Change Proposal for ISSUE-127
- There is no "goal to make link relations handling inside HTML consistent with link relations in other contexts". There is no particular benefit or use case to making link relations consistent between Atom and HTML. The two technologies are distinct and do not share link handling logic, even when a single software package implements both.
- The proposal claims to simplify the spec by removing a degree of freedom, but in practice it merely trades one degree of freedom (the ability for a relationship on <link> to be defined differently than on <a>) for two others (the ability for a link relation to be allowed on <area> but not on <a>, and the ability for a link relation to have an effect on an element but not be allowed on that element).
- The ability for a link relationship to be allowed on <area> but not <a> has no use case and would be more confusing than a link relationship having different meaning on <link> and <a>, since <a> and <area> are interchangeable in a way that <link> and <a> are not.
- A link relationship having a defined effect is not something that should be encouraged. If it is needed, it would be better to handle it as a special case in prose, since that would not give the appearance of it being a regular expected occurrence.
- Fundamentally, the change here is editorial, and should be left up to editorial discretion.