I was just reflecting on what I’d like to see improved upon in Ruby
libraries, and it struck me that I don’t really have a big problem with
the packaging. I can usually find what I am looking for eventually.
The problem I have is with speed and scalability consideration. I am not
saying that Ruby libraries are all slow or unscalable. The problem is I
don’t know if they are fast or scalable or if the author does either.
There are rarely benchmarks or discussions on design tradeoffs that would
make something perform inadequately under load.
It would be nice if I was evaluating, say, a web application framework for
Ruby if I could look at it compared to a couple of other ones in terms of
speed, maintainability, performance under load, memory usage, etc.
I know that benchmarks are vague, monstrous, and sometimes useless things.
It’s true also that just because some says they designed something with
speed and scalability as primary considerations it doesn’t mean they have
succeeded. Of course, it’s also true that if you have particular needs
you must meet, you should test all of the available options yourself
before committing to any one. I’d just like to see in writing somewhere
some comparisons and justifications to help me choose which library to
use. Sometimes an author knows his or her software is slow or broken in
certain situations, and never says anything because they don’t imagine it
being used in those situations. Sometimes an author knows what they have
is the best out there, and modesty keeps them from saying so.
I don’t have a remedy for this. I am not sure it is really a problem.
Just a thought I had.
–Gabriel