Hi Alan,

Thanks for that quick note. It is very helpful. I will look into this
further and also consult other people.

Most important part that I missed:
------------------------------------------------------------------------------

rand.c: By Bob Jenkins.  My random number generator, ISAAC.  Public Domain.
MODIFIED:
  960327: Creation (addition of randinit, really)
  970719: use context, not global variables, for internal state
  980324: added main (ifdef'ed out), also rearranged randinit()*
010626: Note that this is public domain*
------------------------------------------------------------------------------


In the meantime, I would request the authors and WG to continue work on
addressing the review comments.

Thanks,
Ketan


On Thu, May 15, 2025 at 8:00 PM Alan DeKok <[email protected]> wrote:

>   Just a quick note:
>
> > On May 15, 2025, at 9:48 AM, Ketan Talaulikar <[email protected]>
> wrote:
> > Major Concern:
> > The reference to ISAAC has to be normative for this document and that
> reference is today a private webpage. This is not something that could be
> considered a stable reference for an RFC publication. We will need to
> figure this out and please check in my detailed comments for some
> suggestions.
>
>    RFC 5475 and RFC 8290 reference other personal web pages by the same
> author.  So there is at least some precedent for this reference.
>
> > ...
> > <major> The first reference has got to be normative. Additionally,
> > the reference is to a private blog site and its availability is suspect
> > post publication as an RFC.
>
>   RFC 8290 depends on a hash function from the same web site, and likewise
> has the reference as informative.
>
> > Please see if there is a better/stable reference.
>
>   There isn't.
>
> > Another option is if the actual algorithm or code could be placed in the
> > appendix (subject to copyright considerations - I am not an expert on
> this
> > and so we will need to consult - but if the inventor could agree for the
> > portion of the paper/code can be included in the appendix).
>
>   The ISAAC code is in the public domain:
> http://www.burtleburtle.net/bob/c/rand.c
>
>   I believe it could be included in the draft as-is without modification.
>
>   I've implemented a bare-bones version of the ISAAC paging method
> described in the draft.  The source code is on the GitHub repository:
> https://github.com/mjethanandani/bfd-secure-sequence-numbers/tree/v20/isaac
>
>   I'm will to either place that into the public domain (for modifications
> to the ISAAC rand.c file), or assign copyright to the IETF.
>
>   I'm not sure it would be good to include all of that code into the
> draft, but it would certainly help with interoperability.
>
>   Alan DeKok.
>
>

Reply via email to