Re: How do these 4 different methods of creating classes compare?

[Safwat Halaby <softwatt@...>]

On 12/14/2015 06:36 PM, Alex Silva wrote:
> These are the kind of questions where assuming gives you nothing. You
> need to measure it.

While it is generally true that often performance discussions without
benchmarking are ineffective, this is not the case.
There are many objective answers to many of my questions. E.g.

- approach 2 and 3 are either identical or not identical, regardless of
benchmarking.

-Metatable access is either slower or not slower than regular table access.

-The way metatables are implemented gives a definite answer to the
question of whether approach 1 uses less memory than the remaining
approaches.

Attachment: signature.asc
Description: OpenPGP digital signature