Re: [boost] Reference documentation: one approach

From: Joel de Guzman (djowel_at_[hidden])
Date: 2002-10-30 18:01:15

----- Original Message ----- 
From: "Anthony Williams" <anthony.williamsNOSPAM_at_[hidden]>

> Joel de Guzman writes:
>  > 
>  > Is there a (E)BNF definition for LaTeX?
> 
> I doubt it, as macros can change the meanings of each and every character that
> is read subsequent to the invocation of the macro. You can actually make TeX
> take XML as input by preloading a set of macros.
> 
>  > If so, where can
>  > I find it? And also, how good is the definition? Not
>  > like C++ I hope.
>  > 
>  > May I also add the requirement that the target language should be
>  > 
>  > --> formally defined in (E)BNF
> 
> Why? I presume for the following:
> 
>  > --> easily parsable
> 
> Still why? If there are existing tools to convert the output, or we can write
> something that does (e.g. use LaTeX itself to generate other stuff, much as
> TeX4ht does), why does the language have to be parsable with a generic parser.

Because I want to write tools that inputs and/or outputs the
target language, and I want to be able to do it as painlessly
as possible. 

In this day and age, why would anyone want to standardize on
a language that has no formal specification? 

--Joel