Penguin
Note: You are viewing an old revision of this page. View the current version.
An XML document is WellFormed if it parses correctly. This means mainly that all tags must be closed, that they must be nested correctly, and that all attribute values must be quoted. This is not valid
<p>A paragraph <strong>here</strong>. <p>And another one there.
The paragraph tags are not closed, and the nesting for the emphasis and strong tags is wrong. To be WellFormed, this piece of the document has to be written like so
<p>A paragraph here.</p> <p>And another one there.</p>

In the second fragment, neither the <b> nor the <c> tag are closed. Unlike SGML, XML does not allow tags to be automatically closed when the enclosing tag is closed. This is the reason why the <p> tag in HTML/XHTML gives people grief---in HTML you only need to put in the opened tags while in XHTML you need to put in both the opening and the closing tag.

If you want non-nesting overlapping ranges you cannot use something like

<a> ... ... </a> ...

but should use something like

<a id="1"/> ... <b id="2"/> ... <a id="1"/> ... <b id="2"/>

instead, and then you can reconstruct either of the tags as necessary. This flattened form also allows certain kinds of diff-like reasoning which are much more difficult on trees. An alternative to this, which avoids flattening the markup, is to do
<foo type="a" partof="1"> ... </foo> <foo type="a" partof="1" partof="2"> ... </foo> <foo type="b" partof="1"> ... </foo>

This is bulkier (especially when there are a very large number of overlapping structures) but much easier to process with XSLT. The choice between these is the subject of a minor ReligiousWar. StuartYeates is in favour of the former, AristotlePagaltzis in favour of the latter. The whole problem was solved in SGML using the CONCUR feature, which appears never to have been implemented and was dropped from XML.

WellFormed XML differs from Valid XML in that Valid XML has been (or could be) checked against a Schema or DTD.