diacritical (caret) cropped when equation shrinkwraps (FM9)

Report · May 17, 2011

When I use some diacriticals in equations, such as n with a caret over it, the top of the diacritical gets cropped when I shrinkwrap the equation. It does not show in the document, does not print, and does not show when converted to pdf. Using micropositioning to move the diacritical and the letter closer together does not fix this because micropositioning leads to the anchored frame containing the equation being resized also, still cropping the top of the diacritical. The only "solution" I have found is to put a blank superscript by the letter (i.e. click one of the superscript buttons in the equation palette but then not enter anything for the superscript. Is there a better way to solve this?

Report · May 17, 2011

V. interesting workaround with the blank superscript character.

Some troubleshooting questions:

What specific version as shown in Help > About?
What O/S and version;

What specific font(s) are you using? Have you changed equation fonts from FM's default, or applied any character formats?

Does the problem occur with all equation sizes?

What printer do you have set as your default printer when the equations are created and when working in FM in general.

Could you post a screenshot of a problem equation?

Just for clarification, do the characters show correctly when you unwrap the equation (?) and then get cut off when it's wrapped again?

http://help.adobe.com/en_US/FrameMaker/9.0/Using/WS6741C353-0FB0-47a4-90B0-591700106C91.html

Report · May 18, 2011

Thanks for your reply. My responses are below. I also got a suggestion

from Johannes Krueger (who had the same problem in FM8p277 on Win XP),

that in Equation Sizes I reduce the vertical spread to -5%. I haven't

tried this, and am concerned it will make other parts of equations look

wrong.

Andy Taylor

On 5/17/2011 6:12 PM, Sheila Carlisle wrote:

V. interesting workaround with the blank superscript character.

Unfortunately it doesn't always work, and even when it does it often adds unwanted spacing.

Some troubleshooting questions:
What specific version as shown in Help> About?

9.0p237

What O/S and version;

Windows 7 Pro

What specific font(s) are you using? Have you changed equation fonts from FM's default, or applied any character formats?

ordinary text is Times Roman (12 pt). Equation fonts are default: math font = Symbol, Equation Variables = "as is". No character formats in these equations; I do use Symbol font within the text some.

Does the problem occur with all equation sizes?

Yes.

What printer do you have set as your default printer when the equations are created and when working in FM in general.

System default is an HP LaserJet 1320N but I use Print Setup in FM to switch to Adobe PDF. It doesn't seem to make a difference whether I create equations before or after changing the printer.

Could you post a screenshot of a problem equation?

Don't know whether this will be visible, but it's a png file of a screenshot showing both the document open in FM, and a pdf created from it.

Just for clarification, do the characters show correctly when you unwrap the equation (?) and then get cut off when it's wrapped again?

This is interesting, actually. Often when first creating the equation the caret is cut off even in the unwrapped view, as well as after wrapping it. But then if I unwrap it the caret shows properly.

Message was edited by: Andrew D Taylor -- to add the response to Sheila's question, which did not get posted the first time

Report · May 20, 2011

Yikes, your 9.0p237 needs to be updated, there are several significant changes since that release. Either through FM's Help menu, or download manually from:

Updates - Adobe - FrameMaker : For Windows

FM updates are not roll-up, so they need to be applied individually, in sequence.

Report · May 18, 2011

I see the same problem in FM8p277 on Windows XP, starting from blank paper: The shrink-wrapped frame crops off the upper half of the hat, while other diacritics like tilde, vector or arc remain fully visible when shrink-wrapped.

You can try the following:

In the Equations palette, go to Equations > Equation Sizes. Reduce the vertical spread to -5,0% and click set.

You'll have to check though whether this change has unwanted effects on other math elements.

Johannes

Report · May 18, 2011

This does seem to solve the problem. Since I usually only use medium equations, I'm going to redefine the "large" equation setting to the same sizes as "medium" but with this reduced vertical spread. This way for equations without carets I can use the standard "medium" settings and not have to worry about unwanted effects of the reduced vertical spread.

Report · May 20, 2011

hmmm, I'm not a math whiz by any means, but I just found this blurb about the Combining Circumflex Accent which is unicode 0302, aka "hat" that is used for math equations, and it appears that the character isn't a "pointy hat" character (ewww, couldn't resist that) to start with, it does have the top point cut off, so maybe it is actually being displayed correctly.

Here's a link that shows a graphic of the character:

http://www.fileformat.info/info/unicode/char/0302/index.htm

http://www.unicode.org/charts/PDF/U0300.pdf

The regular caret and circumflex characters that are for text, not math, are pointy:

http://en.wikipedia.org/wiki/Caret

http://en.wikipedia.org/wiki/Circumflex

Report · May 20, 2011

No, the math symbol is the same as the combining circumflex accent. I

don't know why the one in the first link you sent has a flat top; the

unicode chart shows a pointy top, as does the Wikipedia article (which

does mention the use of this symbol in math). And even the flat topped

one in the 'fileformat.info' link has the two upstrokes touching. When

FM crops the symbol, they don't touch. And when I do the -5% vertical

spread fix that Johannes told me, as well as when I unwrap the equation,

it shows the full pointy hat, so that clearly is what FM means it to be.

I did the updates you mentioned in your other reply; they had no effect

on this problem but I'm sure otherwise are worthwhile. I thought FM was

automatically checking for updates, but obviously not.

Report · May 20, 2011

Sorry, which unicode chart do you mean shows a pointy top? The PDF I referenced above shows this:

And here's another page:

http://unicode.org/faq/char_combmark.html#12b

This is what I get from Word 2010:

Report · May 20, 2011

You're right. I was referring to

http://www.unicode.org/charts/PDF/U0300.pdf which you sent me, and what

I was looking at in that was actually for 005E circumflex accent, rather

than 0302 combining circumflex accent. And even the 005E accent, if

looked at closely enough, has a flat top. I think how much the top is

flattened actually is a font design choice.

But as I noted before, all of these with a flat top have the two parts

firmly connected, which the cropped form in FM does not have. And when

it is not cropped (with Johannes' fix or because of large elements in

the equation which give a larger wrapped size), it has a flat top as in

the various examples you found, but still with the parts connected, not

like the cropped version.

The issue really is that with default equation sizing, the cropping

causes the character to not be recognizable as a caret.

Report · May 20, 2011

I see what you're saying about the two sides not touching when the equation is wrapped.

A different (incredibly tedious) workaround would be to highlight the equation's anchored frame, do Select All (ctl+a) to select the equation, then alt_gp to view the Object Properties. Change the Alignment to "Manual" and then change the Top alignment to make the object sit just a little bit lower in the frame.

In my test file the object is created as Top=.217", so changing it to .25" moves it down just enough to show the two lines meeting. Using any number less than .25" (e.g. .223) doesn't work, the object doesn't move and the value stays at .217.

If this seems like a good solution it would likely be possible to have a Framescript written to automatically adjust all equations in the doc, and the script could be incorporated into the menu bar and even assigned a hotkey.

On my systems (XP and Win2008Server) with FM's default symbols and equation characters this still is showing the flat-top character, though, not a pointed one, I've never seen a pointed one in these tests in FM, at least not when working at a large zoom size. I'm constantly doing Ctl+L (actually a lower-case "ell") to refresh the display, too.

About the display in Acrobat, if you zoom up/down in size do you see the pointy vs flat-top appear and disappear? That's what I see, which I think the same effect as a long-standing issue with Acrobat not being able to smoothly zoom between some display percentages. People often see this effect when horizontal lines in a table appear to be randomly thick or thin.

SHeila

edit control+ell keyboard shortcut: Sheila Carlisle

Report · May 23, 2011

I tried your suggestion, increasing the Top= alignment setting in Object

Properties. It does make the entire caret visible, but it moves the

entire equation too far below the text line. So for now I'm going to

use Johannes' fix with the negative vertical spread.

It's interesting, and maybe helpful for finding a proper fix, that this

problem doesn't occur with other combining diacriticals, some of which

actually rise higher than the caret. It seems that somehow the

algorithm used to determine the frame size for the shrink-wrapped

equation isn't getting adequate information about the size/location of

the caret, but does for the other diacriticals.

Report · May 23, 2011

Here's a different workaround:

In the equation frame, before it's shrinkwrapped, draw a horizontal graphic line of, say, 1pt. just above the top of the diacritical. Make it shorter than the equation. Shrinkwrap the equation. Delete the horizontal line.

edit: or make the line white, so you might not have to delete it (although personally I would hate having random white lines floating around in a frame, it would make me crazy).

Report · May 23, 2011

I vaguely recall this as being an issue for many FM versions, that's what led me to look at the caret to check the pointy-ness factor.

Rather than FM not making the shrinkwrap correctly, I think it's more an issue with where the caret is located within the bounding box of the glyph in the font.

You may find a better result by trying mathematical fonts, perhaps Design Science has some glyph pages that would show a better position for the caret.

http://www.dessci.com/en/dl/fonts/?src=hpfnt.

Report · May 23, 2011

You're suggesting the problem is in the design/specification of the font

and that trying other fonts might give a better solution. This makes

sense to me. But how do I change the font? For "Math Symbols" on

"Equation Fonts" the only choice is Symbol, and it is unaffected if I

redefine a "Symbol" font in the character designer.

I don't think your other suggestion, that FM has intentionally moved up

a pointy-top caret so that it will be truncated, is right. When the

equation is unwrapped, or when Johannes' fix with negative vertical

spread is used, the caret shows as flat-topped, so there'd be no need

for such a weird fix, the character itself is OK.

Report · May 23, 2011

Unfortunately I'm getting on thin ice about changing the equation font. I really hope Ken Benson might see this thread, he's always been a good source of info on FM and math.

I just noticed in the file \configui\mathcmds.cfg there's a section on diacriticals with this:

Not sure where this could go, but as a test you might try experimenting with the number (with a backup copy of mathcmds safely stored, of course!) and seeing what impact it might have on the equation editor -- if you change the number to, say, 1501, and a Tilde shows up when you specify the caret diacritical in the equations palette then at least you'd know that the code point numbers in mathcmds do directly influence the equation.

drat, I just tested this, doesn't seem to make a difference.

Report · May 23, 2011

You've probably already seen this, but

http://help.adobe.com/en_US/FrameMaker/9.0/Using/WS6741C353-0FB0-47a4-90B0-591700106C91.html

Report · May 23, 2011

I had seen it, but had missed the detail about changing the font for

math symbols. Apparently only certain fonts can be used. Unfortunately

it seems the only one I have that can be used is Symbol. I may try one

of the others (Mathematical Pi or Universal Greek) if I can find a free

trial, but even if it works I'm not keen to have to buy a font just to

fix this.

Report · May 23, 2011

Another blurb about editing the math fonts:

http://kb2.adobe.com/cps/321/321244.html

Report · May 23, 2011

>> It's interesting, and maybe helpful for finding a proper fix, that this

problem doesn't occur with other combining diacriticals, some of which

actually rise higher than the caret.

Ah, the penny just dropped. I wonder if FM is doing some font skullduggery, creating the cropped appearance by using a pointy-hat character and moving it up by using an internal character adjustment (analagous to the GUI Alt+arrow micropositioning) to touch the bounding box and thus appear as a "correctly" cropped caret.

This could be an issue if FM treats the caret as a reserved character, as FM does with several other code points, and this was the only way for the equation editor get the mathematical caret -- I'll bet the equation editor hasn't changed since about 1994.

btw, babelMap is an excellent free unicode font viewer. http://www.babelstone.co.uk/Software/BabelMap.html

It's interesting to see where the caret is located in different fonts, and in fact whether it is pointy or not.