wollman@SADYE.UVM.EDU (Garrett Wollman) (05/23/91)
Here's another caption idea that I have recently had... Presently, the caption displays something like the "Real Name" of the sender of the message; it takes (so far as I can make out) the first "word", the "initial" of the second "word", and the last "word." However, this does not work very well for a number of cases. 1. When I receive a message from root, it will be listed as being from "EMBA C. Facility". Similarly, messages from MAILER-DAEMON are listed as being from "Mail D. Subsystem". I would argue that it would make a bit of sense for the caption-building code to recognize certain names which are likely to have non-name "names", and display them as if no "name" had been specified. I would suggest in particular that "root," "daemon," "MAILER-DAEMON," "uucp," "postmaster," "ftp," and "SYSTEM" all should receive this treatment. 2. Sometimes, especially in a mailing list like Info-Nets, I receive messages from people who have lots of miscellaneous junk in their "names." I would propose that the following happen (maybe some of them already do?): Everything after a comma is ignored. Any "word" which is numeric (ignoring punctuation) is ignored. This won't always help (for example, Ric Nauen's name comes over as "1159-DIRECTOR/OVERSEAS OPS" but then he's from a militar system anyway). But, it might make things a bit nicer to use. [I'd still like to see someone deal with % and ! addresses in a sensible manner--assuming a well-defined meaning can be worked out. However, it is fairly obviously that the Argentinian posters on Info-Nets are *not* employees of Thinking Machines, Incorporated. Similarly, many BITNET posters appear to be coming from CUNYVM.CUNY.EDU; I know that this is not the case, but that might not be so for a new user.] I would like to say, though, that I think the "name" feature in general is quite nice; if only people would stop putting extraneous junk in their From: fields. -GAWollman
nsb@THUMPER.BELLCORE.COM (Nathaniel Borenstein) (05/23/91)
The code that tries to extract the real human name from the "From" field is, inevitably, a set of heuristics. We tuned them for quite a while, and then moved on to other things. What it is good at, therefore, inevitably reflects the range of From: fields we were seeing at CMU circa 1987. It isn't surprising if what you're seeing at UVM in 1991 is different, or that our heuristics are not adequate for all those cases. In general, I found it a tiring and losing effort to constantly update those heuristics, so eventually I just stopped. Wouldn't it be nice if there was a standard header for "Sender-Real-Name" that everyone used? In the absence of such a standard, though, we have heuristics. The code that implements the heuristics is pretty straightforward, and can be found in ams/libs/ms/prettyn.c. I would encourage anyone who cares enough & has the energy to go fiddle with that file. In a sense, you can't really do any harm, since all you can do is alter the set of things that show up on the caption -- it isn't like you could cause mail to vanish or anything....