The whole point of the patent is that it's for cross-cultural communication, not just for English names only. It's not a totally unreasonable idea. It sounds like it looks the user's gender up based on where they are from and what their name is. Odd spellings would likely be classified as unisex, unless there were a general rule for naming conventions (e.g. In North America, names ending in 'i' are likely to be female.) Furthermore, you could build up your 'odd spelling' database by recording the gender people select for themselves.
The example ozamosi posted below
would be covered fairly well by this patent: Robins in North America would be classified as female, but Robins from Sweden would be classified as male.
My criticism of the invention's effectiveness is that it's not completely fool-proof, and would inevitably assign the wrong gender for people with the spelling typically adopted by the opposite gender. It might be a worse "faux pas" to address a male as female (or vice versa), than to leave assumptions of their gender out of the picture. Of course this might vary from culture to culture, and I really don't know about that. It might be more effective to just force the user to input their gender, but this would have to be done on every client, which could be problematic.
Of course, I'm not sure whether we should be assisting the enforcement of "societal conventions" based on differences in gender, but that's a different topic from the invention's effectiveness.
By the way, here's the relevant part:
an expansive list of names compiled from those used in many different cultures catalogued according to gender (that is, male, female, or unisex), a list of rules for associating a username not included in the list of names with a particular culture, and a list of rules derived from naming conventions that are employed in many different cultures catalogued culturally, linguistically, nationally, regionally, and/or according to other relevant anthroponomastic criteria.