[pdftex] TeX as a composition server?

[James Quirk]

Ross,

> By providing the loop in TeX, he can explore with all kinds of things 
> to get an idea of the timings for different tasks. As well as file I/O 
> there is page-composition, headers & footers, indexing, etc. which can 
> all be varied and tested.
> 
>  In the actual situation, these may be able to be optimized by different 
> kinds of technique, for best performance. You want to make the actual 
> looping aspect as innocuous as possible, yet retain flexibility to be 
> able to test other things.
Agreed.

> 
> It is precisely this kind of thing that makes TeX continue to be relevant. 
> Workflows can be stitched together, as you have done with AMRITA, taking data 
> from whatever source and producing a high-quality, feature enriched, 
> representation within a PDF document which is accessible to everyone 
> (through commercially produced, but still essentially free, readers.
> 
> Personally, I do not see any part of TeX that needs to be deprecated.
I have to tread carefully so as not to get myself into too much trouble, 
but I would argue that there are elements of TeX-the-mindset that need to 
be deprecated, which I can elaborate on offline. And should any of my 
concerns stand up to scrutiny, they can always be forwarded on here. One 
of the less contentious concerns, which has already been raised by Ivo, 
has to do with the number of active developers. On that score ConTeXt
is a much healthier enterprise than say LaTeX, or at
least so it appears to this bystander.

> Rather, it is an ever-expanding suite of software, much of it being developed 
> to be able to prepare documents for the ever-expanding realms of display 
> devices and software techniques appearing on the market, over which there 
> is no real control whatsoever --- and neither should there be. (They 
> will live and die on their merits.)   
No. I have to disagree. Historically, software has lived and died on the 
commericial clout of its backers and not on its actual technical merits. 
For instance, it's generally accepted that Fortran won out over Algol 
because of IBM's backing. Although I'm sure Robin is better placed than 
myself to adjudicate on the accuracy of this statement, for I only started 
to program in 1980. Anyhow, we're seeing the same sorts of battle 
now with PDF versus HTML5, and much of it is ill-informed. 
For those who are interested, Adobe's side of the argument is given at:
www.adobe.com/enterprise/pdfs/html5_flash.pdf .
 

James