It's quite likely that CP852 came from IBM, since they were the source of most localization/internationalization (L8N/I8N) data at the time.<p>Microsoft, basically, had no clue: they relied on local partners to request L8N-related features. Then, they tried to figure out if those requests made any sense. For some markets, like Japan, there were existing industry standards that made this easy-ish, but for most other markets, it was pretty much guesswork, and, based on expected dollar-value, they would send out consultants to try to figure things out.<p>So, why was IBM a valuable/likely data source? Simple: they had been selling typewriters into most markets since, like, forever. So, they had to manufacture Selectric "typing elements" (the wheels and/or balls that produced the actual characters) to match the local market needs.<p>CP852 is most likely a more-or-less literal mapping of the Selectric typing elements that were most popular in the "Eastern European" markets.
Very interesting read as for some reason I've been fascinated with old code pages lately.<p>Can anybody point me to some more resources about this general topic? The Wikipedia articles about code pages are all so generic and uninformative.<p>Or maybe answer some of my specific questions: When was the first DOS with support for Icelandic shipped? When for Hebrew? Why are there three different code pages defined for Greek and were they all supported by Microsoft/IBM?