Re: draft-nerenberg-imap-binary-00.txt

I read that spec.  It addresses point 3 from my previous post, and 
partially addresses point 1.  However, it introduces additional syntax 
issues:

(1a) I'm not convinced that the syntax model is appropriately general for 
similar conversion operations.  While I'm not opposed to binary moving 
forward by itself, it shouldn't move forward until we are confident that 
we're not creating lots of special case syntax variants.

(1b) I'm not convinced the restriction to leaf nodes is a good idea.  For 
example, a disconnected client which is caching entire messages could 
recognize a significant performance benefit if it fetched the entire 
message with all base64/QP encoding that's not in a security multipart 
removed.

(1c) It does not specify a syntax to request the size of a decoded body 
part.

A minor nit -- I wouldn't call a binary literal "literal8".  Vanilla IMAP 
literals already support 8-bit (when appropriately labelled).

		- Chris

--On Wednesday, November 29, 2000 11:33 -0700 Lyndon Nerenberg 
<[email protected]> wrote:

> Chris, have you seen the -01 version of the draft? It contains two
> substantial changes:
>
> 1) The command syntax was changed to "FETCH n BINARY[x.y]"
>
> 2) The literal8 syntax is now "~{nnn}CRLF..."
>
> The -01 document was submitted last week but isn't yet in the
> repository. For now you can FTP it from:
>
> ftp://ftp.messagingdirect.com/pub/staff/lyndon/drafts/draft-nerenberg-ima
> p-bina ry-01.txt
>
> --lyndon
>