Re: Fonts again in ConTeXt

[Patrick Gundlach <pg@levana.de>]

Jens-Uwe Morawski <morawski@gmx.net> writes:

Hello Jens and others,

>> >>Well, in my opinion, the berry typesript is sub-optimal, since it
>> >>maps to the urw variants that are a) shipped with a few TeX systems only
>> >>and b) (even worse) the urw variant helvetica (nimbus sans) is of no
>> >>real use for the german people out there (broken germandbls).
>> >
>> > then, how many people do have the adobe palatino's?
>> 
>> None that I know of. But this is IMHO irelevant: the current TeX
>> practice is:
>
> No. No.

Well it is the way psnfss works. That means all LaTeX users do it
this way and this is about 99% (wild guess) of all TeX users. So it
is current practice.


>> 1) TeX macro package (aka LaTeX ;-) asks for "Adobe Helvetica", it
>>    gets the font metrics from "phv....tfm". Then, in the map files
>>    the users decides whether to use the original adobe fonts or the
>>    substitues. 
>
> the MAP-file is a nice place for the real font selection. 

I agree with that. (Did you mean it this way?)

> How should i tell "normal" users that their font-selection from the
> document is overwriten since some stupid and wrong decisions have
> been made in the past.

That is easy: You don't have the real helvetica? Then your dvi
processor (or pdfTeX) switches back to a substitution. This is the
way every printer works, that don't have the Linotype fonts built in. 
So what wrong decisions have been made?

>
>> 2) TeX macro package (aka ConTeXt ;-) asks for "URW Nimbus Sans"
>>    a) on TeXlive (and perhaps any other I don't kown of) it gets the
>>       metrics from "uhv....tfm". Then, in the map files the user has
>>       no choice; the dvi/pdf driver should take Nimbus Sans.
>
> that's correct, since *it is* URW...

>> I think it is wrong behaviour to explicitly ask for the substitution
>
> it's no substitution, it is the font that is ask for. You asks for
> URW and you get URW. Simple, isn't it?



Well asking for URW, you get URW; that's clear. But there are three
steps involved:

1) the user: he or she requests the typescript [times][..]. What does
   he want? URW? Or what? IMO I request Times-Roman, and not a
   substitue. 

2) ConTeXt (or other TeX macro packages): When "Helvetica" is
   requested, it is mapped to uhv..., which is not Helvetica but
   Nimbus Sans. -> IMO this is wrong behaviour. The user request
   should be carried through the whole system right to the end, and
   let the last possible part in the printing process decide "to
   substitute or not to substitute".

3) The dvi processor: uhv-> get Nimbus Sans. phv-> get Helvetica,
   when you don't have that (default) get Nimbus Sans. 

The way with current ConTeXt it is now: I ask for Helvetica and get
Nimbus Sans. There is no way around this. 

> The way the psnfss works is wrong, since one asks for Adobe and gets
> URW. Please read some threads in the PDFTeX-ML and you can see what
> confusion this creates.

I did a quick search for Adobe URW, but found nothing. Could you give
me some hints?

>> and leave the user no choice about which variant (adobe, urw) to use.
>
> The choice is made with a new typescript. This is the common procedure
> when a new font is installed. 

I am *not* talking about new fonts! I am talking about some base
fonts that are shipped (or better: not shipped) with the TeX system.

> So, if you have Adobe Palatino then install it as you did for other
> fonnts too. But, stop! Sorry, there are always entries in the
> current pdfTeX MAP-file that map your Adobe fonts back to URW
> variants. So, first you to have edit the MAP-file. 

This is what the mapfile is for!?!




Patrick