Cache structs, not queries.

Thus far I've helped install two of the largest ColdFusion-based content management systems for two different clients, and in comparing both one thing is evident: they both just love structs.

Structs are used to return content. They're used to hold site definitions. Section definitions. Generated page information and definitions. Arrays of structs hold navigation lists. Flags. Constants. Paths. Titles. Keywords. Anything and everything.

In fact, even when a database query returns, say, information about the current site (name, domain, url, etc.), in both systems that information is quickly transfered to a struct and it, not the query, is cached.

Why? Glad you asked:

1. Structs are relatively lightweight containers, whereas queries tend to be quite a bit larger and are much more resource intensive.
2. Caching a query might entail maintaining an open connection to your database, reducing the number of connections it can handle.
3. Lookups and existence checks are faster than doing query-on-query searches, providing fast random-access to key/value pairs.
4. A cached query may count as one of your CF Administrator "cached queries". A struct doesn't.
5. Structs can hold more stucts, arrays, and other complex data types, allowing data to be nested and presented in a logical fashion.
6. Values can easily be altered, changed, and updated as needed.
7. Substitution into the HTML is quick and easy. Just do #request.page.title# and you're done.
8. With one assignment you can transfer a large amount of cached data into the current request scope for immediate access. No need to individually set two dozen variables by hand each and every time.
9. They're easy to dump and examine for debugging and development.

Take a lesson from the big boys. Just think of structs as fast little ram-based INI files, and your code will be "structured" in no time.