OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [office-comment] Support for overlined text in ODF


Martin Whitaker wrote:
> Thomas Zander wrote:
>> On Wednesday 09 May 2007 08:27:54 Leonard Mada wrote:
>>> I expect it too to be continuous, but where on the vertical axis would
>>> you position it?
>> ...
>> If we use the typographical rules then, actually, we don't have to 
>> (re)define it :)
>> By that I mean that if you ask 10 typographic experts and the fast 
>> majority will come up with the same answer we don't need to write 
>> that answer down. It is up to the implementer to gather that 
>> information.
>>
>> With my typographical background, I'd say;
>> If you consider one line of characters a block, from top of font to 
>> bottom (with the baseline in between) and the linespacing being 
>> excluded from this block, then the overline would IMO be at the top 
>> of that block.
>> If it grows thicker it then eats away at the linespacing (50%) and at 
>> the tops of the ascenders (50%).
>>
> Basing the overline position on typographical rules seems to me to be
> the correct approach. I'm not a typographic expert, so can't contribute
> an authoritive opinion. For what it's worth, though, OOo already contains
> code for rendering overlines (used instead of underlines when certain
> CJK fonts are selected). This code uses the "internal ascent" font metric
> (if available) and positions the overline at 50% of the internal ascent.
> If the internal ascent metric is not available, it estimates a value,
> based on the font height. This gives reasonable results with the fonts
> I have tried.

To my knowledge, there is no consensus on what actually overline is. I 
have searched the internet some while back. I did search it these days 
again, but did not find any hard written definition. I might be wrong 
and, in that case would politely ask to be pointed to the internet resource.

The problem with overline is not trivial. IF I remember correcly, the 
font height is the vertical distance between the maximum descent and 
maximum ascent.

However,
  1.) the ascent and descent may *NOT* contain the distance for letter 
accents.
       (they can, BUT they may also NOT)
  2.) Additionally, the individual letters of a font may have differing 
heights for ascenders and descenders (usually not, BUT occurs in some 
fonts). I have used such fonts in mathematical writing, where the letter 
"f" and "F" were disproportionally taller than the remaining letters.
  3.) for many entirely lowercase words, we may want to base the 
'overline' on the x-height (especially for fonts where the height of 
uppercase letters is 3-5x bigger than of the corresponding lowercase 
letters)
  4.) superscript => might extend well beyond the computed overline for 
that particular font-word combination

As there is no universal definition, I still believe that writing a 
robust definition (taking account of the specific cases mentioned 
previously) is the better way.

The most basic question is:
Q: *Should the 'overline position' be different for different words*?
[My answer is definitely YES + option for IDENTICAL overline!]

Or, more basically, should we allow the user to choose for a fixed 
position (with some limits) for the 'overline', too?

IF we take only the ascenders for the letters making the word into 
account, overline positions computed for "xxx", "xi", "input" and 
"INPUT" may be completely different. A user may choose however to use 
the same height for the overline, because that would look nicer (with 
some fonts). With other fonts, that would look totally silly, so he may 
want to use the first method.

These are just some thoughts and I hope that in the end we adopt a 
flexible solution.

Kind regards,

Leonard Mada


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]