At 4:44 PM +0200 8/17/09, Heiner Strauß wrote: >> On Mon, Aug 17, 2009 at 06:18:45PM +0800, Erich Dollansky wrote: >> > On 17 August 2009 pm 18:09:06 cpghost wrote: >> > > On Mon, Aug 17, 2009 at 10:25:29AM +0200, Polytropon wrote: >> > > > By the way, where did I read that #define macro names have to >> > > > be unique within the first 6 (six) letters? :-) >> > > >> > > The 6 letters limit was actually a restriction of earlier >> > > linkers and it affected all identifiers of linkable objects >> > >> > I did not know that linkers resolved macros those days. >> >> Of course they didn't. But knowing that linkers restricted the >> identifiers' length to 6 chars, it made sense for preprocessors >> to restrict them as well before passing them to the compiler >> and linker. >> >> Actually, it's a bit more complicated than that, but the basic >> restriction came from the linkers, the preprocessors only inherited >> it. >> >> > Interesting. >> > >> > Erich >> >> Regards, >> -cpghost. > >Putting the symbol names in one word helped the linker / loader a lot. >Live was so easy. > >Heiner > >C (one word = 32 bit) .NOT. (some word processor software)
As something of an ancient curmudgeon these days, I've enjoyed this discussion. As speculation on my part, perhaps the six character limitation is less a software issue than an early architecture issue - DEC's PDP-6/10 design used 36-bit words and packed six characters (clearly from a limited subset of the then current ASCII) per word, making simple searches very effective through symbol tables with a simple word level compare loop. While likely not all that closely related to the issue, I recall a technique I was introduced to on Control Data systems called COSY, in which one punched binary coded Hollerith cards with two characters per column encoded (six bits per character). Of course, such cards required excellent handling equipment (which Control Data had) because a stack of cards punched with 960 holes in each one had lots of opportunity for hanging chads. -- Walter M. Pawley <w...@wump.org> Wump Research & Company 676 River Bend Road, Roseburg, OR 97471 541-672-8975 _______________________________________________ firstname.lastname@example.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"