Re: CHAT: Corrupt messages (was CHAT: Doug Ball, PhD)

2008/12/15 Philip Newton <philip.newton@...>

> On Mon, Dec 15, 2008 at 00:44, David McCann <david@...>
> wrote:
> > I get my CONLANG messages in a daily digest. If the problem doesn't
> > occur with individual emails, it could be that the server at Brown is
> > just taking the individual emails, each with its own encoding, and
> > sticking them together.
>
> Ah, that could well be the case -- that the list server isn't 'clever'
> enough to re-encode all messages into one consistent encoding.

We had the same discussion a little less than three months ago, where David
McCann was stumped by base64 in his digest. I still haven't checked for
myself how a digest actually looks at the MIME level, but I think you're
right.

With the open source MIME support libraries available these days, it would
be a SMOP to make a digest where all the mails are converted to Unicode and
encoded in utf8/base64 -- and there can't be many mail clients left that
won't handle that.

> In that case, naturally my software will assume
> > that the same encoding occurs throughout.
>
> Quite a reasonable assumption, and indeed, the only possible one,
> since you can't just put a new email 'header' (announcing a new
> Content-Transfer-Encoding or charset, for example) into the middle of
> the email's body. (Unless you use MIME multipart digests.)

Like the multipart/digest MIME type, for instance. It's almost as
lightweight as the current digests, and just as human readable even if the
mail reader has no support.

--
Lars

Reply