The next stage that we will need to add some day to the mach try perf tool is the actualy task information. At this stage, we are making educated guesses about if tasks should exist in the given category. We need to add a verification method for this which would involve running fzf on the categories at some point. It's unclear if this would be possible though because it could be slow.

We should see if we can make use of an attribute field to denote the categories a set of tests is contained in. This could be an approach we could use to cross-reference the selected tasks with those that actually exist.

I've also been wondering if we could use an attribute directly to set what tasks exist in a category and go from there. I don't believe this would get rid of the decision matrix that we use because we still need to be able to provide the users with clear names of category combinations (e.g. Pageload linux, Pageload desktop). If the tasks would be available on platforms that we don't want developers using, then that would defeat the purpose of the categories because we would still need to perform some set of fzf queries at some point, or some filtering somewhere. So having an attribute might not provide any benefit, but it's something to look into a bit more.