Re: [rfc-i] Unicode in xml2rfc v3

[Henrik Levkowetz <[e-mail]>]

Hi Lars,

The way to go, given the current limitations, is to provide the equations
within <artwork>.  This would work both for unicode text artwork (with the
limitation that it will use a monospaced font) and svg artwork, if you
can produce that from the math formula input.

I have an idea of how to improve math support substantially, and I strongly
believe it's needed and appropriate, but for now the approach above is what
is available.

Looking to the future, I also mentioned Math support in the new work section
of my implementation notes:
https://tools.ietf.org/html/draft-levkowetz-xml2rfc-v3-implementation-notes-11#section-5.1


Best regards,

	Henrik

On 2020-12-01 18:35, Lars Eggert wrote:
> Hi,
> 
> we're at the moment revising RFC8312, the CUBIC congestion controller for TCP. 
> There is quite a bit of math in RFC8312, and even more in the papers that 
> describe CUBIC, some of which we're planning to roll into RFC8312bis.
> 
> That math uses a bunch of greek letters, which we needed to ASCIIfy for RFC8312. 
> One thing I was planning to do for the bis was to reduce the difference between 
> the paper and the RFC, by using the greek letters in the RFC. Since, you know, 
> xml2rfc v3 is supposed to handle Unicode.
> 
> Except, as most of you probably know but I didn't, you can't just use Unicode. 
> You need to wrap it in a <u> tag, and it must be rendered with "num" in the format.
> 
> Which means that while I can write a formula in kramdown-rfc2629:
> 
> ~~~ math
> W_{est} = W_{est} + α_{aimd} * \frac{segments_{acked}}{cwnd}
> ~~~
> 
> and have it automatically rendered in SVG as (via tex2svg):
> 
> and in ASCII (via asciitex) as
> 
>                         segments
>                                 acked
> W    = W    + α     * -------------
>   est    est    aimd        cwnd
> 
> but I cannot use "α<sub>aimd</sub>" in the text of the RFC - it gets rendered as 
> "&#945;_(aimd)".
> 
> And using "<u>α</u><sub>aimd</sub>" gets rendered as the even uglier ""α" (GREEK 
> SMALL LETTER ALPHA, U+03B1)_(aimd)". I can of course play with the format 
> string, but since there is no way to not at least use "char", I have basically 
> no option here.
> 
> I had really been hoping that v3 would enable better math. Unicode is not just 
> for names...
> 
> Thanks,
> Lars
> 
> 
> 
> _______________________________________________
> rfc-interest mailing list
> [e-mail]
> https://www.rfc-editor.org/mailman/listinfo/rfc-interest
> 

_______________________________________________
rfc-interest mailing list
[e-mail]
https://www.rfc-editor.org/mailman/listinfo/rfc-interest