WebIDL-1 reported as WD about specref HOT 6 CLOSED

Yeah, WebIDL-1 redirects to WebIDL as of ee2359a.

This is tied to the conversation about leveled specs lead by @deniak.

What's the use-case for linking to WebIDL-1?

from specref.

There seems to be two issues in one.

The first issue is that the status reported by specref for WebIDL is incorrect: Specref says the document published on 15 December 2016 at https://www.w3.org/TR/WebIDL-1/ is a WD, whereas it is a REC. Shouldn't [WebIDL] report the same info as [WebIDL-20161215], the latest published spec?
http://www.specref.org/?q=webidl-20161215

The second issue is the ability to specifically link to WebIDL-1 or WebIDL-2 instead of just WebIDL. Two use cases come to mind for specs that are to be published as PR/REC:

1. Linking to WebIDL-1 to assert that all IDL concepts used in the spec are stable to ease Director's transitions.
2. Linking to WebIDL-2 to point out specific terms that are not yet "stable".

One concrete example: the Presentation API uses NotAllowedError, which is defined in WebIDL Level 2. I'd like some way to say somethink like "all IDL terms are defined in [WebIDL-1], except NotAllowedError, defined in [WebIDL-2]". Or "all IDL terms are defined in [WebIDL]. At the time of publication, NotAllowedError is defined in the ED of [WebIDL-2] and considered stable".

My understanding is that the leveled spec proposals could make "/TR/webIDL/" link to the latest level (i.e. the Level 2 WD here, once published), but "/TR/webIDL-1" would still point to "WebIDL Level 1".

@deniak, proposals do not seem to say anything about which shortname PR/REC specs should use for references. What is the expectation?

from specref.

Just to clarify, the current proposal actually has 3 options and we haven't decided yet whether /TR/shortname should point to the latest level or the most stable version.

Regarding WebIDL, I think there's a small problem. I confirmed with @ylafon and at some point, there was a shortname change from webidl to webidl-1 (see WebIDL-1 history) so technically, /TR/WebIDL should point to /TR/WebIDL-1. If we agree on this, I will fix the error.

from specref.

@deniak this is precisely why it's key to come up with a consistant story for spec levels and enforce it. Levels should be tags in Git parlance (branches when you need to run things in parallel for a little while). WD's and the like should just be tags along the way.

There is something weird going on with WebIDL, though, as /TR/WebIDL-1 points to the December 15, 2016 Rec, and /TR/WebIDL points to the April 19, 2012 CR.

On the Specref side, I'm experimenting with treating WebIDL as a "leveled" spec. Every version is stored under the WebIDL shortname, and WebIDL-1 points to WebIDL-20161215.

from specref.

@tobie it should now be fixed. As per @ylafon's request, /TR/WebIDL now redirects (HTTP 307) to /TR/WebIDL-1

from specref.

Note that it was a request from the chairs until we figure out what to do with /TR/WebIDL (and have the edcopy published as WDs).
Note the use of a temporary redirect so that we can change what it redirects to (i.e.: no hard choices were made)

from specref.