This is a remarkably stupid workaround for a remarkably stupid problem: If you're running a Turkish version of Windows, you can't safely read XML documents;the encoding in the XML PI is read with culture sensitive code, and in the turkish charset 'i' and capital 'I' aren't the same letter - so any encoding that has an 'i' in its name and you're out of luck. The workaround suggested: Uppercase your encodings.
I'm sure I'm reading to much into this, but its clear from the workaround that at least the writer does not think naturally of data as something derived from the internet, but rather something local that you can control. It's easy to read a corporate failure into this: Microsoft does not understand the distributed nature of the internet. Anything that is not tight coupling and full stack is just alien to MS.
The real workaround is prudent advice even if you're a Dane and don't suffer this particularly arcane problem: The localized windows versions can't really be used to develop software at all.
Posted by Claus at July 16, 2008 2:46 PMGood thing the string "utf-8" doesn't contain any i's...
yup - but a pity "latin1" does.
Posted by: Classy at July 16, 2008 8:39 PMYep, and iso-8859-9 too, the Turkish standard character set!
Posted by: Morten Høybye Frederiksen at July 16, 2008 8:49 PMI was just going to mention iso. What's more this also happens in the database libraries, e.g. DataTable. So let's suppose you have a column in the database called "id", crazy I know, and you accidentally try to read by name as "ID" - then you're screwed again.
Posted by: Classy at July 16, 2008 8:51 PM