On Mon, Dec 14, 1998 at 07:19:17PM +0100, Hans Aberg wrote:

> >It's madness, in my view, to modify the TeX engine
> >in order to allow inclusion of particular graphic formats.
>
> What modifications do they do in pdfTeX?

I don't understand the question.
Since TeX does not understand graphics in any format while pdfTeX does,
there is obviously a modification, as indeed is clear from pdftex.ch

@# {data structure of images; for |pdf_image_node| and |pdf_ref_form_node| the
identifier of XObject will be store in |obj_info| of coresponding object in
|obj_tab|}
@d pdf_image_node_size      == 5 {size of whatsit node for image}
@d pdf_image_info(#)        == mem[# + 4].int {pointer to hold data structures
in external \.{libpng} library}
@d obj_image_ptr            == obj_aux {pointer to image structure}

However, my main point was that since pdftex is a single monolithic program,
every additional graphics format -- PDF, PS, TIFF, etc --
must involve further modification to pdfTeX itself.

> >While Thanh's pdfTeX is a marvellous piece of work,
> >which satisfied an urgent need,
> >in my view it is fundamentally misconceived,
> >and will in time be replaced by a version based on possibly extended DVI.
>
> But will an extended DVI suffice as a new byte-code for WWW publishing?

I take more or less the opposite view to that generally expressed here.
In my view, it is up to browsers to accept
generally accepted formats like PDF or DVI --
it's not up to the outside world
to try to convert information into the format expected by the browser.

I would expect browsers to define something like a Java interface,
with the understanding that they will display anything
which implements this interface.

--
Timothy Murphy
e-mail: [log in to unmask]
tel: +353-1-2842366
s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland