But some seemingly obvious flaws might not be, when seen from the perspective of longevity and evolution rather than short-termism. Cadenced game design is a process of designer and player co-discovery, not just agile development efficiency. It's a process of building a durable cultural form as much as a stable product.
Take the Words With Friends dictionary as an example. It's terrible -- rudimentary and incomplete, failing to recognize common words, plurals, tense changes, and other inflections. No serious player will fail to encounter this limitation, but that doesn't make it a game design problem, exactly. After all, a limited dictionary might be a welcome play constraint; for example, Scrabble's rules prohibit abbreviations partly to reduce the number of viable two-letter plays (often key to expert play). Rather, the dictionary could be seen as an opportunity with many possible solutions.
Words With Friends could strive to offer the most complete word game dictionary around. That could take place through the use of a better dictionary, or through continuous updates, or even by using human computation to suggest and validate rejected words that should be included.
Or, if its creators really wanted to embrace Zynga-style monetization, the Words With Friends in-game store (did you even know it existed?) could sell custom dictionary add-ons: Disney/Marvel, Election 2012, Particle Physics, Molecular Gastronomy, Proust -- whatever.
Or, following Draw Something's model, Words With Friends could release limited edition collections of words, keyed to current trends or events. Like in Bookworm, these words might offer a bonus if played in a particular game. No matter the case, the dictionary's shortcomings suggest many possible avenues for future development, not just one obvious solution.
In fact, the dictionary reveals another of the game's quirks: while Scrabble is a game about knowing words, Words With Friends is a game about finding words. Thanks to the game's lack of penalties for plays that don't find a match in the game dictionary, players can try out endless possible combinations of letters until one of them works. The game's asynchronous nature tends to magnify this play pattern; none of the social anxiety of long turns exists in a distributed play session. And besides, each player has his or her own private screen for play, thus making it possible to hide experimental moves in a way that wouldn't be possible on a coffee table.
What to do with this unexpected situation? One answer would be to revel in it. Zach Gage's indie word game hit SpellTower features word-finding as a core mechanic, eschewing both time constraints and vocabulary exertion in favor of an open invitation to try as many hypothetical moves as possible before committing to one. But Gage -- who admits that a hatred for traditional word games partly motivated SpellTower -- had to devise a completely new design to offer an experience based on finding words over than manufacturing them.
Instead, Words With Friends might embrace its encouragement of word discovery, but add orthogonal elements to downplay its tendency to take over games among well matched, mid- to high-level players.
One answer already exists in Zynga's other mobile hit, Draw Something, which demonstrates every stroke of a player's entire drawing while presenting the result to a competitor. This revelatory experience is certainly part of the excitement and appeal of the game, but it also serves a design purpose, implicitly challenging players to guess a drawing as early in its creation as possible (even though the game offers no explicit rewards for doing so).
A similar approach might be possible in Words With Friends, but with the opposite result. By storing and displaying all of a player's trial moves, including loose tiles placed on the board experimentally as well as word "guesses" rejected by the dictionary, an player would gain a partial view of an opponent's tiles, as well as his or her placement penchants.
Thus, a balance could be struck between the boundless experimentation the game currently allows and the closed, touch-play effect of traditional tabletop Scrabble. Discovery would still be possible, but in a form dampened by revelation.
Or, Words With Friends could fully embrace word discovery, making it a more central part of the game. Such a result is harder to imagine, but not impossible. For example, EA's mobile edition of Scrabble offers a hints system, which automatically selects the best word on the board.
This solution may go a bit overboard, as successes like a seven-letter "bingo" really ought to be limited to player accomplishment. Still, you can imagine a version of Words With Friends that might highlight words of a specific length crossing a particular tile, playable over a particular bonus, or in a particular region of the board, inviting controlled experimentation. Such invitations could simply offer hints, or they could come with bonuses akin to those suggested earlier for special per-match words.