Where to split Directrory Groupings? A-F | G-K | L-P

Question

I'm looking to build a "quick link" directory access widget. e.g. (option 1)

0-9 | A-F | G-K | L-P | Q-U | V-Z

Where each would be a link into sub-chunks of a directory starting with that character. The widget itself would be used in multiple places for looking up contacts, companies, projects, etc.

Now, for the programming part... I want to know if I should split as above...

0-9 | A-F | G-K | L-P | Q-U | V-Z
10+   6      5     5     5     5

This split is fairly even and logically grouped, but what I'm interested to know is if there is a more optimal split based on the quantity of typical results starting with each letter. (option 2)

e.g. very few items will start with "Q".

(Note: this is currently for a "North American/English" deployment.)

Does anyone have any stats that would backup reasons to split differently?

Likewise, for usability how do users like/dislike this type of thing? I know mentally if I am looking for say: "S" it takes me a second to recall it falls in the Q-U section.

Would it be better to do a big list like this? (option 3)

#|A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|T|U|V|W|X|Y|Z

Answer 1

I would suggest one link per letter and hide the letters that don't have any results (if that doesn't asks for too much processing power).

Answer 2

As a user I would most definitely prefer one link per letter.

But better (for me as a user) would be a search box.