[rfc-i] No, constraining to a custom SVG profile is not trivial

Julian Reschke julian.reschke at gmx.de
Mon Jan 20 22:04:18 PST 2020

On 20.01.2020 22:30, Brian E Carpenter wrote:
> On 21-Jan-20 08:32, Phillip Hallam-Baker wrote:
>> It is not just the greyscale that is the issue. There are numerous issues in the diagrams that result from the chosen profile.
> No colour or greyscale was a choice, not an issue. Because people wanted both printability and accessibility, the choice was made to get rid of colour-impaired sight problems and cheap printer problems.
> The only other big problem I'm currently aware of is scalability. There are some interactions between browsers and elements like viewBox, width="724.0" and height="485.135549872".

I implemented experimental down-conversion of SVG in
<https://github.com/reschke/xml2rfc/blob/master/prep-xml2rfc.xslt> and
found that the lack of markers prevents the use of even trivial GraphViz

> I also discovered that any any <?xml ...> or <!DOCTYPE ...> declarations must be removed from the SVG file.

That I would call a bug. It's not backed by anything in the specs.

> Hopefully the Temporary RFC Series Project Manager can coordinate some systematic approach to identifying SVG issues, both in tooling for the existing subset and possible updates to the subset. As Leonard Rosenthol hinted, unrestricted SVG is not really an option.

While I agree with that, there are *much* more pressing issues to
address (such as the fact that we publish "canonical" XML which does not
conform to any agreed-upon vocabulary definition).

Best regards, Julian

More information about the rfc-interest mailing list