Re: Align argument in #+attr_latex for tabular env. is whitespace sensitive

[Nicolas Goaziou <n.goaziou@gmail.com>]

Hello,

"Sebastien Vauban"
<wxhgmqzgwmuf-geNee64TY+gS+FvcfC7Uqw@public.gmane.org> writes:

> For my information (and not only for me, I guess), is there a fundamental
> reason (I guess yes) why width is a first-class parameter for tables and an
> option among others for images?
>
> Looking from a 50 ft perspective, I would even imagine the opposite: images
> will have to be scales in almost all cases (that is: have their image
> attribute set), while tables will often take the space they need to.
>
> Is there a reason why images can't have their own attributes as well?  I would
> find that much, much more easy to apply, if we don't have to write the same
> "functional" request in two different manners, depending on the
> context.

It's LaTeX's fault!

In tables, width is an argument between curly brackets. In other words,
it has a specific location among the command arguments. On the other
hand, in images, width is just an option between square brackets. And
the same brackets can contain other options.

So, the easiest implementation was to follow LaTeX's conventions. An
argument within curly brackets gets its own keyword. Optional arguments
are stuffed within an `:options' keyword.

With care[fn:1], "width" option could be extracted from `:options'
keyword and get its own `:width' keyword. Do you want to give it a try?


Regards,

[fn:1] By deciding what to do when width is specified in both :options
and :width keyword, for example.


-- 
Nicolas Goaziou