Notes on Published Recipes
There are any number of recipes available for cat food both on the Internet and in print. Some are designed by veterinary nutritionists and may be assumed (subject to the credibility of the source) to be nutritionally balanced and complete when prepared exactly as directed. For normal, healthy, adult cats these recipes are probably ok as long as they like the way they taste. There are, however, serious limitations on the use of these diets.
Possible Design Flaws
These are in most cases probably not critical but should be mentioned because possible cumulative effects might be significant in some cases.

  • The bases for all dietary design programs are "The Nutrient Requirements of Cats"published and updated by the NRC/AAFCO, and the USDA Nutrient Database. It's apparent that at least one of the popular software packages uses older versions of these resources which don't include updates that in some cases are significant.

  • To simplify matters these recipes generally take a "one-size-fits-all" approach to vitamin supplementation, simply stating that "a human multivitamin" should be added (or a standardized amount of a proprietary product that's being promoted). Not all products have the same constituents and this strategy can lead to over-supplementation of some items while leaving deficiencies in others. A more accurate method is to calculate what's missing after the basic ingredient list has been established, then determine what additions are needed.

  • Some recipes that propose to aid in the treatment of clinical conditions often ignore well-established principles in favor of their creators' beliefs of what a cat "should" eat. Their reasoning is generally based on a personal philosophy rather than any scientific basis.
  • Limited Ingredient Choice
    Cats are individuals and have individual tastes. Some like turkey but not chicken, some prefer beefy flavors, some go crazy over fish. For a food to be successful, the cat has to like it. This is especially important for cats with medical conditions that affect their desire to eat.

    Published recipes are almost universally generic, using one primary meat source like chicken or hamburger. They may or may not indicate that substitutions can be made, but unless the substituted ingredient is analyzed in context the integrity of the original recipe design is negated. There are no valid "rules of thumb" for substitutions. Not only are nutrient profiles different between chicken and beef, they vary depending on body part -- a chicken leg is different from a chicken breast in all respects: protein content, amino acid profile, energy density, etc. A pound of turkey breast isn't equivalent to a pound of chicken breast. There are considerable differences in the vitamin and mineral content of chicken liver, beef liver, calf liver and pork liver.

    Yields also differ based on whether the meat was measured raw or cooked, and vitamin content can vary considerably based on cooking method. If a recipe calls for "chicken breast, meat only, rotisserie", that recipe is not valid for "chicken breast, meat only, raw". Conversely if it calls for raw ingredients cooking it will change the analysis, requiring compensations to be applied.

    Some published recipes even allow the substitution of seafood, which has some very unique characteristics. Fish should never be used in a recipe designed for any other meat source.
    The only way to ensure the accuracy of a recipe design is to do the calculations based on the data for each specific ingredient as reported in the USDA Nutrient Database or other well-based authority which includes adjustments for each ingredient based on cooking method, etc.