service worker client definition is redundant

[isonmad]

8b483b0 changed the definition of a service worker client to "a type of <a>environment</a> or <a>environment settings object</a>" which is redundant, since every environment settings object is itself also an environment, so this confusingly implies they don't have an is-a relationship. Also, the 'type of' wording from #832 (comment) c11942b seems redundant and potentially confusing, isn't it more straightforward to just say "a service worker client is an environment [that...]"?

[jungkees]

isn't it more straightforward to just say "a service worker client is an environment [that...]"?

I'm not sure if we can assume that kind of a polymorphic behavior in the spec language.

I'm actually curious if we can make a service worker client just be a type alias to the environment (i.e the base type) and use it to dynamically reference the states depending on the actual type (either environment or environment settings object) in runtime. @annevk, @domenic, would it make sense?

[jungkees]

Note that we have some places in the algorithms where the logic branches depending on the types: 4ce01b5#diff-27b79860afe28f01aed4f1f6228367faR1278. (In this particular case, an environment doesn't make sense for checking of the secure context.) To deal with this sort of requirements with polymorphic approach, we'd have to define the algorithms on the base type and override them on the sub type (assuming the polymorphic behavior works for the specs in the first place.)

[isonmad]

I'm not sure if we can assume that kind of a polymorphic behavior in the spec language.

Like @annevk said in #1045 (comment) that's already endemic to the specs.

To deal with this sort of requirements with polymorphic approach, we'd have to define the algorithms on the base type and override them on the sub type

Except you don't really need virtual function esque overrides to do that at all. Just do the equivalent of if (x instanceof SubType) {...} else {/* must be SuperType */...}. As long as you put the if-else-conditions in order from most-specialized to most-general types (which isn't the case right now, which is #1045), it works like you'd expect.

[jungkees]

Just discussed with @domenic in the f2f. We agreed what @isonmad suggested in the above comment. I'll try to update the spec by using the base type with the most-specialized-to-most-general-types language.

[jungkees]

Closed by b67c445. Thanks @isonmad! I'd like to add you to acknowledgements if you're okay. Can you provide your name if so?

[annevk]

@jungkees we credit isonmad as "isonmad" in various WHATWG standards. I think that's their preference.

[jungkees]

I added "isonmad" to the acknowledgements. Thanks for help!