On Jul 12, 2010, at 18:10, Daniel Gavelle wrote:

> Carsten and Colin,
> 
> The codes for repeated bytes in Carsten's draft looks similar to the LZ77 
> compression used in Deflate.  

Of course, backreferences are being used in LZ77, so I'm not claiming anything 
in my draft is in any way new.

> Maybe Deflate could be considered for the ICMP compression.  

As the half-draft says, I thought about that.
Too complicated for my taste (even if you take out most of the gratuitous 
choice that is in RFC 1951).

> Nodes may need Deflate code anyway, either for HTTP compression or X.509 
> certificate compression.

I'm mostly thinking about nodes here that are too limited for either HTTP or 
X.509.
While deflate is not *that* complicated, the eight specially crafted bytecodes 
are about two orders of magnitude simpler.
(Maybe one can make them even simpler, but I didn't invest much time in this 
demonstration of concept.)

I should probably also clarify that I wrote this draft to un-stall HC, not to 
add complexity to ND.

Gruesse, Carsten

_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to