split html for pagination

[arfa]

I have set up a nice javascript pagination system using explode() with a pre-set seperator. This divides the root file by 'item' (an item comprises a title and sub-text)

For some pages the items are quite long and I would like the option to have a page split according to text length. The pagination script requires an array so I have been playing with:
$split_text=chunk_split($str,22,'|');
$split_array=explode('|',$split_text);

The only problem here is it ignores both words and html tags.
The only word-related php I can find is: str_word_count and this also ignores html tags.

I guess I could explode using a (full) stop but this snags if the page includes a path reference - eg. img src=../images/pot.gif
Ditto splitting by space.
In short, there seems too many if{} else{} permutations to try this approach.

So, how to split (or explode) a bunch of html in relation the actual text?
Any suggestions or solutions would be much appreciated.

cheers - arfa