I think that there's a bit of unfortunate terminology here though, in that "Windows" could be taken as a generic term representing any windowing user interface (i.e. including Gnome etc. on Linux).
Rarely heard, if ever.
I have, I assure you. There were /lots/ of windows libraries back in the day.
Any form of text-only interface, yes. Still, I don't think it is our job to give general IT education and terminology.
Until misconceptions impact on our own productivity. I remember elsewhere that somebody argued vehemently that X11 server vs client terminology was /wrong/, that sort of thing is not so bad in private discussion but once Google starts repeating erroneous opinions it get serious.
So that reduces the question to "could the user have found out that the unit is only meant for one (group of) target(s)?".
There is the RTL layout, and notes etc in the docs, but maybe something more prominent is needed. I think that is more constructive than discussing alternative interpretations of unit names. (specially as the interpretations might be depending on similarities in terminologies in other languages)
The bottom line is that if zhe tries to incorporate it into a program compilation should abort with a sensible message. Not "missing dependency", but "you can't do this". [Sigh] And there'd /still be people who'd try to disable the checking code.
I think as a beginner you are better of with a socket suite like Indy. If you want to go further, study it.
Oi! I think you mean /one/ is better off :-)
MarkMLl