This is not a Stata question as such.

My department is a Windows / Stata environment but some of our collaborators do not use Stata, so I have a policy of storing all our data files both as dta and csv in the hope that most other statistical programs can read csv files (and also as a backup in case the University stops supporting Stata). However our IT is now recommending that we use xml instead of csv largely on the grounds that it supports Unicode and is a more modern format!

I have no experience with xml but I get the impression that it produces large datasets and is not as flexible as csv.

I would be grateful for any any advice / experience in choosing a suitable data format.

Thank you.

Eddy