Specify deprecation model

[nigelmegitt]

Netflix has requested that IMSC 1.1 should indicate that when TTML2 features are used in the same document at the same time as their non-TTML2 equivalent and deprecated features [from IMSC 1.0.1] the TTML2 features prevail.

The motivation is "this ensures that future versions of IMSC can effectively remove the features marked as deprecated".

[palemieux]

See https://lists.w3.org/Archives/Public/public-tt/2017Nov/0107.html

[palemieux]

The request above is not easily achievable in many cases, e.g. if condition replaces itts:forcedDisplay.

[nigelmegitt]

Thanks for the reference @palemieux

[nigelmegitt]

The request above is not easily achievable in many cases, e.g. if condition replaces itts:forcedDisplay.

You mean that it is complex to do in the spec and might require additional authorial support?

An authorial mechanism would be a construct or pattern that checks for feature support for condition and if present, uses it to exclude the itts:forcedDisplay style and instead to make a further test of the forced parameter, and then use the outcome of that test to set tts:visibility.

[palemieux]

It is too complex for processors to detect the presence of both, and apply precedence rules.

[nigelmegitt]

@palemieux that would be a comment worth making on w3c/imsc#279

[skynavga]

@palemieux your statement seems wholly subjective and without factual support

[palemieux]

@skynavga Per w3c/ttml2#482 (comment), itts:forcedDisplay would be deprecated in favor of syntax along the lines of the following:

<style xml:id="forcedStyleFalse" condition="parameter(forced)==false" tts:visibility="false"/>
<style xml:id="forcedStyleTrue" condition="parameter(forced)==true" tts:visibility="true"/>
<style xml:id="forcedStyle" style="forcedStyleFalse,forcedStyleTrue"/>
...

<p style="forcedStyle">This text is only visible if forced==true</p>

so how would a processor enforce precedence rules when encountering the following:

<p itts:forcedDisplay="true" style="forcedStyle">This text is only visible if forced==true</p>

[nigelmegitt]

The deprecation model needs to define what deprecation should mean in this context. I think from our meetings that we are all agreed that deprecated features are permitted but are expected to be prohibited by future versions of the specification.