Hi Jim, I see where you're coming from. The underlying assumption with DITA is that each XML document is standalone, which is why the DTD enforces some #FIXED values so all attributes are known. To process a standalone DITA instance you would either have to have the #FIXED attribute values in the XML, or have the XML parser apply the DTD so the #FIXED attributes are respected. I guess another approach would be to hard-code DITA #FIXED attributes in to your application so you don't need the DTD or attribute values defined but that doesn't smell like good architecture to me. I know this doesn't answer your question but might help give perspective. In that context it seems safest at point of parsing (eg. SAX) to add the #FIXED attribute values so the document is serialised in a known state. IMHO, the authoring tool should be hiding the @class attributes anyway as these represent fundamental DITA infrastructure and not something authors should be playing with. Perhaps there is a tool which is able to remove (de-dup?) attribute values which are defined as #FIXED in the DTD? It would be a bit of a pain to hand-roll your own tool to do that. Cheers, Gareth Chief Architect, GPSL
From: Jim Tivy <jimt@bluestream.com> Date: Monday, 10 June 2013 20:47 To: "xml-dev@lists.xml.org" <xml-dev@lists.xml.org> Subject: [xml-dev] Processing and Serialization of FIXED attributes Hello The DITA XML DTD has a number of #FIXED declarations in the DTDs that imply fixed values for class attributes. When this is processed in SAX and serialized the DITA class attributes are also serialized. It would be cleaner, for future XML authors to not have to see these class attributes inlined into the XML document. Do XSLT or other serializers commonly take out FIXED attributes or does the infoset mark them so they are not serialized – or is the usual form to serialize these. In
http://www.w3.org/TR/xslt-xquery-serialization-30/#xml-output it states “MAY” · The
reconstructed tree MAY contain additional attributes and text
nodes resulting from the expansion of default and fixed values in its DTD or schema; also, in the presence of a DTD, non-CDATA attributes may lose whitespace characters as a result of attribute value normalization. My question is, what do most java based implementations of XLST processors, SAX processors and associated serializers do with respect to inlining or stripping out #FIXED attributes. Jim |
