Re: [Cellar] namespace considerations for ebml's definition of EBML Schema

Hi,

My understanding is that a HTTP URL with meaningful data about the
format is a plus. But it's not mandatory. It has to be distinct enough
from other namespaces existing in the world. Using a urn: would be
less useful IMO.

If we want to use a better URL that people can refer to in the future
then https://datatracker.ietf.org/doc/draft-ietf-cellar-ebml/ would be
the one to use. Or even https://tools.ietf.org/html/rfc8794 when it's
published. I don't know how permanent the URLs are, and we can never
guarantee a URL will be there in 100 years.

I don't know if the namespace we use in the XSD is normative. But even
if it is the value itself doesn't mean much.

If you want a IETF URN rather than the IETF URL for this RFC, then
"urn:ietf:rfc:8794" sounds nice. It matches the ABNF in RFC648.

I searched all (current) RFCs and there are a few "urn:ietf:rfc" in
them, like in RFC 8759:

  The namespace "urn:ietf:rfc:8759" is as defined by [RFC2648].

There's also "urn:ietf:rfc:7807" and "urn:ietf:rfc:7351".

If this normative we may add a sentence similar to RFC8759. Our text
doesn't imply it's normative:

  The following provides an XML Schema [@!XML-SCHEMA] for facilitating
verification of an EBML Schema described in (#ebml-schema).

Le jeu. 16 juil. 2020 à 20:03, Dave Rice <dave@dericed.com> a écrit :
>
> Hi cellar,
> This is a late comment in regards to the EBML document, but I’m considering this dialogue.
>
> Begin forwarded message:
>
>
> From: Karen Moore <kmoore@amsl.com>
> Subject: [AD - Murray] Re: AUTH48 [KC][SW]: RFC 8794 <draft-ietf-cellar-ebml-17.txt> NOW AVAILABLE
> Date: July 16, 2020 at 1:53:35 PM EDT
> To: "Murray S. Kucherawy" <superuser@gmail.com>, Steve Lhomme <slhomme@matroska.org>, Dave Rice <dave@dericed.com>, Moritz Bunkus <moritz@bunkus.org>
> Cc: RFC System <rfc-editor@rfc-editor.org>, Alexey Melnikov <aamelnikov@fastmail.fm>, cellar-ads@ietf.org, cellar-chairs@ietf.org, Steven Villereal <villereal@gmail.com>
>
> 3) In the XML Schema for EBML Schema, the URL https://ietf.org/cellar/ebml
> is provided, but it doesn't seem to be a valid site. Please confirm that this
> is okay.
> -->
>
>
> I don't suppose we could host it on the IETF website. I'm not sure if
> a valid URL is necessary here, this is just a marker for a namespace.
> I think it only needs to be unique so that it's not mistaken with an
> XSD with a same name.
>
>
> Within the EBML Schema example xml and the EBML Schema XSD the value of https://ietf.org/cellar/ebml is used as an undefined and stand-in namespace. From reviewing other RFCs which define XML Schemas, I see that they don’t use this style of URL but instead use a URN guided by RFC2648.
>
> Should we consider adding a section such as the one quoted below and changing from https://ietf.org/cellar/ebml to urn:ietf:params:xml:ns:ebml?
>
> ### `<EBMLSchema>` Namespace
>
> The namespace URI for elements defined by this specification is a URN as
> defined by [@!RFC2141] that uses the namespace identifier 'ietf' defined by
> [@!RFC2648] and extended by [@!RFC3688]. This URN is:
>
> ```
> urn:ietf:params:xml:ns:ebml
>
> ```
>
>
> and changing from https://ietf.org/cellar/ebml to urn:ietf:params:xml:ns:ebml. This would follow the recommendation at https://tools.ietf.org/html/rfc3470#section-4.9.
>
> Dave Rice
> _______________________________________________
> Cellar mailing list
> Cellar@ietf.org
> https://www.ietf.org/mailman/listinfo/cellar

-- 
Steve Lhomme
Matroska association Chairman