[rfc-i] Should Figures be allowed to float?

Paul Hoffman paul.hoffman at vpnc.org
Fri Jan 10 12:10:03 PST 2014


On Jan 10, 2014, at 11:10 AM, Jim Schaad <ietf at augustcellars.com> wrote:

> Let me be clear, when I am talking about floating, I am not talking about
> figures that got to the top or the bottom of a page.  The current move, for
> better or worse, is to move away from paged formats so that is not something
> that makes sense in the current context.  Instead I am talking about figures
> which are not full width of the display page and can move up and down next
> to and around text so that the display is both more dense and so that
> pictures and text can appear alongside of each other and not stacked.

Thanks for the clarification. I'm pretty sure you mean "be wrapped", not "float".

Given the narrow margins of the current format, this seems like an edge case for v2. Can you clarify if you are proposing this for v2 or v3? Even if it is for v3, the floating would not work for the text-only non-canonical output because it make it harder to read.

> To do so does require that that the figure know where in the resulting
> document that it needs to go, but that can be relatively easy to cover in a
> number of ways.
> 
> A figure would be restricted to a paragraph, section or list by default
> because that is what the containing element of the figure would be.  It
> might be necessary to allow for a container object to be created that has
> content but no display area to allow for a figure and an amount of text to
> be associated together. But that can generally have some type of generic
> association by distance.   That is a figure that is tagged "floating, right,
> center" could be placed at its current location in the text, but be placed
> on the left margin.  The preceeding and next t element would then be
> centered on the figure with wrapping text if that is reasonable, and shoved
> to the top and the bottom if the display area is too narrow.
> 
> I don't think that there is a huge amount of logic that needs to be created
> to deal with this.

Possibly true, now that we are talking about wrapping text from the current container (and possibly the following container). However, I'm still not seeing much benefit, but a lot of downside for restricted-width screen.

--Paul Hoffman


More information about the rfc-interest mailing list