Switch preservation format extension from .tab to .tsv

[posixeleni]

Suggestion from Archivematica and Leanne Trimble (ScholarsPortal):
Archivematica is not able to recognize the .tab file extension, because there are multiple kinds of files that can have that extension (e.g. http://fileinfo.com/extension/tab). She is wondering whether there is any chance that Dataverse could consider switching to .tsv, which is basically the same thing but with a unique file extension (https://en.wikipedia.org/wiki/Tab-separated_values).

[pdurbin]

I like this idea (I used .tsv for files in https://github.com/pdurbin/vim-tsv ) but from a quick look at the code my impression is that ".tab" appears in many places in the code. It doesn't seem to be defined once, from what I can tell.

[axfelix]

Hrm, if we're talking about /preservation formats/, i.e., not something the user ever sees or probably thinks about, can we not just do a quick tsv-to-csv call internally wherever it'd be an issue? https://github.com/Jach/tab2csv

I know tab-separated is still fairly common in social sciences and the sort of disciplines that encounter STATA, SPSS, and other files that Dataverse is developed against, but it seems silly to be hamstrung by the file extension of such a generic and easily-transformed format.

[axfelix]

Clarification: by "internally wherever it'd be an issue," I mean "in the bridge code between Dataverse and Archivematica," not "replacing every instance of hardcoding in the existing Dataverse codebase" :)