26703 2014-09-01 18:28:23 +0000 In case an algorithm applies conditionally, the conditions should be defined in some kind precondition step 2016-03-16 18:00:55 +0000 1 1 1 Unclassified WHATWG HTML unspecified PC Linux RESOLVED WORKSFORME blocked awaiting response from smaug to comment 1 (please remove this when replying) P2 normal Unsorted 1 bugs ian annevk ian mike contributor oldest_to_newest 110826 0 bugs 2014-09-01 18:28:23 +0000 As an example, when one follows http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#processing-model-4 it gives no clues that the algorithm might not apply to some input element types. In order to know that, one needs to look at the specific type, for example http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#hidden-state-%28type=hidden%29 and see that 'autocomplete' is listed as 'do not apply'. Then one needs to find the definition for that http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#do-not-apply and from there to actually understand what the whole thing means, http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#common-input-element-attributes To get better readability, I suggest having a precondition step in each algorithm which applies conditionally. So, before "1. If the element has no autocomplete attribute, then jump to the step labeled default." there would be something like Precondition: abort this algorithm immediately if the following preconditions don't hold: 'autocomplete' attribute <a href="http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#common-input-element-attributes">applies to</a> the element Precondition should be used consistently, so that whenever it isn't there, reader would know that there are no preconditions. I can see how the current spec is easier for the spec author, but it certainly is rather hard for follow for the reader. And as with code, specs should favor readers. 111755 1 ian 2014-09-18 17:31:50 +0000 The specific section cited here now starts with an explicit statement that it's only for <input type>s that apply. Are there any other cases that are confusing? 125542 2 annevk 2016-03-16 18:00:55 +0000 Olli, if you still find something confusing, please reopen or file an issue on GitHub. Thanks!