See http://www.lmnl.net/index.html

What problem does it solve that is not already solved?

''This appears to be a similar setup to XML, with the dubious capacity for allowing "ranges" (elements which can overlap).  I fail to see the importance. -- StoneCypher''

It adds layers, in which you can mark up ranges over things other than elements (like ranges), and more interestingly annotations (like XML attributes) can have structure. Imagine if XML attributes were XML tags which could also have attributes. Some parts are messy but it has some interesting ideas. -- StevenMurdoch

What they are attempting to do is to create a system for allowing honest MarkUp of HyperText, something HTML doesn't let you do. One of the severe weaknesses of XML is that the data model doesn't allow overlapping of tags, and forces everything, including database tables, into a tree. While it may make some things easier, you are forced to abandon everything to get the tree. You can't have multiple dimensions or layers, you have to fit the tree, its OneSizeFitsAll, and it sucks!

TedNelson has great ideas about HyperText, and this is somewhat closer to the vision. I like the fact that you can use the existing markup of LMAL documents to extract portions, somewhat like transclusion. Its a compromise, but I think its useful, and practical. (Especially considering the source may get tweaked, even a spelling correction could throw off something using absolute offsets).

I dare anyone to figure out how to add comments to this document without copying it, or modifying the source. You can't because HTML doesn't support MarkUp as defined by common English. You can't link the the middle of a document with HTML either, unless there is an anchor. You certainly can't transclude. 

In summary, HtmlSucks, and this is part of the cure.  -- MikeWarot

Perhaps YouAintGonnaNeedIt.