david_goldsmith(_at_)taligent(_dot_)com (David Goldsmith) writes:
In the context of MIME, why is it necessary to specify the line separator
sequence at all?
Content types by their very nature restrict what can be placed in
bodies labeled by such content types. It is the communication of such
restrictions from composers to readers that make content types useful.
If MIME does not specify the line separator sequence for the type
"text/plain" (and thus other types under the top-level "text" type),
the type ends up having no meaning. All of the type information then
gets moved to the character set and one ends up having to have
knowledge of the particular character set in order to do anything with
a body part.
There are cases where MIME readers and gateways need to know the line
separator sequence of a text/plain object, but won't necessarily know
about all the possible character sets. Minimally conformant readers,
for example, need to be able to save such objects to files, converting
the line separator sequence to the local newline convention.
As an extreme example, I claim that without a 0D 0A line separator
restriction on text types, a charset of "GIF" could be legitimately
registered. Such a thing could work, I hope people can understand why
it would be problematic.
--
_.John G. Myers Internet: jgm+(_at_)CMU(_dot_)EDU
LoseNet: ...!seismo!ihnp4!wiscvm.wisc.edu!give!up