[rfc-i] What the text version is used for (was Re: The <tt> train wreck)

Paul Kyzivat pkyzivat at alum.mit.edu
Sat Aug 21 20:00:58 PDT 2021


On 8/21/21 7:26 PM, Brian E Carpenter wrote:
> On 22-Aug-21 10:27, John Levine wrote:
>> It appears that Robert Sparks  <rjsparks at nostrum.com> said:
>>> There have been several times we have had long threads where people have
>>> told us what they use the text formats for. They've been compelling, at
>>> least to me, for helping ensure the format continues to be produced.
>> Early on I thought the text version was a barbarous relic, but as I
>> talked to more people I see that it remains useful mostly because we
>> (for a very broad version of we, not limited to the IETF) do not have
>> adequate tools for using non-text versions. We don't have xmldiff
> Have you tried rfcdiff on xml2rfc files? It works remarkably well
> except for needing to wrap long lines. That could be fixed. But of
> course what we really need is a diff between *rendered* html
> files, and that is science fiction afaik.
>> or
>> xml or html grep. Several people have said that when they're
>> commenting on a draft or I-D, they copy and paste rectangular regions
>> of text, which is generally impossible in HTML in a browser. (You can
>> copy, but try to select a rectangle.)
> Huh? I don't know how to select a rectangle from a .txt file which
> is not full width. Also, the cases where you actually *need* to
> preserve the layout will be artwork or code. You can select those
> from the rendered HTML with no problem.

I'm one of those people. Call me lazy, but here are a couple of use cases:

* As you mention, rfcdiff of *rendered* text. The side by side format is 
very convenient for tracking the changes in a doc I follow and don't 
author. A diff of the xml would be next to useless to me.

Sometimes the side by side diff doesn't show enough context.In that case 
it is trivial to find the same place in the .txt.

Once I see something in the change that I want to comment on, I can copy 
a block of consecutive lines from the .txt and paste them into an email 
to give context to a comment I want to make. The copied content is 
conveniently indented to set it off from my comments. From that email 
others can easily find the same place in the .txt.

If there were convenient tools to do the same using html renderings then 
I might no longer be attached to the .txt.


More information about the rfc-interest mailing list