ISSUE: character-height is wrong. [was: Re[2]: [cgmo-webcgm] QUESTION about text-font SP]

[Lofton Henderson <lofton@rockynet.com>]

At 04:41 PM 7/8/2005 -0400, Benoit Bezaire wrote:
>[...]
>But I agree with Dave, this is not a 'text-font' issue but a
>'character-height' issue.

I disagree that it is a character-height issue.  (As I pointed out in 
previous message, changing text-font will almost always lead to a mismatch 
between the font and the restriction boxes in the metafile, sometimes to 
dramatically bad effect.)

However, there *is* a serious issue about character-height, and I think I 
finally understand it.  Therefore, I'm going to put aside (for now) arguing 
about whether my original question is a restriction box or character-height 
issue.

Here is the problem with character-height.  The current definition, which 
is more or less what has been around since Jan-Feb drafts, says:

>character-height is the cap height of a character (the distance from the 
>baseline position to capline). Relative values of this this style property 
>value will increase or decrease the metafiled-defined value of the 
>CHARACTER HEIGHT attributes in the target object (APS or picture) by the 
>given "%" value; absolute values (NVDC) will define a new character height.

Okay, so suppose I have a metafile with this stuff:

BegAps 'myRTexample' 'para' StList;
BegApsBody;
   RestrTextType boxed-cap;
   CharHeight 50;
   RestrText 250 50 (5000,5000) "ABCDE";
EndAps;

Question.  What would be the effect of this SP setting?

...apsid.setStyleProperty('character-height', "100");
or equivalently,
...apsid.setStyleProperty('character-height', "200%");

Answer.  According to ISO CGM:1999 ... NOTHING!

Reason.  We have defined that character-height redefines the metafile 
CHARACTER HEIGHT attribute value.  According to CGM:1999, the restricted 
text parameters for boxed-cap (which is by far the most common Restricted 
Text Type, IMO) override the CHARACTER HEIGHT to achieve an exact vertical 
fit of the baseline-to-capline and the height of the restriction 
box).  This is true for 5 out of the 6 defined Restricted Text Type values 
(only 'justified' is bound to CHARACTER HEIGHT).

The Style Property character-height, as we have defined it, does not meet 
the need that we seem to have defined for a text-resizing Style Property.

PROPOSAL:

1.) Replace character-height with text-resize (or text-size, I don't care 
which).  Details...

2.) If text-resize is "%", then it adjusts the text restriction boxes 
(heights and widths) by that amount.  This would give the exact result that 
I would expect for the very common left-base alignment, RTT boxed-cap.  (As 
well as for the other alignments, with the possible exception of 'justified').

3.) Do we still want to allow NVDC?  Benoit questioned it, and it does 
complicate things without adding much useful functionality.  If we do allow 
it, and if the text-resize value is NVDC, do something like 
this...  Compute the ratio (effectively it's a percent) of the new NVDC 
value and the restriction box(es) height(s) in the target object.  Then 
resize the restriction box(es) as in #2.  With RTT boxed-cap, this leads to 
the resized text being the height specified by the new NVDC value.

4.) What about 'justified'?  I'm not sure.  You could say that the 
text-resize values affect both the restriction boxes, and the CHARACTER 
HEIGHT attribute values as well.  That would take care of 'justified'.

IMO, this proposal for text gives us the same sort of simple style changes 
that are within the scope of Style Properties, similar to resetting line 
widths etc.

>I don't see why changing the font should
>have an impact on the restricted text box.

See previous message...

-Lofton.

>--
>  Benoit   mailto:benoit@itedo.com
>
>
>Friday, July 8, 2005, 4:28:33 PM, David wrote:
>
>CDW> This is the same question, I remember  hearing from Franck
>CDW> and Benoit relating the changing the character height, also.  Are
>CDW> we saying anything in that area about what happens to the
>CDW> restricted text box?  I don't think we are addressing it there
>CDW> after a quick look.  Should we?
>CDW>
>CDW> thx...Dave
>
>CDW> -----Original Message-----
>CDW> From: Lofton Henderson   [mailto:lofton@rockynet.com]
>CDW> Sent: Friday, July 08, 2005 9:18   AM
>CDW> To: cgmo-webcgm@lists.oasis-open.org
>CDW> Subject: [cgmo-webcgm] QUESTION about text-font SP
>
>
>
>CDW> About the   text-font Style Property issue,
>
>
>CDW> ISSUE: text-font Style
>CDW> 
>Propertyhttp://www.oasis-open.org/apps/org/workgroup/cgmo-webcgm/email/archives/200506/msg00165.html
>CDW> This     is Dieter's proposal for the text-font Style
>CDW> Property, and discussion thread     - discuss
>
>
>CDW> the Wednesday telecon decided to put include the
>CDW> text-font Style Property in the next Editors Draft.  Here are the
>CDW> words   that I put in:
>
>
>CDW> text-font: Specifies a replacement     font for all text in
>CDW> the target object. If the characters that are needed     for all
>CDW> text in the target object are available in the specified
>CDW> replacement     font, and if the specified font is available,
>CDW> then use it for all text in     the target object. Otherwise,
>CDW> ignore the specified font.
>
>
>CDW> QUESTION.  Do we intend to say anything about the boxes
>CDW> of the Restricted Text elements in the case that font replacement
>CDW> happens?  Or do we just keep silent about it?  (This is one place
>CDW> where the can of worms starts to open up.)
>
>CDW> -Lofton.