Re: [xmlroff] xmlroff's "story"?

[Tony G. <Ton...@Su...>]

Stefan Seefeld <se...@sy...> writes:
> Tony Graham wrote:
>> Stefan Seefeld <se...@sy...> writes:
>>
>>>Mauro C. wrote:
>>>
>>>>--- Tony Graham <Ton...@Su...> ha scritto:
>>>>
>>>>>I started the "What needs to be done to make xmlroff
>>>>>a success?"
>>>>
>>>>IMO a wrapper exetension for PHP and other scripting
>>>>language.
>>>
>>>What would that wrapper do that can't be done by calling
>>>system('xmlroff ...') ?
>> If the scripting language's XML DOM implementation is a thin layer over
>> libxml2's XMLdoc, then it becomes possible to pass a DOM straight to the
>> thinly wrapped libfo.
>
> I understand that. What I fail to see is that there is indeed a need for
> this level of optimization (or whatever else a programmatic interface to
> libfo provides), at this time.

Some people on this list do want that (and if they can help in implementing
it, so much the better), and it was among the most emphatic responses to the
question of how to make xmlroff a success.

You (AFAICT) and I don't use Windows much if at all, yet we are discussing not
whether but how to do a Windows version of xmlroff.  Not because we want to
use xmlroff on Windows ourselves, but because it could put xmlroff in front of
more people, in the hope that some of them will be enamoured/exasperated
enough with xmlroff to want to improve it.

The same argument can be applied to a SWIG wrapper over fo-libfo-basic.h.
While you're not convinced of the necessity, you can just put it down to the
"feel good" factor for script programmers: if it looks like it's built into
the language, they're more likely to use it, so they're more likely to be
enamoured/exasperated enough with xmlroff to want to improve it.

Regards,


Tony.