Automatic figure placement and figure styles

[DylanMuir]

<start latex war>
I don't like LaTeX. I want to be able to set up a bunch of rules on how to style a document, then just type and see my document as it follows those rules. Mellel does this admirably for text (as do some other word processors). I don't want to have to write code, then compile my document (twice!), then spend a day searching Google to see if someone's made a package to fix some weird formatting condition, then compile my document (twice!) again.
</end latex war>

But there's one (particular) thing which LaTeX does well, and which Mellel (or any other word processor) does not; namely automatic figure placement. I know there's a way to go yet until Mellel figure handling is perfect (wrapping, etc.), but as an academic writer (natural sciences) I'd like to be able to use Mellel and throw everything else away.

So here's my suggestion: Figure styles (similar to OpenOffice) and automatic figure placement.

I'd like to be able to define a "full page width figure" style for example, and set options such as "only at the top or bottom of a page". Or define a "half page width figure" style and set that the figure should appear on the inside of a page, with the caption on the outside. Or a "full page figure" style.

I could import images into the document and assign them a figure style, without specifying their exact position in the text. Then I'd have a pane showing a list of figures referred to in my document, and be able to insert figure references in the text. Figures would be automatically placed in the text somewhere near the first reference (this would of course be configurable in the figure styles).

I'm certain this is possible; LaTeX has a set of algorithms for figure layout which does exactly this. As far as I'm aware this feature is not available in any other word processing system.

Of course in-line placement should still be available, as well as the possibility to force figure placement to a particular location. I understand this is a major feature, and would take quite some implementation effort... I think if it were included in Mellel, every academic writer using a mac would switch at once...

D.

[DylanMuir]

What happened to the flame war which should have erupted over this suggestion? I thought at least someone else would have suggestions for improvement of figure handling in Mellel...

[Reiner]

Sounds very reasonable from my point of view. But I would add one more feature to complete this:

If image/figure-styles would also enable auto-title-attaching this would be great:
- Attach image-auto-title to figure (or not)
- auto-title before image
- auto-title after image
- keep auto-title and image on same page/in same column (or not)

[joewiz]

Figures would be automatically placed in the text somewhere near the first reference (this would of course be configurable in the figure styles).

Very cool idea. This sentence, in particular, makes it clear that the system would be tied in with the cross-reference system which is somewhere in Mellel's future. Reiner's auto-title suggestion also implies cross-referencing, if I understand correctly. Is there anything else that needs to be in place for any elements of this plan to be adapted in Mellel? Image wrap-around? Additional image options? It also shares a lot with the 'Marginalia' discussion of some time ago. In addition to 'figure styles', would you want overrides to the styles for custom placement and options?

I bring these questions and points up in order to consider where this idea might fit into Mellel's overall development. Figure styles are a neat idea, but are there prerequisites? Or would it be possible for core elements to be introduced now before other capabilities are added to Mellel?

[DylanMuir]

The way I imagine the system working would require captions to stay with figures, and text wrapping (at least at the top and bottom of a figure/caption block). I guess that implies some sort of "frame" system; text would wrap around frames, and frames would enclose a figure and its caption.

I think the auto-titling that Reiner mentions is exactly the captioning which I already use within Mellel, and the captions which I imagine would work with this figure system.

References are an obvious pre-requisite, but as I understand that's in the pipeline.

The ability to force a figure to a particular place seems like a simple extension of this system once the machinery is in place. One could say "this figure must appear in this position on this page" and everything else would wrap around it. Or a page could be dedicated to a figure, and the rest of the document would flow over that page. This is very different to the current "locked to a paragraph" behaviour of other word processors, or to the "as a character" behaviour that Mellel has currently.

D.

[cyberbryce]

The way I imagine the system working would require captions to stay with figures, and text wrapping (at least at the top and bottom of a figure/caption block). I guess that implies some sort of "frame" system; text would wrap around frames, and frames would enclose a figure and its caption.

I think the auto-titling that Reiner mentions is exactly the captioning which I already use within Mellel, and the captions which I imagine would work with this figure system.

References are an obvious pre-requisite, but as I understand that's in the pipeline.

The ability to force a figure to a particular place seems like a simple extension of this system once the machinery is in place. One could say "this figure must appear in this position on this page" and everything else would wrap around it. Or a page could be dedicated to a figure, and the rest of the document would flow over that page. This is very different to the current "locked to a paragraph" behaviour of other word processors, or to the "as a character" behaviour that Mellel has currently.

I agree, I'd re-cast my vote for simple frame-related functionality (no need for nesting or arbitrary page-layout). It seems like frames with the placement and text-flow functionality DylanMuir mentions would suffice for proper figure styling and placement. Per Reiner's suggestion, the frame-style could specify that an auto-title is automatically inserted at the beginning or the end of the resulting frame when you apply the style.

In terms of integrating with existing functionality, would it also suffice to allow figures to be "mentioned", or does it have to involve cross-referencing?

--Bryce

[DylanMuir]

I must admit that my understanding of the Mellel mentions systems is a touch fuzzy, but I think that a mention can
a) only contain a copy of the text of the mentioned auto-title and
b) changes its content to reflect the current auto-title of whatever section it's in.

A reference, on the other hand, has to refer to a specific object, and should have options such as displaying "above"/"below" or just a figure number instead of only "Figure x".

Maybe someone else has a better idea of how this could mesh with the current mentions system.

D.

[nicka]

And mentions can only be in headers, not the body text.

[joewiz]

I must admit that my understanding of the Mellel mentions systems is a touch fuzzy, but I think that a mention can
a) only contain a copy of the text of the mentioned auto-title and
b) changes its content to reflect the current auto-title of whatever section it's in.

A reference, on the other hand, has to refer to a specific object, and should have options such as displaying "above"/"below" or just a figure number instead of only "Figure x".

I think the chief obstacle to Mentions being used is that, as the Guide says, they are "can be created only in the header and the footer of a page, one mention per paragraph" (p. 280). Furthermore, Mentions can also only include the following elements: the text itself of the auto-title to which the mention refers, the number to which its auto-title corresponds, the current page number, free text and tabs and line breaks, or a document variable.

So the restriction of Mentions being stuck in header/footer and kept to one-per-paragraph would have to be done away with in order to encompass the auto-title ideas proposed in this thread. The auto-title could be placed below the figure and include the title of the figure; and the mention could be placed in the text, in the form: "See Figure (#)" or "See Figure (#), '(Title)', on page (#)."

As you observed, though, Mentions currently can't say "See Figure (#) (above/below)", or "See Figure (#) on (previous/next/opposite page)", or "See Figure (#) in Chapter (#), page (#)." Thus, there is room to expand the options for Mentions to handle your ideas.

Or perhaps the Redlers will decide to keep Mentions as they are -- and address this need via the cross-reference system?

[DylanMuir]

If any of the Redlers (or other assorted developers) are reading this, it would be nice to have some idea of what you think -- is the proposed system possible, desirable, etc? Or do you think it's a stupid idea and I should stop hoping? ;)

D.

[Ori Redler]

If any of the Redlers (or other assorted developers) are reading this, it would be nice to have some idea of what you think -- is the proposed system possible, desirable, etc? Or do you think it's a stupid idea and I should stop hoping?

D.

Reading and liking it. It's a cool ideas... As for when we'll be able to do it -- that's a different opera, but I've added this to our suggestions/to-do list.

Ori