ARPAVAX:mo (11/07/82)
Your comments about case sensitivity in 822 "host" and "user" names are partially incorrect and generally misleading. The 822 specification makes statements about the "interpretation" of fields. Let me explain further. 822 specifically says "user" names ARE case-sensitive and "host" names are NOT case sensitive. But NOWHERE does it say you must send the "host" name case-folded. It merely states the default interpretation of a host name is case-insenitive. I agree this is probably a botch (and there are those of us lobbying to get that changed). So, for the moment, your admonishment is generally correct - case distinctions in host names are difficult to deal with (but could be special-cased in the UUCP domain!!!), but user names are required to be case-sensitive. This doesn't mean Unix sites have to stop folding case for user names, but they can allow mixed case user names (oh, the things that would break!!!). But it DOES mean Unix mailers had better not screw with the case of user names when sending mail if you EVER intend to get mail to someone on a Multics system. -Mike
furuta (11/09/82)
Interesting that this plea for avoiding case sensitive site names should come in right now. But since it did... Can someone tell me how it is that I'm supposed to address a message to a site with a capital letter in its name? How can I prevent the various mailers from deciding to collapse the case of the name? --Rick
goutal (11/09/82)
Well, I can certainly see that for the moment there is a problem. However, if the ARPAnet standard, or whatever standard or system, merely translates things to some prearranged case, then the fact that decvax or whatever doesn't do the same thing is not the fault of the system that chooses an uppercase name. It is the fault of the software on decvax (sorry, Armando). In other words, if ALL systems were as consistent as their proponents claim to be, we wouldn't have this problem. I just can't see banishing some system just because some other system can't deal with the name of the first. Note that this could easily metamorphose into the old general debate about case sensitivity. It just reared its ugly head again. In a different form. Actually, I guess I would conclude that it's the fault of the implementor of the ARPAnet standard, such that an implemen- tation modifies the form of an address that it passes on, rather than doing the case conversion only for purposes of local comparison. If the ARPAnet site just left the case alone (except internally), then uucp sites could continue to be their old case-sensitive selves (just like I am light-sensitive -- who am I to throw stones?). Would't that work? Would that violate the ARPAnet standard? _ Kenn (decvax!)goutal