I have just got the latest version of the draft. I am most annoyed by
a problem of terminology : the expression 'body part' has two
meanings. It means both 'the part of a message that is not the header'
(as defined since the dawn of ages by RFC 822), and 'one section of
the body part of a multipart message' (this is new). This is
unnecessarily confusing.
Want an example ? See this :
If a Content-Transfer-Encoding header field appears as part
of a multipart message body part's headers, it applies only
to the body part. If the body part is itself of type
|||||||||
Quiz : what is the exact meaning of this occurence of 'body part' (you
have three seconds to answer) ?
Two solutions seem possible :
1- replace all occurences of 'body part' meaning 'one section of the
body of a multipart message' by the expression 'part-encapsulation',
which is used in the BNF the denote that kind of object ;
2- replace both the same occurences in the text and
'part-encapsulation' in the BNF by 'body-part'.
3- (I said two but...) as 2-, but replace by something else
(body-section ? Run a contest for the best name ?).
Anyway, the expression used in the text should match the BNF. The
second solution seems weaker, since one must still be very careful not
to mix body part and body-part.
The example would then become (with part-encapsulation) :
If a Content-Transfer-Encoding header field appears as part
of the header of a part-encapsulation in a multipart,
it applies only to the body part of that part-encapsulation.
Another related problem is that the expression 'the body or body
part', which is used at several places, is not correct (even with body
part replaced by part-encapsulation or by body-part or ..). The
problem is that the current expression implies that the body of a
message and a section of the body of a multipart message are at the
same level, which is not true. It is the body part of a section that
is at the same level as the body part of a message. So that expression
should be replaced by 'the body of the message or of the
[body-part][part-encapsulation] '.
Is this nit-picking ? While I must often recognize that I am indulging
at such an activity, I don't think it is the case here. MIME is a
rather complex document, and having such a major ambiguity in the
terms used could lead to some bad results.
/AF