From: Brad Templeton (brad@templetons.com)
Date: Mon Jul 02 2001 - 04:22:21 CDT
On Mon, Jul 02, 2001 at 09:29:56AM +0100, Clive D.W. Feather wrote:
> Brad Templeton said:
> > But once again, if a site has declared a policy of not wanting articles
> > larger than a megabyte, does that mean the poster should split into parts
> > to get around the policy?
>
> That depends on why it has the policy. It may be that they don't want
> *pieces* larger than 1Mb, because of the way they handle them internally.
> You don't know.
And that's why you do want a format where each article specifies both
the size of the component and the size of the final article.
The problem with message/partial, as defined by the mime group, is that
it doesn't tell you in each component what the total number of
components is or how big the final result is. As such, there is no
way your tools can, without gathering all the components first, examine
them to decide if they should be fed to you at all.
>
> - software MUST be able to handle articles up to 1Mb;
I agree, though today that is too small a number, and in a few years
it will be way too small.
> - agents MUST NOT prevent the use of message/partial just because the
> whole message or a given piece would be under some limit;
I'm confused. Why would anybody care if a message was under a limit?
The reason for saying "don't do a multi-segment message if your
article is less than [Size]" is to reduce or in fact eliminate the need for
things like message/partial outside of groups for mpegs. Meaning
the groups are far cleaner, and the software is simpler to write unless
you want to write it to reassemble and handle large mpegs.