“Macro Caching”

While working on the new weblog plugin this weekend, I found that (as is too often the case) there were a few things that could either be done "right" or "fast," but not both.

I opted for right -- maximum flexibility -- but was more disappointed with the performance than I expected.

The solution was macro caching. The output of a macro can now be stored after the macro has run. The next time the macro is requested, the cache is returned (almost instantly) instead of running the macro again.

With only a few lines of code, Conversant's slowest macros can now be made its fastest. For example, the insertQuery macro (which searches the database and renders the results into the page) -- perhaps the coolest macro Conversant has (so far) -- was occasionally also the slowest. With caching, it's one of the very fastest. The insertQuery macro is a great tool for knowledge management, and its new speed should make it even more popular among Conversant administrators.