Re: Where we headed?

On 06/04/2013 03:31 PM, Tony Graham wrote:
> On Tue, June 4, 2013 2:01 pm, Jirka Kosek wrote:
>> On 4.6.2013 2:40, Arved Sandstrom wrote:
>>
>>> 1. Truth be told, since XSLT is declarative and works well in this
>>> space, I am convinced that an bettter FO formatter would actually be
>>> written in something like F# or Scheme or Haskell.
> Back in the early 1990s, Uniscope had a transformation language for SGML
> written in Scheme and a Scheme preprocessor and TeX macros for formatting.
>
> Some days I think of implementing a formatter in Erlang.  It's appealing
> to think that you could just fire off each page sequence and they could
> fire off their blocks and they could all just sort themselves out, but the
> reality might not be so appealing.
>
>> RenderX in past tried to develop FO engine in Common Lisp:
> Uniscope for a while also had a SGML parser written (before my time) in Lisp.
>
> Regards,
>
>
> Tony.
>
>
>
In this particular case, an XSL-FO formatter, I'd probably go with F#. 
It's impure, which is practical. It's got great support for concurrency 
and parallelism, and pretty good support for actors if you chose that. 
And there are pluses like Units of Measure ( 
http://msdn.microsoft.com/en-us/library/dd233243.aspx), so you could 
have the compiler actually help make sure that when you're doing 
calculations with FO areas etc with dimensioned numbers, that you're not 
mixing apples and oranges.

Arved