Re: Request for comments: Delegation of urn:ogf

[Freek Dijkstra <[e-mail]>]

Hello Mykyta,

Thanks for reading our draft and providing useful comments!

> I see RFC 2141 didn't use ABNF for definition, so it's impossible to 
> import the <URN chars> rule from their.  It may be displayed as follows:
>>        SUBNAMESPACE-SPECIFIC-STRING  = 1*uchar
>>        uchar       = trans / pct-encoded
>>        trans       = ALPHA / DIGIT / other-uchar
>>        other-uchar = "(" / ")" / "+" / "," / "-" / "." /
>>                      ":" / "=" / "@" / ";" / "$" / "_" /
>>                      "!" / "*" / "'"
>>        pct-encoded = "%" 2HEXDIG

You are correct. Indeed RFC 2141 uses BNF instead of ABFN, so you can
not import that as-is.

I'm a bit hesitant to add this to the draft, though. In practice, people
who want to use ABNF can easily translate the BNF from RFC 2141 themselves.

The reason I'm hesitant is that such definition (or should be) in
RFC2141bis, not in this document.

Given that RFC2141bis is not finished, I don't know what will be in
there. For example, your definition deliberately excludes RFC 2141
<reserved> from allowed chars. RFC2141bis may make a different decision.
I hope to avoid such potential ambiguity by simply referring to RFC2141.

> Moreover, such definition will require changing the following text (and 
> the whole Section 2.11):
>>     It is
>>     RECOMMENDED that these documents forbid the assignment of URNs
>>     containing characters in the<reserved>  set ("%", "/", "?", and "#")
>>     as defined in [RFC2141].
> My proposal is to mention that RFC 2141 <reserved> chars are 
> deliberately excluded from the allowed in <SUBNAMESPACE-SPECIFIC-STRING> 
> because of restrictions made by RFC 3986, explicitly Section 2.2 of it.  
> Such characters should be percent-encoded if occur.

This is a useful clarification, I'll gladly include it, thanks.

> Section 2.5, Ancillary documentation.  I think a here should be a 
> reference to your GFD.1 publication, describing the GFD series.

I referred to the URL page that lists the series. Indeed, it is better
to refer to GFD.1 instead.

> So far, I see, OGF has published 189 GFDs.  Is your draft going to 
> provide a reservations for something like "legacy" GFDs, eg. 
> "urn:ofg:legacy:<number>" or this will be left to OFGF Director?  
> Anyway, your document should mention the relationships between 
> so-far-published GFDs and OGF URNs.

No, this draft will not defined how to define subnamespaces. It is
planned that this is done in a GFD document. A draft of that document is
provided at http://forge.ogf.org/sf/go/artf6478.
I expect that is needs a revision or two.

> As an editorial comment: all ABNF productions should be enclosed in 
> angle brackets, as recommended by RFC 5234:
>>     Unlike original BNF, angle brackets ("<",">") are not required.
>>     However, angle brackets may be used around a rule name whenever their
>>     presence facilitates in discerning the use of a rule name.

Ah, thanks. This indeed makes it more readable. I'll fix this too.

Freek.