[scribus] Why scribus lacks letter-spacing? was: how does scribus do justification?

[Tornoci Laszlo]

On 01/28/2010 06:56 PM, avox wrote:
> Tornóci László-2 wrote:
>> In Scribus 1.3.5 you can set (Properties, Text, Advanced settings): Word
>> Tracking (that's word-spaces) minimal and normal value (I guess this is
>> the mean
>>
> Not exactly. Normal is the value that is used if no justification is
> necessary,
> minimum is % of normal and maximum is unlimited (that's implicit in
> linebased
> justification algorithms). Allowing stretching AND shrinking of spaces gives
> 2 break possibilities for each line instead of just one as with th esimple
> algorithm.
> As you can verify with Scribus the result is quite drastic.
Yes, that's clear.

>> ), and you can set Glyph extension (that's character width) min
>> and max value. There is no way to set letter-spaces. So we have #1 and
>> #3 but no #2.
>>
>> It is well known, that misusing letter-spacing is ugly. Setting it to
>> more than +/- 3-5% from 100% can lead to a change of "color" of the text
>> (the average grey impression of the text when not directly focusing on
>> it). But you can also misuse #1 and #3 as well! There is a wide belief
>> that letter-spacing is from the evil (based in part on the above
>> misquoted opinion of Goudy). I rather believe my eyes. There are lots of
>> examples in the book "J Felici: The Complete Manual of Typography" of
>> the same block of text set with different combination of H&J methods #1,
>> #2 and #3. It is clear, that a combination of these methods gives the
>> best results if used properly. I simply don't believe anyone can spot 3%
>> letter-spacing.
>>
> So you would be content if letter spacing was limited to +/- 5% ?
>
> But what is the base value? Normal letter spacing is 0pt, so we can't
> base it on that (well, in fact we already do ;-) )
> You mean +/- 5% of the character width? Or +/- 5% of a given character, e.g.
> 'M' ?
>
> Give me a definition and I might be inclined to implement that.
I am definitely not a font expert, so I have no idea what would be the 
best definition. However, I don't think there is much difference between 
any reasonable definitions. You change a setting, look at the result, 
and either you like it, or change the setting again. So the setting 
doesn't have to be exactly precise.

Here is what the Felici book (what I cited in my last letter) says:
"Generally, these values are expressed as some percentage of *normal*. 
For word spaces, *normal* means the width of a word space as it's 
defined within the font being used. For letter spaces, *normal* is the 
distance between characters when their bounding boxes abut but do not 
overlap. Spacing adjustments based on tracking and kerning controls are 
added on top of the spacing values calculated by the H&J program."

If I understand it correctly, this definition of letter space depends on 
the character pair considered. It may be possible, to use the value of 
an arbitrarily selected character pair.

>> So my humble request for the developers: PLEASE, AT LEAST CONSIDER
>> including the ability to change letter-spacing. You don't get as much
>> flexibility with it as with #1 and #3, but it is useful. I think Scribus
>> should be the best layout program, and this is a useful feature.
>> I understand the implications: the H&J algorithm becomes more
>> complicated, possibly slower, but I still think it should be implemented.
>>
> No, that would be a small, well contained change. I dont think it would make
> a difference in spsed, either.

That's good news. Thank you for taking the trouble to do this.

					Yours: Laszlo