On Aug 25, 2011, at 12:15 PM, Marc Petit-Huguenin wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> (Sorry for the convoluted title, it seems I can no longer say 
> "Interoperability
> problem")
> 
> It is about the size of i in section 6.3.4 (and d in section 8).  I remember
> Cullen saying that it is 4 bytes, but it seems that the spec was never updated
> to reflect that.

Uh, I think the resolution was the other way around for one and as you say for 
the other. Let me separate the two..

First for the size of i, in section 6.3.4. The max value of this is controlled 
by the max-node-multiple values with is limited in the XML grammar to xsd:int. 
It never shows up on the wire so I think this is defined as 4 bytes. Given the 
need to interate through all values up to the limit computing a hash each time, 
people will need to stick to small values. 

On the other one, d in section 8, I think we decided on 8 bits was enough. Back 
in draft-ietf-p2psip-base-13 the relax NG for turn-density defined it as and 
xsd:int which did not match the unit8 in for d in section 8 but we moved the 
XML grammar to be xsd:unsignedByte  in version 14 of the draft.  

So, I think we decided 4 bytes for the first and 1 byte for the second. I think 
the doc is consistent with that. Is there a part that is inconsistent where I 
need to update it. 

Cullen

_______________________________________________
P2PSIP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/p2psip

Reply via email to