Re: Silly question

1998-06-15 19:20:17
On June 15, 1998 at 18:19, Riccardo Cozzani wrote:

if in the body of the message are some pieces of html code MhonArch quote
them, so I can see in my browser only the code, not the result of it.
Can I turn off this option? I think not (not found in resources), but if
you have some suggestions, please let me know.

Al gave a fairly complete answer.  I'll try to elaborate (and
correct as needed):

MHonArc basis its conversion of MIME content-types (aka media-types).
Therefore, if a message is labeled "text/plain", it will be treated as
text/plain.  If a user wants to send HTML text, then they should label
the text with a "text/html" content-type.

Now Al wrote:

Warning: that is not the way the world is going.  The
conventional wisdom is that mail clients (and hence archivers
like MHonArc) process messages as text/* and sniff for HTML tags
on suspicion that they may have a web-formatted document in the

This is incorrect.  Not even Netscape's mailer does this with
text/plain data.  At the most, a nice mail viewer will sniff out text
that looks like URLs and make them selectable (MHonArc does this).  At
the worst, the link may be bogus, but it does not alter the rendition
of the data in a way contradictory to the given content-type.

In general, some heuristics may be applied to text/plain to provide
some added benefit for the user, but caution should be observed to not
make assumptions on the data-type from what it is labeled as.  This
would be bad practice, and undermine the whole idea of labeling message
data with content-types.

    [Side note: Mailers that use proportional spaced fonts
     to render text/plain messages with auto-line wrapping are in
     error.  This goes against all the known mail standards.  PC based
     mailers are known to do this, and it is just plain WRONG.]

If you have software that kicks into "HTML-mode" when it sees tags in
text/plain data, senders would have no way to represent raw HTML code
in their message.  This would be a major pain for lists like this one
(where sample resource files with HTML markup occurs).  Software
should never contradict what the underlying data says it is (and
which a human -- the sender -- says it is).

MIME gives ways to include HTML data in messages.  If the whole message
is HTML, then the main content-type can be text/html.  MHonArc can
handle this.  Also, you can use multipart/* type to intermix HTML with
plain text (or other data types).  MHonArc will handle this also.

Now, with all that said, MHonArc is not draconian in its implementation
of MIME and in message conversion.  For example, the text/plain
filter does support options that technically contradict how text/plain
should be rendered.  However, you, as the archiver, make that choice.
By default, MHonArc does the right thing.

If you really need the ability for MHonArc to try to auto-detect HTML
markup in text/plain messages, then you can register your own filter to
do such a thing (it does require some knowledge of Perl).  How to do
this is covered in the MIMEFILTERS resource page.


<Prev in Thread] Current Thread [Next in Thread>