[XeTeX] XeTeX 0.4 available

[Hans Hagen]

Hi Ross,

>Yes; but in XeTeX there is no such ability yet.
>XeTeX does not allow you to directly construct PDF coding,
>and there is no PostScript --> PDF conversion step, to get it that way.

by reading the docs i feared that already -)

(btw, i still support pdfmarks, but haven't distilled a file in years)

i think the best way to go is to let the xtex dvi/pdf driver support the 
dvipdfmx specials; that would safe us a lot of trouble and also provide a 
quite compatible feature set

>Back to cropping, scaling, etc.
>These occur in the options to \includegraphics, so are naturally thought-of
>as being applicable just to the imported graphic, at least in that context.

sure, but pdfliterals can do the same in a more general way; so, instead of 
making graphic placement more powerful by extending its options, i'd go 
adding pdf litral inclusion; otherwise we're going back in time some 15 
years -)

>But while there is not such a general method, and there may be an easy way
>to do these things for single images, then that's the current question.
>What we really need to see is documentation on the interfaces that
>Jonathan is using, so as to find what is easy to implement, and what is not.

sure, but since i expect an engine to have the more generic options, i'm 
not going to spoil my macro set / driver concept by using the rather 
specific for each driver different graphic extensions; i just wait for the 
generic option then

>But this is going beyond Jonathan's stated aim of being
>  "simple to use the advanced typography features of OS X"
>  --- though ultimately we would like to get there.

sure, but we no longer liv ein a text only world, each doc we make today 
(at pragma) has color, graphics, etc

>Hans, what do you think will be easiest in the short term,
>and best in the long-term ?

just start with:

   \special{pdf: literal [direct] ...}

or

   \xtxpdfliteral [direct] {....}

because that would already permit much (transformations, color, and 
graphics); object constructors are next (i can post one feature request per 
day to keep things going -)

>I think that 'pdfmark' should be do-able, at least for short-term,

not really, interpreting them is no fun and specials are simpler

>and that complete integration with pdfTeX could be a long-term goal.

dvipdfmx special compatibility is enough at the moment

Hans