RE: X.509 Extensions Enhancements

["David A. Cooper" <[e-mail]>]

At 01:22 PM 6/28/01 -0700, Carlin Covey wrote:

>David,
>
>I am persuaded that X.680 & X.690 do intend the "delete all
>trailing zeros" interpretation, despite the ambiguities in
>crucial sentences.
>
>I appreciate your taking the time to respond to my posting.
>You have stated your position well, and I think you have probably
>captured the essence of David's logic.

Yes. The way that I read X.680, when dealing with a NamedBitList, all of the following bit strings must have the same meaning:

10110000
1011
10110
101100000000000000000000000000000000000000000000000
1011000000000

Since they all have the same meaning, we must select one as the "distinguished" encoding for all of them.

>However, it seems to me that any agreed-upon number of trailing
>zeros would have served the purposes of DER.  I agree
>that the shortest such bit string is more rational than longer
>encodings.

I suppose the rule could be "include the minimum number of trailing zeros necessary to make the length of the bit string a multiple of 8", but the simplest rule is just to require that all trailing zeros be removed.

>By the way, Bodo and Tom have both pointed out that the
>last bit must be a 1, so the final bit (and indeed any string of 1's
>at the end) is redundant if you know the number of bits that are
>represented in the encoding.  It may have been possible to create
>a shorter encoding that was incrementally more rational.  ;>)

Perhaps, but X.509 requires a distinguished encoding that is also a BER encoding. I suppose we could define a set of distinguished packed encoding rules and then specify an X.509 version 4 that used those rules, but doing so wouldn't be very rationale. :-)


Dave