D&D helper uses different types of databases to generate words, riddles, names, and more. To see what a particular type of database something is, just refer to this chart:

Pick

Instead of "generating" something, this file contains a really big list of words, names, or whatever. It then just picks one from the list.

Advantage: The results are really nice. They are actual words/names instead of ones that were assembled. This also works for other things, like random riddles.

Disadvantage: The database is larger. Usually much larger than the Gen or Small types. It also has a very finite number of names that can be generated. Gen and Small can generate a lot more using a smaller database.

Gen

Using a letter pair algorithm, this will assemble a word for you. It also knows what letters should be added onto the word while it is building it. For example, in English, the letter 'q' is almost always followed by 'u'. Gen keeps track of that information.

Advantage: Words and names that are generated in this fashion look like they belong to the language.

Disadvantage: Even though it is usually smaller than Pick, the database can still be larger than Small.

Small

This is the same as Gen, except it doesn't keep track of the likelyhood that a specific letter follows the first one. It just keeps track of wheter it happened at all.

Advantage: The database is smaller than Gen. Small can also produce wilder results because it has less restrictions.

Disadvantage: The words are sometimes unpronouncable, don't look like they belong to the intended language, and frequently give really odd results.

Special Notes: This database format usually produces very poor results and will only be used if I can not get enough data to use a Gen style database.

PSR

This takes a tree of rules and expands each branch in order to come up with the result. If written properly, this can make extremely realistic names without the size of Gen or Small. It can create sentences, paragraphs, plots, and much more without the size of Pick.

Advantage: The database is usually significantly smaller than an equivalent Small or Pick database.

Disadvantage: It takes time to create the complicated rules, which must be created by hand.