[docbook-apps] Re: Proposal: stop shipping entity sets with DocBook V4.4

[Karl Eichwalder <ke at gnu dot franken dot de>]

Michael Smith <smith@xml-doc.org> writes:

> To the Pro list, you might also add:
>
>   - named entities are a legacy convenience feature, inherited from
>     SGML, that users should not be encouraged to continue using

Using XML as a file format does not imply you do not want to process
the files using propr SGML tools.

> I personally think we ought to be encouraging users to move to doing
> their authoring and validation using tools that support Unicode

Often this isn't an option.  It might work for DocBook where usually not
that many "strange" characters are required.  For sure, I don't want to
enter things like &ndash;, &mdash;, &nbsp;, etc. literally, because on
the screen they are often hard to distinguish.  Using Emacs from CVS,
you can enter strange Unicode characters (âââÂâÂâââ), but it is easy to
mess up your files with these things â even if you arrange special
display features as (use a colored underscore for displaying the
unbreakable space, etc.). 

> and that are driven by modern schema languages (RELAXG NG and W3 XML
> Schema), which don't support named entities unless you declare them in
> an internal DTD subset in each and every one of your doc instances.

The easiest way to work around those idiosyncrasies is flattening your
XML file before before processing it with "modern" XML tools.  I guess
xmllint can do this job (--noent).

-- 
                                                         |      ,__o
                                                         |    _-\_<,
http://www.gnu.franken.de/ke/                            |   (*)/'(*)