🚀 go-pugleaf

[note my Followup-To: comp.std.internat]

In article <CKD.92Dec16130204@loiosh.eff.org>, ckd@eff.org (Christopher Davis) writes:
>  VA> You don't send 8 bits to a site which does not understand it. It
>  VA> simply makes no sense -- they don't have appopriate fonts on their
>  VA> terminals anyway.
> 
> How do you know if arbitrary-site understands it or not?  YOU CAN'T.  And
> remember, the terminals may be several years newer than the mailhost.

Also, the terminals needn't be directly connected to the host,
or even owned by the same organization as the host (thus the
capabilities of the fancy terminal may have no relevance for, or
its owner may have no influence on, the gateway host software).
But everybody knows that... :-)

> MTAs are not MUAs.

...which is of course a more concise way to put it.  Thanks!

I'm personally not too bound to 8 bits as the ultimate data unit
size.  Today's efforts of 'stomping out' 7-bit transmission
channels for being obsolete in the presence of superior *8-bit*
channels (even as it's probably a good idea) may seem amusing
sometime in the future, when 8 bits have likewise been stomped
out in favour of 16 bits, or whatever is the fashion.  So, I
prefer not to be too quick labelling every 7-bit application as
'buggy' today, especially when they weren't yesterday.

What?  Are there *still* 8-bit processor architectures out there
in the suburbs? ;-)

What we really need is to separate the application issues from
the transport issues.  The application writer (or the user) need
not, and *should* not, care about how the lower layers of the
communication medium are implemented.  The user may have specific
demands in terms of speed and functionality, but it shouldn't
matter to me whether this article is transported via an 8-bit
channel, a 7-bit channel, a deck of punched cards on a jet plane
or whatever, if it just reaches its destination unmodified and
within reasonable time.  By allowing the lower layers to pick
any communication channel they like, the best performance is
achieved.  Ask DHL...

Haven't people studied the OSI network reference model?
--
Anders Andersson, Dept. of Computer Systems, Uppsala University
Paper Mail: Box 325, S-751 05 UPPSALA, Sweden
Phone: +46 18 183170   EMail: andersa@DoCS.UU.SE