Subj : Re: MMTerm?
To   : Jim Hanoian
From : William McBrine
Date : Wed Apr 18 2001 06:03 pm

-=> Jim Hanoian wrote to William McBrine <=-

 JH> initial root of the all caps header.  Also, PCBoard messages had
 JH> a size limit of 64k (I think it was) which was later "updated" to
 JH> permit messages of a liberal 100 lines.

64K is much more than 100 lines, actually -- figure 80 chars (ignoring that
most lines are slightly shorter) * number of lines = 8K for 100 lines.

 JH> I remember TomCat, and WildCat still uses the "T" command to get into
 JH> mail, doesn't it?

Indeed. They took away the TomCat name, but kept the key for the benefit
people who were already in the habit, and made up a new string to fit it
("Transfer QWK mail", or something like that).

 JH> [OPX doors are] Much more consistent that the ways that different QWK 
 JH> doors worked with different BBS types, right?

In some ways; but overall, not really. Take the line endings (please) --
in QWK, every line ends with character 0xE3. This is unambigious. In Blue 
Wave, they're defined as CRs to terminate each paragraph, with LFs to be 
ignored if present. In OPX, I discovered that they can be either LF, CR, 
CRLF, or even a misused "soft CR", depending on the door. It's psychotic.

Then there's the EXTAREAS.DAT file. In some doors, this is used to list any 
areas after the first 256. But it's not actually necessary, because other 
doors just add those areas to BRDINFO.DAT along with the first 256.

I could go on and on with this. There's also the internal inconsistency
found in every implementation, because it's part of the spec: there's a mix
of C and Pascal-style strings, for example. (It's mostly Pascal-style, but
the message headers are lifted whole from the Fido specs, which use C.)

.... MultiMail: http://multimail.sourceforge.net/
--- MultiMail/Linux v0.40
 * Origin: COMM Port OS/2 juge.com 204.89.247.1 (281) 980-9671 (1:106/2000)

.