Re: Publishing WebVTT as a Proposed Recommendation (CfC)

From: Gary Katsevman <me@gkatsev.com<mailto:me@gkatsev.com>>
Date: Thursday, 13 June 2019 at 06:32
To: Pierre-Anthony Lemieux <pal@sandflow.com<mailto:pal@sandflow.com>>
Cc: Cyril Concolato <cconcolato@netflix.com<mailto:cconcolato@netflix.com>>, Philippe Le Hégaret <plh@w3.org<mailto:plh@w3.org>>, Nigel Megitt <nigel.megitt@bbc.co.uk<mailto:nigel.megitt@bbc.co.uk>>, TTWG <public-tt@w3.org<mailto:public-tt@w3.org>>
Subject: Re: Publishing WebVTT as a Proposed Recommendation (CfC)

On Tue, Jun 11, 2019 at 4:15 PM Pierre-Anthony Lemieux <pal@sandflow.com<mailto:pal@sandflow.com>> wrote:
"Text combine upright: These are all implementation issues. The
text-combine-upright CSS property hasn't been whitelisted by them."

Does it mean that there is not two implementations that pass the test?

text-combine-upright is made available to WebVTT via the CSS extension "feature". Currently, no browser allows this CSS property but I have a proof-of-concept in vtt.js. Given that other CSS properties are allowed and available in multiple implementation, I think this is an implementation issue and should not block the spec.

In general non-implementation of a feature very much does block progression to PR. That’s what the CR exit criteria say. If text-combine-upright is a specification feature and we cannot demonstrate its implementation in two independent implementations, it cannot go through as a normative provision in PR.

Is this feature already present in another Rec so that we can effectively include it by reference and know it has been through the W3C process? We have historically only tested incremental additions, but in this case the CR exit criteria do not make any provision for this – they imply that every feature must have two implementations that pass reasonable tests.