[luatex] luatex - font encoding for type 1 fonts

[Hans Hagen]

On 6/25/2016 2:06 AM, Reinhard Kotucha wrote:
> On 2016-06-24 at 10:06:57 +0200, Hans Hagen wrote:
>
>  > On 6/24/2016 8:36 AM, Mojca Miklavec wrote:
>  >> (I would argue that the best way forward in the long run would be to
>  >> eventually add the missing features to LM.)
>  >
>  > We had that discussion ... differences in positioning of some accents
>  > between aer, plr, csr, vnr, anyr fonts (plus german hackery to get
>  > umlauts lower) and the principles behind lm / gyre was to get a uniform
>  > consistent font. There have been discussions at BT about these issues
>  > and representatives of countries present in the end agreed that it's
>  > also a matter of getting accustomed to.
>
> I can at least confirm that the German and Vietnamese accents are
> placed correctly now in LM.
>
> However, in the first beta releases of LM the Vietnamese accents were
> placed incorrectly and later the correct positions of accents were
> derived from the fonts created by Thành.

Sure they were betas ... even now things get fixed (there are still bugs).

> If there is an acute or grave accent above another accent, it's
> usually moved to the right, sometimes an acute accent is moved to the
> left.
>
> I've spent two months in Vietnam and I can assure you that these
> horizontal movements are a must. It's neiter a matter of taste nor a
> matter of getting accustomed to.

At BT the discussion was more about how high should the diaeresis 
(umlaut) and similar accents or .. be above the u a e v s ... should we 
have several variants, should it be language dependent, what if one 
mixes languages ... it was compared to other fonts and one of the 
conclusions was that in most cases it was done consistent across latin 
scripts.

> As far as I understand, the problem with Czech is related to kerning
> pairs.  I suppose that it's no big deal to add some kerning pairs,
> especially if only Czech is concerned.

At BT the discussions i refer to were about the heights of accents (and 
consistency of the common ones among the cm derived fonts \^ \" \' etc etc.

Kerns can of course be added (there often are already more kerns in lm 
than in the originals) and in case of doubt can even be language 
dependent and it's mostly a matter of telling the lm folks (i don't 
remember seeing proposals for that). Btw, lm is also better for Dutch 
wrt kerns.

> If people report that they have to use csr instead of LM because LM
> doesn't support Czech properly, I strongly recommend to contact the
> authors of LM.  Not everything is a matter of taste.

Of course that also applies to other fonts. But the original question 
was more about how to support an old school 8 bit font in some special 
encoding (a dozen or so in the tex world) in an unicode / utf environment.

Anyway, I again looked at an easy way to load with the old csr and I am 
surprised about the availability: several packages, vf and non vf ... 
it's all geared to a specific kind of use I guess. It's a bit of trial 
and error case,

Hans

-----------------------------------------------------------------
                                           Hans Hagen | PRAGMA ADE
               Ridderstraat 27 | 8061 GH Hasselt | The Netherlands
       tel: 038 477 53 69 | www.pragma-ade.com | www.pragma-pod.nl
-----------------------------------------------------------------