FO preprocessor

I'm writing some kind of "merge" preprocessor for a FO 
document.
The FO document would contain XML tags from a new namespace 
(the mrg namespace) melted with the actual content.
The merge preprocessor would generate a new FO document 
with FO tags and content untouched but would process mrg 
tags for instance by querying a database.

An example:
<fo:block>Hello <mrg:field name="who" 
table="people"/></fo:block>

would be converted to
<fo:block>Hello Jordi</fo:block>

Then FOP would be used to generate a PDF from the processed 
FO.

Sometimes it is desirable to generate non-merged documents 
so I might write a FOP extension to cope with mrg tags.

Question would be: experienced FO guys would say that this 
is a good approach?

*jordi

---------------------------------------------
Aquest missatge ha estat enviat des del servei
de correu gratuït de VilaWeb.  http://correu.vilaweb.com

Aquest servei és públic i, per tant, el contingut
del missatge no té cap relació editorial amb VilaWeb.