[xml2rfc] Re: Is xref working as it should?

nobody at xyzzy.claranet.de (Frank Ellermann) Thu, 03 July 2008 03:20 UTC

From: "nobody at xyzzy.claranet.de"
Date: Thu, 03 Jul 2008 03:20:19 +0000
Subject: [xml2rfc] Re: Is xref working as it should?
References: <g4i0m4$fl1$1@ger.gmane.org> <486C8AFD.9000706@gmx.de><g4i5fn$15m$1@ger.gmane.org> <486C9B88.7070102@gmx.de>
Message-ID: <g4i94f$e6f$1@ger.gmane.org>
X-Date: Thu Jul 3 03:20:19 2008

Julian Reschke wrote:
  
>>> You would need to say
>>>    <xref target="RFC5226"/> section 4.1
 
>> Yes, I have that now, but that leaves the
>> semantics to best guesses of rfcmarkup.

> Well, that one needs to guess anyway, as it
> only has the TXT format as input, right?

At the moment, yes, and it gets it right for
[RFCxxxx] section y.z

I vaguely recall that it sometimes links the
section y.z to a section y.z in the same RFC,
no matter if it exists or not, for plausible 
variations of "y.z in RFCxxxx".  Test:

http://tools.ietf.org/rfcmarkup?url=http://hmdmhdfmhdjmzdtjmzdtzktdkztdjz.googlepages.com/draft-ellermann-idnabis-test-tlds-08.txt#section-5

The <xref target="RFC2860">section 4.3</xref>
text output won't work with rfcmarkup.  That
is why I asked, I think the output could be
improved for RFC targets.

> The raison d'?tre of rfc2629.xslt is to 
> avoid the HTML output mode of xml2rfc :-)

Testing the HTML output with a modern browser,
it's okay.  With an older browser it wasn't, I
guess the XSLT output also won't pass with an
older browser (IOW without CSS).

> It's just that many authors are sloppy in 
> quoting other specs, making it for the readers
> harder to find out what's being referenced
> than it needs to be.

ACK, and that could be automatically tested if
<xref> would work as I expect it.  In theory -
I hope a future rfcmarkup variant will accept
the XML source as input. 

 Frank