On 01/02/12 05:53, Andy Moreton wrote: > Microsoft themselves rarely refer to it as the "Windows API"
My impression is otherwise. For what it's worth, the Google search '"Windows API' site:microsoft.com' has about 68,900 hits, which is nearly as many hits as "Win32 API" at the same site. And outside Microsoft, a Google search suggests that "Windows API" is used more than twice as often as "Win32 API". > (a less specific term in any case). I'm no expert in the Windows API, but I suspect that in many (most?) cases the less-specific term is appropriate in Gnulib. We can change the documentation to say something like "32-bit Windows API" for the part where the number of bits actually matters and is not already obvious from the context. That'd be a further improvement that is largely independent of the proposed change. > Using the term "win32" does not connote winning or losing No doubt it has that connotation only to some people, and not to others. But we should be communicating to as many people as possible, including the people who see the connotation. > The term "woe32" is insulting rather than descriptive This seems to be inconsistent with the previous quote. If "win32" doesn't connote "win", why would "woe32" connote "woe"? I used "Woe32" because that seemed to be the second most commonly used Gnulib term for this notion: it's currently used 68 times in Gnulib. "Woe32" is short for "Windows Operating Environment, 32-bit". The phrase "Windows Operating Environment" itself is perfectly respectable, being used about 20,000 times on microsoft.com according to Google. I personally might have capitalized the whole thing, e.g., "WOE" for the generic term and "WOE32" when it's important to refer to the 32-bit variant; but the patch defers to existing convention. Admittedly the Microsoft marketing department is not likely to approve of the acronym "WOE", but we're under no obligation to do Microsoft's marketing.
