Re: [xml2rfc] RfcMarkup not authoritative, HTML is gone?

[Henrik Levkowetz <[e-mail]>]

Hi Paul,

On 2019-09-16 17:32, Paul Kyzivat wrote:
> On 9/16/19 10:46 AM, Anders Rundgren wrote:
>> On 2019-09-16 16:27, Paul Kyzivat wrote:
>>> On 9/16/19 1:22 AM, Anders Rundgren wrote:
>>>> On 2019-09-14 11:41, Julian Reschke wrote:
>>>> <snip>
>>>>>
>>>>> RfcMarkup has no official standing (and yes, I like it as well, and it
>>>>> has served us well for a very long time). It apparently generates 
>>>>> XHTML,
>>>>> but the content is served as text/html on tools.ietf.org. It might be
>>>>> good to change it to produce valid HTML5.
>>>>
>>>> This in interesting and but also rather confusing since this is the by
>>>> far most used method for communicating RFCs by developers.
>>>
>>> Do you have any data to backup this claim?
>> 
>> No, OTOH, since I mostly work with JSON-based stuff, the documents are 
>> fairly recent.
> 
> I don't recall every having seen a reference to an RfcMarkup.

rfcmarkup is the script that does the htmlization of text drafts and
rfcs.  I've pulled the interesting bits out as a python lib,
https://pypi.org/project/rfc2html/ .

>> What I'm sure that I have newer seen in wild are references to PDF RFCs 
> 
> Nor have I.
> 
> The most common references I see are to either the "HTMLized" form, the 
> plain text, or the datatracker page.
> 
> Personally I much prefer to work with the HTMLized form. It gives easy 
> ability to jump around in the document, the extra info in the header 
> including links to related tools, etc. and yet the text is almost 
> identical to the plain text form which makes it easy to jump back and 
> forth between a diff and the full document. It is also convenient to cut 
> and paste bits of the document into email. I much prefer this format to 
> the HTML format generated directly from xml.
> 
> When a new version of a document is published, I first pull up the 
> HTMLized format. Then from there I follow the link to open a diff in 
> another tab. That way I can skim the changes and easily refer back to 
> the full document when I need more context.
> 
> I hope that the HTMLized format will continue to be available.

That is my plan, yes.  I see no reason to stop the scripts that does that
job.

> It could 
> however be directly generated from the XML rather than from the 
> plaintext. Generating it from the plaintext sometimes results in defects 
> regarding references.

Yes.  When I can find the time, I thought I'd at least try to do a CSS
stylesheet that applies a style to the v3 html that makes it come out
very close to the htmlized version.

Generating html which has *exactly* the same line breaks as the text output
will require writing a almost completely new formatter that does the text
format line breaking, but inserts the proper links everywhere the regular
HTML v3 renderer does so.  I don't know when, if ever, I'd have that much
time to spare.


Best regards,

	Henrik