Re: Could someone please clarify the spec for alignment-adjust

Dear Dave,


Am 07.07.2007 um 07:56 schrieb Dave Pawson:
> Please don't post to multiple lists at the same time.

Reason for the cross-post was that I first thought this was a bug in  
FOP, and then realized this is a bug in the spec.

>> I am now a little confused about the spec on alignment-adjust.  My
>> Reference is xsl-fo 1.1 [1]
>> 7.31.22 "vertical-align"
>
> Something wrong.
> http://www.w3.org/TR/xsl11/#alignment-adjust
> 7.14.1
>
> or
>
> http://www.w3.org/TR/xsl11/#vertical-align
> 7.31.22 "vertical-align"
>
> I'll assume the latter.
>> [...] shift-direction to bottom-to-top
>>
>> Which together results in: a positive value will shift towards bottom
>> (lower the box), while a negative value shifts towards top (raise the
>> box),which is contradictory to the specification before.
>>
>> I may have missed something, but I believe this is contradictory.
>> Could someone please clarify on this?
>
> All relative to the baseline, see
> http://www.w3.org/TR/xsl11/#area-alignment  the figure?
> The alphabetic baseline is the horizontal reference point for
> vertical placement of a glyph.
> To tweak this, we shift this
> http://www.w3.org/TR/xsl11/#baseline-shift
> (suggestion is for graphics etc)
> The assumed 'positive' direction (shift direction) is pos = up for
> Western scripts.
> That makes superscript move nearer the top.
> Hence setting this inverted might make sense.

Assuming Western shift direction (bottom-to-top) alignment-adjust is  
specified to work in the opposite direction as baseline-shift. This  
is fine, as long as one knows about it, and first I've used baseline- 
shift without any problems.

However, the "graphic" I want to align in this case is a mathematical  
formula, which does have an "internal textual structure". According  
to the note in <length> in 7.14.3 I am supposed to use alignment- 
adjust, so I tried doing so and ran into this problem.  One advantage  
of using alignment-adjust is that I can specify the baseline in terms  
of percent of the inline-object instead of the surrounding element,  
which was suggested by another developer [1]

>> [...] shift-direction to bottom-to-top
>>
>> Which together results in: a positive value will shift towards bottom
>> (lower the box), while a negative value shifts towards top (raise the
>> box),which is contradictory to the specification before.
>
> So now 'super' would *lower* the graphic when baseline shift is set to
> a positive number.
> As to why you'd want to do this ....
> Mmmm.

For simplicity. Assume a "graphic" containing text: It does have the  
baseline somewhere in there. If I would like to specify where the  
real baseline is, i would find it natural to specify a percent value  
between 0 and 100 (or a length between 0 and the height of the  
object) rather than a negative value.

> HTH

So is it safe to assume that 7.31.22 is incorrect, and it should say:

<percentage>[...]
alignment-adjust="-<percentage>" (in the case of bottom-to-top)

and
<length>
alignment-adjust="-<length>" (in the case of bottom-to-top)

?

And if so, what what would I need to do to get this on the list of  
errata?


[1] http://issues.apache.org/bugzilla/show_bug.cgi?id=42785

Max Berger
e-mail: max@berger.name

--
PGP/GnuPG ID: E81592BC   Print: F489F8759D4132923EC4  
BC7E072AB73AE81592BC
For information about me or my projects please see http:// 
max.berger.name