I explored the similar issue some time back.

You could look at my findings at,

http://gandhimukul.tripod.com/xml/xml.html

Please see, item no, 6.

Regards,
Mukul

On Oct 10, 10:56=A0am, Chris W <chrisw_j...@yahoo.com.au> wrote:
> Hi All,
>
> I have hundreds of small XML files of the form (extrabeous stuff removed)=
:
>
> <?xml version=3D"1.0"?>
> <!DOCTYPE page PUBLIC "-//LOCAL//DTD PAGE 0.1//EN" "page.dtd">
> <page>
> =A0 <graphic boardno=3D"entityname1" />
> =A0 <graphic boardno=3D"entityname2" />
> </page>
>
> that I would like to process into this form:
>
> <?xml version=3D"1.0"?>
> <!DOCTYPE page [
> =A0 =A0 =A0 =A0 <!ENTITY =A0entityname1 SYSTEM "entityname1.gif" NDATA gi=
f>
> =A0 =A0 =A0 =A0 <!ENTITY =A0entityname2 SYSTEM "entityname2.gif" NDATA gi=
f>
> =A0 =A0 =A0 =A0 <!NOTATION gif SYSTEM "image/gif">
> ]>
> <page>
> =A0 <graphic boardno=3D"entityname1" />
> =A0 <graphic boardno=3D"entityname2" />
> </page>
>
> That is, I'd like to load each file, find all the boardno attributes,
> insert an ENTITY declaration, insert a NOTATION declaration, and write
> the result to a file. =A0The XML markup is unchanged, just the internal
> DTD is altered. =A0Finding the boardno attributes in a DOM is trivial, bu=
t
> manipulating the internal DTD subset and getting it to file is eluding me=
.
>
> Apart from doing the DTD manipulation as a text file, any suggested tool
> sets/approaches. =A0Perl, Python, Java, whatever.
>
> Regards,
> Chris W