Re: WG Chair Nits & start of WG Last Call: draft-ietf-secsh-publickeyfile-06.txt

[Joseph Galbraith <galb-list%vandyke.com@localhost>]

Ben Harris wrote:
> I took a printout with me on a train yesterday, and it's now covered in blue
> ink.  Most of my suggestions are in the form of a diff, but there are a few
> things that merit explanation.

Thanks.

> I think having things specified as being permitted (or encouraged) "for an
> interim period," without specifying the length of that period, is pointless. 
> My patch removes such references on the assumption that the spec can be
> amended if the interim period ever ends.

I have an alternate proposal.  The draft has been around for a while
now; how about we declare the interim period as ended.

I've edited the draft accordingly.

> The case-sensitivity or otherwise of Header-tags is unspecified.  Since
> RFC-822 headers are case-insensitive, I think this spec needs to be clear.

Specified as insensitive.

>    Existing implementations may not correctly handle unrecognized
>    fields.  During a transition period, implementations SHOULD generate
>    key file headers that contain only a Subject field followed by a
>    Comment field.
>
> It's unclear whether this means that implementations SHOULD emit both
> fields, or at most one of each in that order.

Removed completely as 'transition period' is now over.

>    Contain a user specified comment which will be displayed when using
>    the key.
>
> I think "will" needs to be replaced with "MAY", "SHOULD", or "MUST", but I
> don't know which is intended.  The sentence should perhaps also be recast in
> the active voice to make it clear who's doing the displaying.

Done.

> There should perhaps be an example of wrapped lines.

Done.

> There's no mechanism for defining new headers.  I'm not sure whether there
> needs to be, but it seems to be usual to at least allocate some namespace
> for private use.

How about:

   New headers that are of the from "x-" are considered experimental,
   and may be used without IETF consensus.

   All other headers are reserved for use only by IETF consensus.


> It's traditional to list RFCs in the References section in numerical order.

Done.

> I don't think the Trademark Notice is required, since neither of the authors
> is affiliated with ssh.com.

Removed.  I wasn't sure.

> Now for the dull fixes -- I've avoided rewrapping paragraphs so as to keep
> the diff intelligible.

Thanks... since I don't work on the text, but rather an XML source,
I applied manually.

Hopefully I didn't mess anything up.

> @@ -172,7 +172,7 @@
>  
>     In order to use public key authentication, public keys must be
>     exchanged between client and server.  This document formally
> -   describes the existing public key file format, with few exceptions.
> +   describes an existing public key file format.
>  
>     Where this document departs from current practice, it also suggests a
>     mechanism for backwards compatibility.

Done.

Also removed:

-  Where this document departs from current practice, it also suggests a
-  mechanism for backwards compatibility.

>     Existing implementations may not correctly handle unrecognized
> -   fields.  During a transition period, implementations SHOULD generate
> +   fields.  Implementations SHOULD generate
>     key file headers that contain only a Subject field followed by a
>     Comment field.

Removed this paragraph entirely.

> -   During an interim period implementations MAY include the quotes.  If
> +   Implementations MAY include the quotation marks.  If
>     the first and last characters of the Header-value are matching
> -   quotes, implementations SHOULD remove them before using the value.
> +   quotation marks, implementations SHOULD remove them before using the value.

Here I did exactly as your patch did and left in the MAY include quotes;
this doesn't seem to be that large a burden to just leave permenantly.

Thanks,

Joseph

PS. I'll wait a day for other comments and then send an updated version.