Re: draft-ietf-impp-cpim-msgfmt syntax

On Mon, 2002-11-11 at 20:30, Dave Crocker wrote:
> Let's say a small message is a kilobyte with RFC2822 syntax and 2 kilobytes
> with an XML syntax. That's 100% extra overhead. The only problem is that 2
> kilobytes is still small enough to not matter.

Perhaps in isolation.

Each MIT zephyr server handles an average of 63 packets per second. 
(Zephyr is a UDP protocol, so about half those packets are acks and half
are requests.  But there are two servers, so we'll let those cancel.)  A
savings of 1K may not matter, but a savings of half a megabit might.

I don't favor changing the draft because:

  * header:value is more readable than XML when nesting is not required.

  * header:value takes very little code to implement.  (RFC 822 takes a
lot of code to implement, because it is remarkably permissive.  But this
is way simpler than RFC 822.)  It would take more code to interface with
an XML parser than it would to implement header:value.

  * Although header:value is not the most compact representation because
it does not take advantage of field order, it only uses space for
legitimate reasons (extensibility and readability).  XML wastes space to
no particular end.

  * (Not a new argument.)  This working group appears to be nearing
completion, and these drafts have been around for long enough to have
been implemented.  Changing the encoding would be disruptive.




  [reminder: [email protected] for non-technical discussions, please]