[texhax] which TeX?

[Doug McKenna]

Pierre MacKay wrote -

>I should think that the "compatible" requirement would militate against 
>any rewrite that did not start with Pascal.

Given that it's possible to automatedly translate Pascal programs to C 
(my recollection is that TeX has been so translated), I doubt anyone 
would feel required to start with Pascal.  At the very least C or C++, 
one would think, to keep things as fast as they currently are.

>THe conservatism about the WEB version is part of the promise that the 
>TeX engine will continue to produce the same output from output from any 
>given input sinsince TeX82 was stabilized (and for the most part, even 
>before).  Additions to that engine and its feature---I do mean 
>feature---of archival stability really must adhere to the requirement 
>that if you call it TeX, it should pass the TRIP test, and I think that 
>it would be extraordinarly difficult to do that in a new language.  

A reimplimentation of TeX would have to pass the TRIP test, but that's a 
different matter from creating a new typesetting language (e.g., like 
ANT) (if that's what you meant).  I was wondering more about a 
TeX-compatible engine, re-implemented in something besides WEB/Pascal, 
that not only passes the TRIP test (a necessary but not sufficient 
condition for compatibility) but also sufficiently (within some 
acceptable epsilon anyway) preserves the legacy of all the myriad TeX 
documents out there.

>Probably not worth the effort.

Yup.


Doug McKenna