DOCBOOK-APPS: Processing html/wml files to Docbook-SGML

Hello,

I have the unthankfull task to include some WML (Website
Meta-Language) files into a Docbook SGML document. The files are also
avaible as (quite plain) html, but then contain some small tables. So
such a convertor would also help.

Most problems (in times of work and making errors while editing) are
the missing </li> in lists, unclosed <p>s and some strange tags
resulting from WML (and my naive way of doing it last year ;-). These
are mainly <key CTRL> constructions which need to be
<keycap>CTRL</keycap> afterwards.

I guess such filtering could be done in Perl with regular expressions, 
but I really have no clue about this topic, so any hints, examples or
pointers to helpfull tools (unix/linux) are very welcome, I am going
crazy... 

Thanks,
Carsten.