> The translator documentation that you pointed to is a nice overview from a > user's point of view, but I have (so far) been unable to work out how to > actually *write* a translator. (I also haven't had time for a month > or two to try..) Would it be possible for someone (preferably someone who > knows something about this topic ;-) ) to write some quick documentation on > how to write a very simple translator? A "hello, world!" type of program, > for example, would be very useful; perhaps a tiny translator that just prints > the above message when 'cat'ed, or shows a file containing "Hello, world!" > when "ls"'ed?
I am getting the impression that you have not read the Hurd Reference Manual (in doc/hurd.texi). It is far from complete and (might well have some inaccuracies, I'm not sure), but Thomas and Gordon have spent no small amount of effort in documenting the system from the programmer's point of view. I think there is certainly enough there to sink your teeth into and come up with specific questions rather than just expressing a complete lack of direction. And for goodness sake, man--read the source! Are you honestly telling me that after reading trans/null.c (the source for the /hurd/null translator that implements /dev/null) you cannot figure out for yourself how to produce a trivial "hello world" translator as you described above? We endeavor to make the system easy to program, and some of us are working on improving the documentation (I think Gord is pretty much in charge of that). But I don't think anyone should expect to know how to program the system without poking around a bit in its source code and seeing how things are done in practice. I don't even think that should be any kind of goal. I mean, that's why we make the software free and ensure that everyone can see the source code.
