Re: [apps-discuss] An XML RFC primer?

Hi Tom,

On Jan 11, 2011, at 6:08 PM, t.petch wrote:

> ----- Original Message -----
> From: "Murray S. Kucherawy" <msk@cloudmark.com>
> To: <apps-discuss@ietf.org>
> Sent: Monday, January 10, 2011 8:57 PM
> 
> I've got a project on the go that will result in a few different I-Ds for
> consideration, one of which contains a specification for a new XML-based
> reporting format.  Can someone suggest some RFCs that have done a good job of
> the same sort of thing in the past that I could use as a template for such work,
> or a BCP or similar?
> 
> <tp>
> 
> Netconf chose XML and XSD, and probably wishes it hadn't.  I had a quick scan of
> archives and cannot find any decent summary of just why it doesn't work; Randy
> Presuhn is always the most articulate as to just why XML gets it wrong, its lack
> of coherent intellectual underpinnings.  OK, Netconf has a more challenging
> objective, of replacing aspects of SMI and SNMP but even so, some at least of
> the problems are generic, such as
> - lack of proper typing

In fact YANG helps a lot, but yes, types like IP addresses are still not built in.

> - naming and namespaces (eg the no namespace, null namespace, ...)

Well, in hindsight, XML namespaces perhaps could have been done in a better way - James Clark has an interesting account on it in his blog - but namespaces as such are indispensible for NETCONF data modeling. For example, JSON has no namespaces.

> - impossibility of validating documents with XSD (possible with DSDL).

I am not a fan of XSD (as you know) but this statement may be too strong. In the particular case of RFC 4741, the problem is more in the design of the XSD schema than in XSD as such.

> 
> I did a survey of many (most?) IETF XML RFC/I-D before we started, and saw a lot
> of options with no guidance as to which way to go.  (I found some documents that
> were plain wrong).  I raised this lack of guidance on the IETF list but got no
> traction.
> 
> Netconf did take advice from the XML Directorate, and they were plain wrong, in

It seems it was sufficient to carefully read the XML spec.

> one key regard, which has turned out to be an issue that has taken years so far,
> and as yet is unresolved (the question they failed to answer correctly was what
> is and is not valid XML? the consequence is an inability to parse XML documents
> transported by Netconf as currently specified).

This has been recently resolved in draft-ietf-netconf-rfc4742bis-05.

> 
> So, if there anything better, with tools available, I would recommend it

My observation is that quite often problems have been caused by how people (mis)use XML rather than by XML itself.

Lada

> 
> Tom Petch
> </tp>
> 
> 
> Thanks,
> -MSK
> 
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss

--
Ladislav Lhotka, CESNET
PGP Key ID: E74E8C0C