I'm still here.  A busy spell, a tricky update, and some experimenting that didn't pan out (yet) makes for a slow release.  Can't have weekly updates forever, can we?  But here it is, your brand new shiny 1.0!

It's all about the menu.  First off we have auto-placement.  Instead of dumping all new entries into an "Other" group, 1.0 will try to find the best group based on directory and put it there.  Also, if that group is alphabetized (even if just the file entries) it will put it in the correct position.

Second, if adding a file to a group makes it too long, it will try to sub-group it automatically, again based on directory.  If you have a C++ project and just dump all the source files in a single directory, this isn't going to help you, but for languages like Java and Perl that require the directory structure to match the package structure, this should work very well.  It only sub-groups when automatically adding files, so it shouldn't be constantly changing your menu out from under you.

Third, auto-titles got a whole lot better.  If every entry in a group starts with the same package or directory name, that will be stripped off.  If the titles are still too long, the middle will be condensed.  Finally, you don't have to worry about forgetting to delete "auto-title" when changing it manually.  Auto-title is now the default, and Natural Docs will detect whenever you change one and apply the "no auto-title" attribute for you.  You only have to delete it if you want to make it automated again.

Take these three features: auto-placement by directory, auto-grouping by directory, and stripping common package prefixes, and combine them with a language that requires the package and directory structures to match, and the auto-menu generation becomes all you need.  You'll still get some benefit in C++, but it depends on you having an organized source tree.  Try it out by renaming your menu file and letting Natural Docs generate a new one.

There's still more.  In the HTML, menu groups will expand beyond the minimum required to show the selection.  Expanding the groups in close proximity to the selection improves usability considerably; try it out in our documentation.

Finally, 1.0 has a number of little tweaks like faster index generation and function prototypes that break so there's only one parameter per line.