RE: No decimal bitstrings

[Black_David@emc.com]

Kindly leave calling of "rough consensus" to the co-chairs.

What I think I've seen so far is that decimal does not work
and needs to be forbidden for *variable length* binary bit
strings whose length cannot be determined by means other
than looking at the value.  SRP_s is an example, as the
length of a password "salt" may vary based on implementation,
but if different length bit strings are used by the initiator
and target, the SRP authentication will almost certainly fail
(Paul Koning was right about this one).

OTOH, decimal seems to work fine for fixed length binary
items where the fixed length is known such as the components
of an IP address (each component is known to be 8 bits a priori,
hence 0x1 and 0x01 represent the same component), or an iSCSI
Target Portal Group Tag (known to be 16 bits a priori, hence
0x1, 0x01, 0x001, and 0x0001 are all the same tag).

<SOAPBOX>
Can we also stop using the word "vote"?  If I see 7 people in
favor of one position, one person in favor of a different one,
and the latter one person has technical correctness on his/her
side, that one person's position will carry the day independent
of how many people are involved.  IETF is about doing the
proverbial "right thing", which is not necessarily the most
popular thing (just wait until you see my WG last call comments
on iSCSI ...).
</SOAPBOX>

Thanks,
--David
---------------------------------------------------
David L. Black, Senior Technologist
EMC Corporation, 42 South St., Hopkinton, MA  01748
+1 (508) 249-6449            FAX: +1 (508) 497-8018
black_david@emc.com       Mobile: +1 (978) 394-7754
---------------------------------------------------



> -----Original Message-----
> From: Paul Koning [mailto:ni1d@arrl.net]
> Sent: Friday, July 05, 2002 2:32 PM
> To: ips@ece.cmu.edu
> Subject: Re: No decimal bitstrings
> 
> 
> I believe there are a bunch of comments in favor of not allowing
> decimal for bitstrings (only for integers) and only one against.
> 
> This seems to be "rough consensus".
> 
>      paul
>