JohnWalsh discovered a problem: Apparently, Antville converts all angle brackets that are not part of a tag to < and &rt;, but doesn't recognize all tags correctly, thus destroying valid html code.

As I said on help.antville, I've made something that extracts all defined tags from a DTD. This way, we can collect all existing (X)HTML tags for a start, so the bug reported above goes away. In the long run, it would be more elegant to have a seperate list of tags for each DTD and use each one where applicable.

Good morning hns; I even managed to sleep for about two hours tonight!

I analyzed some DTDs now; the result is: HTML 4 transitional defines all there is. XHTML 1.0 defines the same tags in lower case, except for two (who knows which two?) and XHTML 1.1 defines the same again, broken up into individual modules (e.g. text Module, Hypertext Module, Table Module, ...), but no additional tags. The total is 91, which is a little less than the existing list in the Helma source, which apparently includes tags from older HTML versions. I can add these, but do we really want to allow <blink>? (By the way, please at least delete the deprecated IE-only <marquee>, this is disgusting.)

I'm so proud of my little program, so here's my interaction with it:

For batch mode invoke: java tagstractor.Extract
 

Entering interactive mode.

input file name: xhtml1.0-transitional.dtd

output file name: tags.txt
Found a total of 89 tags so far.

Process another file (y/_n_)? : y

input file name: xhtml1.0-strict.dtd
Found a total of 89 tags so far.

Process another file (y/_n_)? : y

input file name: html4-transitional.dtd
Found a total of 91 tags so far.

Here's a list of what's missing from the Helma source:

allTags.add("acronym");
allTags.add("bdo");
allTags.add("dfn");
allTags.add("label");
allTags.add("legend");
allTags.add("noscript");

You can just add that to the source, I just noticed it isn't in strictly alphabetical order anyway :-) As you can see, my problem solution overkill really paid off, I found 2 (in words: TWO) tags JohnWalsh didn't find :-)