Re: [Gen-art] Genart last call review of draft-ietf-httpbis-header-structure-18

[David Schinazi <[e-mail]>]

On Mon, May 4, 2020 at 10:02 PM Mark Nottingham <[e-mail]> wrote:

> Hi David,
>
> Thanks for the comments. Responses below; I've committed in <
> https://github.com/httpwg/http-extensions/commit/001023>.
>

Thanks for making the changes! That commit looks good to me.


> > On 5 May 2020, at 11:18 am, David Schinazi via Datatracker <
> [e-mail]> wrote:
> >
> > Nits/editorial comments:
> >
> > In s1.2 (Notational Conventions), I didn't understand what greedy meant
> in:
> >   In some places, the algorithms are "greedy" with
> >   whitespace, but this should not affect conformance.
>
> Hmm, I think we can remove that sentence.
>
> > In s2 (Defining New Structured Fields), perhaps "Reference this
> specification."
> >  should instead be "Normatively reference this specification." ?
>
> Sounds good.
>
> > In s2, the definition of Foo-Example Header seems to be enclosed in
> >  "--8<--" and "-->8--" in the TXT version, could be a bug in the tools?
>
> Our AD commented that it was difficult to distinguish the example spec
> text from the surrounding spec text in the text/plain rendering. These
> "scissor" marks were intended to serve that purpose; I suppose they're not
> as common as they used to be. My assumption is that the RFC Editor is going
> to propose a more suitable way to do this.
>

Ah! Now I see it :) Happy to let the RFC editor decide what's best here.


> > In s3.1.2 and s3.2, in the example, I was confused by "a=?0" and "b=?0"
> until I
> > s3.3.6.
> >    Perhaps reordering sections or adding a reference would help?
>
> I think a reference.
>
> > Should there be some guidance for defining new integer fields that don't
> fit in
> > 10^15?
> >    Is a String the recommended approach?
>
> I'm a little wary of giving a single recommendation here; it depends on
> the use case. It might be that it would be better to use two integers, for
> example, and add, multiply or otherwise combine them. Or it might make
> sense to implicitly multiple (e.g., *100) the value. Or it might make sense
> to yes, use a string -- or binary.
>

Sounds good. I was mainly curious because I defined a sh-integer in one of
my drafts for a value that can in theory go up to 2^62-1, and I wonder if
it's worth the added complexity to support values between 10^15 and 2^62...
https://tools.ietf.org/html/draft-schinazi-masque-connect-udp-00#section-5
But we can figure that out in the context of that draft, no need to answer
that question in draft-ietf-httpbis-header-structure if there's no easy
solution that fits all.