Hannes Magnusson wrote: > I agree with Richard here. The advantages of having all files well > formed XML out weight any argument I myself have tried to come up > with. > > My initial testing of adding DTD to all book.xml files, using custom > catalogs and XInclude shows no noticeable performance hit, it actually > reduces the memory overhead drastically.
That was my primary concern. If this is, in fact, not an issue, I'm all in favor of XIncludes! The DocBook mentions some disadvantages of XIncludes: [1] - XInclude can only have one root, parsed entities can have multiple elements (I don't think this is a problem with PHPDoc) - An XIncluded document with an IDREF that crosses document boundaries still won't be valid. I suppose adding a META id section that defines all the IDs and is discarded when we perform the XInclude would be one way of fixing this, but it seems terribly hacky. I presume we'll still use entities for text-level inclusions, so we'll still need to define a custom Doctype for PHPDoc. [1] http://www.docbook.org/tdg5/en/html/ch02.html#ch02-physdiv -- Edward Z. Yang GnuPG: 0x869C48DA HTML Purifier <http://htmlpurifier.org> Anti-XSS Filter [[ 3FA8 E9A9 7385 B691 A6FC B3CB A933 BE7D 869C 48DA ]]
