheets a lot less if they were testable.
Or didn’t “help” by silent type conversions that skew analyses (which you’ll never notice because, uh, no tests) https://www.sciencemag.org/news/2016/08/one-five-genetics-papers-contains-errors-thanks-microsoft-excel https://www.sciencemag.org/news/2016/08/one-five-genetics-pa...
- kyberias 7y agoWhat in spreadsheets is not testable? It's trivial to write a test that opens the spreadsheet in excel and tests any formula calculations via Excel Automation API.
- hprotagonist 7y agoand where do i store my tests? how do i version them, and the data? can i put it in travis? how many people ever even try?
- kyberias 7y agoYou store them in the same place where you place other tests. It's just code. You version the data as you version any other test data. Never used travis myself, I guess it's some kind of CI. Many people do this. It's perfectly normal. Of course if you don't even use Excel in the first place, I don't know why you're asking.
- l0w 7y agoFinally, thank you! Tests, destribution and versioning are my main concerns. I like where the people at stensila are going: https://stenci.la/blog/humane-sheets/ https://stenci.la/blog/humane-sheets/