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

Julian Reschke julian.reschke at gmx.de
Fri Jan 10 00:53:40 PST 2014


On 2014-01-10 09:03, Riccardo Bernardini wrote:
> On Fri, Jan 10, 2014 at 2:47 AM, Paul Hoffman <paul.hoffman at vpnc.org> wrote:
>> On Jan 9, 2014, at 12:57 PM, Jim Schaad <ietf at augustcellars.com> wrote:
>>
>>> The current layout rules are all based on the assumption that figures are going to be placed in-line in the text and not allowed to float freely and allow for the text to flow around them in those cases where it makes sense.  I have been working on a document where there is a picture and a set of steps.  Allowing the step description to flow around the picture of the steps would probably improve the reading experience for this document.
>>
>> This seems like a minor improvement for a lot of UI pain. The "allow to float" attribute would need to say how far the figure would be able to float; getting the units right on that would be daunting, to say the least.
>
> I would suggest using the LaTeX approach that, in my experience, works
> pretty well (as long as one knows that the float will be moved around;
> float mobility is a major surprise/annoyance for LaTeX beginners).
> LaTeX decides where to place floats (figures and tables, but also
> other "objects" like algorithms, ...) according to an algorithm
> similar to this
>
> *    The float is placed as soon as possible, compatibly with the
> following constraints.
>
> *    Floats are placed at the top or at the bottom of a page or on a
> page that contains only floats.
>
> *    A float cannot be placed before the point it is "declared"
>
> *    Fig. <n> cannot be placed before Fig. <m> if m < n
>
> *    The user can modify this behavior by using some options (e.g.,
> put only on top, only on bottom, put the float exactly here!)
>
> The actual algorithm is a bit more involved and it is described in the
> Leslie's  book.
>
>
>> Can you quantify "probably improve"?
>
> Although this question was directed to OP, let me add my 0.02 Euros.
>
> Clearly, it is difficult to quantify something like this, but I had
> few experiences with documents with non-floating figures and it can be
> quite annoying.  For example, it can happen that a figure is placed
> almost at the bottom of the page with only one or two text lines after
> it.  When you arrive at that page, you read the text before the
> figure, you do not notice the line after it and move to the next page,
> just to find that there is a jump in the sentence.  You go back, go
> forward again, check the page numbers and, finally, you discover the
> line after the figure.    If the document has many figures, this can
> be quite annoying.
>
> Honestly, I do not see any major drawbacks (from user's perspective)
> in having floats, especially if we provide an attribute to say "place
> it here!" for those very very rare cases when it is needed (I write
> scientific papers &stuff with LaTeX since the 90s and I never used
> this option).
>
> I guess that handling floats will make the xml2rfc code a bit more
> complex, but if you think that in LaTeX float handling is written
> using TeX...
>
> Riccardo
> ...

A big -1.

I don't believe we need this complexity. Floating figures are great for 
paginated layouts, but we are moving away from that.

Best regards, Julian


More information about the rfc-interest mailing list