Most search engines do not break words to all possible substrings. However the parser we are using here does break words based on upper case / lower case transitions.
For example you can search for Basic4android Search: Preference and get results with PreferenceActivity.
The method String.ToLowerCase is broken to the following tokens: tolowercase, to, lower and case. None of them match: lowercase. If you search for lower case you will see that it does return. With that said, if you search for lowercase you will still get a good result as the first result.
I wasn't considering that you have to index all the words. I was thinking of a search like the browser does on a page where it just does a string search on the entire block of text, thus finding any embedded instances. Oh well...