[selectors3] better terminology for ~ sibling combinator?

[dbaron]

A recent edit changed the terminology for sibling selectors:

"Adjacent Sibling Combinator" -> "Next-sibling Combinator"
"General Sibling Combinator" -> "Following-sibling Combinator"

I'm a bit hesitant about this change, because it doesn't seem obvious to me that "following" means "any later sibling". I think I occasionally hear the term "following sibling" as a synonym for "next sibling" rather than meaning "any later sibling".

I wonder if there are better options here for the latter combinator. Maybe "Later-sibling combinator" would be better -- although I'm still not sure that it fully disambiguates.

[inoas]

I know it is a bit more verbose but for sure clear?

Next-Immediate-Sibling Combinator
Any-Following-Sibling Combinator

[bkardell]

Maybe just pluralizing helps?
”following siblings” sounds less misleading than ”following sibling”

[inoas]

Immediate-Sibling Combinator vs Following-Siblings Combinator maybe?
Forget it... without next the direction is not specified.

[Loirooriol]

@bkardell The problem with plural is that it doesn't work when reading right-to-left, which is usually the natural way to read selectors. For example, a ~ b > c would be

Select any c element
that is a child of a b element
that is a following sibling [can't use plural!] of an a element.

"Following" is confusing indeed. What about "posterior" or "subsequent"?

• Next Sibling Combinator
• Posterior Sibling Combinator

[SelenIT]

I suppose that it can be read as "...that is one of the following siblings of..." or "that belongs to the set of the following siblings of...". So pluralization is not such a big problem. But the word "subsequent" looks good to me.

[tabatkins]

Tangent: you have an interesting notion of "recent", @dbaron, since that edit was from 2012. ^_^

@dbaron: Are you okay with @SelenIT's suggestion of subsequent? I think it suffers from the same problem as "following", but it's my favorite suggested so far, at least. ^_^

I'm unsure that any single word would solve your concern 100% unambiguously; they can all be interpreted as "the immediate sibling" as well. We'd have to go with multiple words to be fully unambiguous, I think.

That said, I do think the words we are using now are a definite improvement over the pre-2012 terminology of "adjacent" and "general", which don't convey directionality.

[dbaron]

Tangent: you have an interesting notion of "recent", @dbaron, since that edit was from 2012. ^_^

Ah, it was an old edit, recently brought to the working group for republication of the selectors3 REC.

I guess I'm OK with "subsequent", although it's not clearly better than "later" to me, and it goes against Orwell's advice (see the "Pretentious Diction" section).

[tabatkins]

I agree that it's not better than later/following/etc, and also that it's a complicated word for no good reason. ^_^

[AmeliaBR]

I agree with the original concern that "following" can be too-easily read as "immediately following". But not sure "subsequent" is an improvement. I also suspect that "later" could be confusing, because it properly refers to time relationships and we're talking about spatial relationships.

From a plain-language perspective, my vote would be for "Any-Following-Sibling Combinator".

These are definitions, not syntax that people actually have to type on a regular basis, so clarity should win out over compactness.

[svgeesus]

I'm pleased to see this terminology change getting wide review, since that is one of the requirements for transition to a new (no substantive change) Recommendation in Process 2017
https://www.w3.org/2017/Process-20170301/#rec-modify

[tabatkins]

(It's an editorial change, adjusting the non-API-exposed name of a feature in a spec. It's not substantive - it is impossible to cause any change, positive or negative, in impl support as a result.)

[svgeesus]

Yes, that is what "no substantive change" means :)