I don't feel great about explicitly saying that procedural tests have reduced inter-rater reliability. I don't think that's necessarily true, and feel it's too early for us to give up on inter-rater reliability for process testing. Lets first figure out if we can make it work with a high degree of inter-rater reliability. To address this, remove the following phrase: "which higher than that of a procedural test".
I don't think "Text has minimum contrast" is an unconditional test. It's certainly not a great example of it. It relies on the "purely decorative" and "human language" definitions, which have a good deal of grey area. In ACT lingo this isn't an objective rule. Very few rules actually are. I think this needs to be replaced with the "HTML page has non-empty title" rule.
I don't feel great about explicitly saying that procedural tests have reduced inter-rater reliability. I don't think that's necessarily true, and feel it's too early for us to give up on inter-rater reliability for process testing. Lets first figure out if we can make it work with a high degree of inter-rater reliability. To address this, remove the following phrase: "which higher than that of a procedural test".
I don't think "Text has minimum contrast" is an unconditional test. It's certainly not a great example of it. It relies on the "purely decorative" and "human language" definitions, which have a good deal of grey area. In ACT lingo this isn't an objective rule. Very few rules actually are. I think this needs to be replaced with the "HTML page has non-empty title" rule.
Source: @WilcoFiers , survey