[Talk] Demystifying the Knuth-Plass (in French)

[Paulo Ney de Souza]

On Sat, Jan 20, 2024 at 6:57 AM Zunbeltz Izaola <
textnik.typesetting at gmail.com> wrote:

> Paulo,
>
> What is the Knuth-Plass algorithm that InDesign was trying to do? or the
> hz (Herman Zapf) algorithm that implies the scaling of glyphs?
>

That I do not know, with this level of precision. The guys working on the
implementation came to Berkeley
frequently to learn about Knuth--Plass and to compare outputs with TeX and
with old manuscripts, but they
would not talk much about what they were doing. We always assumed they
couldn't talk.... and they probably
were very busy transforming the code of the old application and
the competition of Aldus and QuarkXpress.

Then by mid-2000 when TeX started to be able to use any kind of font and
{microtype} appeared, I remember
it was a real blow for them.

Every year Adobe would put out a new version of the tools and as soon as I
got my copy, I would go for my
standard text to see how things changed. And indeed, they would change
every year, sometime for better
sometimes for worse, and it is hard to recognize Knuth--Plass in them right
now.


According to wikipedia it is not known if InDesign uses the algorithm.
> https://en.wikipedia.org/wiki/Hz-program
> but this resource
> https://www.typografi.org/justering/gut_hz/gutenberg_hz_english.html
> says it is part of the composition algorithm.
>


Interesting how information is hard to come by.

Paulo Ney




>
>
>
>>
>>> > implementing Knuth-Plass in Adobe InDesign.
>>>
>>>   I'm curious about this. Is this coming from your experience as an
>>> InDesign user (like, getting unsatisfactory results) or from some
>>> technical information that you have about their implementation? In what
>>> way is it failing?
>>>
>>
>>
>> Both. Some 15 years ago I was in touch with the group tasked, but
>> information was tightly controlled back then and much more now.
>>
>> I have followed every release since then and the results varied
>> tremendously from release to release and right now it seems they removed
>> everything and are using some in-house developed algorithm -- which has
>> awful results.
>>
>> Paulo Ney
>>
>>
>
> Zunbeltz Izaola
> TeXtnik typesetting
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/texhax/attachments/20240120/bbc57aad/attachment.htm>