Re: [DNSOP] New draft on representing DNS messages in JSON

[Paul Hoffman <[e-mail]>]

On Aug 21, 2014, at 12:53 PM, Jay Daley <[e-mail]> wrote:

> I agree with all of Shane's points below, namely better encoding of data,

I have responded to this one, assuming you mean RR data.

> not making all fields optional,

I don't think Shane made that argument. To me, it sounded like Shane was fine with fields being optional as long as the records could be defined by a schema.

> handling of unknown rrtypes

Sure, but that's already in the document.

> and having a schema for the data.

He preferred an internal schema; I explained why we couldn't do that with current or soon-to-be-expected standards.

>  These are all points (and many more) that were thoroughly addressed in the following draft on representing DNS in XML
> 
> 	http://tools.ietf.org/html/draft-daley-dnsxml-00

Not to be too picky about it, but your document doesn't describe "DNS in XML", it describes "DNS resource records in XML". You clearly say in section 2.5 that you don't try to cover the header data, which is very important for the use cases for this document.

> If it helps then I can produce XSLT that takes data in the XML format specified in that draft and validated against that schema and turn it into whatever flavour of JSON you require:
> 
> 	http://controlfreak.net/xml-to-json-in-xslt-a-toolkit/
> 
> *stands back and waits for the "but json is sooo much easier than xml" brickbats*

There is that. In fact, one could argue that it would be much simpler to cast your document in JSON, if you didn't need a schema. Nothing is stopping people from adopting yours, although I would have called yours as experimental as mine.

--Paul Hoffman