[rfc-i] Does the canonical RFC format need to be "readable" by developers and others?

Julian Reschke julian.reschke at gmx.de
Fri Jul 6 12:35:51 PDT 2012

On 2012-07-06 21:17, Iljitsch van Beijnum wrote:
> On 6 Jul 2012, at 21:05 , Julian Reschke wrote:
>> Wow. Yes, XML is non-trivial. But there are tons of (well-tested) parsers, so you don't *need* to write one.
> Then we're beholden to those parsers to power our home grown private tools that scrape RFCs.

"scrape"? Anyway: I fail to see a problem with tools relying on 
well-tested components.

>> Also, saying "XML is useless, because there is nothing that naturally uses XML" is weird; tons of components use XML; I'm not sure about the "naturally" part; could you elaborate on that?
> Browsers display HTML. Applications that generate or display PDF are very common. Line printers print text. But none of them do XML2RFC out of the box. So if you have an XML RFC, you would still need to have a layer of middleware or a conversion tool in order to display or print that RFC. The same for creating it.

Browsers display XML using XSLT, and we have an XSLT-based 
implementation of the xml2rfc vocabulary. Dunno whether this counts as 

And no, you don't need any tools to create the files, beyond a text 
editor. I do it all the time.

> It's like having Russian as the language for EU documents. I'm sure Russian is a great language, but virtually no-one in the EU speaks it so it's just more work for the translators.

Again, every relevant desktop browser understands XML.

>>> There is a reason word processors don't use it.
>> Actually, many do, internally.
> Yes, but what do we care about the internals of word processors? Unless I seriously missed something, the XML formats of Word etc are sufficiently different from XML2RFC that the fact that both are XML-based is of no practical importance.

They are still XML files.

Anyway, I get it: you don't like XML. Do you have a *concrete* proposal 
what to use instead? JSON? Custom syntax? Constrained HTML? Nroff? Yaml? 

Best regards, Julian

More information about the rfc-interest mailing list