But what's the question? (was: RB clear winner in speed test)

[Monte Goulding]

I agree. It's frustrating to have the rules change all the time. I think if
we had a clear description of the problem then we would get a truer
comparison between the development environments. One thing that might help
would be to have expected output as part of the test.

I have my doubts about even the new test data as it seems to be a pointless
exercise and would be far more enjoyable to actually read the book ;-)

In the end no matter what development environment you choose it's
application design that's the key. Using a database back end like Valentina
this entire problem could likely be cut down to one SQL statement.
Maintaining text metadata in the background would likely cut the problem
down to under 500 milliseconds. Maintaining search phrase metadata could cut
it down even more. It's all a matter of design and knowing your application
requirements. We can't change the fact that RB has a faster substring
searcher than Rev but we can make that fact irrelevant.

Cheers

Monte