Re: SVG12: playbackOrder design

Hi Bjoern,

Bjoern Hoehrmann wrote:

>Dear Scalable Vector Graphics Working Group,
>  In the
>playbackOrder attribute seems ill-designed. The default attribute is
>defined as "all" which requires implementations to assume that seeking
>backwards is possible even when there is evidence in the document that
>this is unreasonable (mutation events are a good indicator).

>  Please
>change the draft such that the default value for the attribute is a new
>value like "auto" which allows implementations to behave in a more
>reasonable manner..

The author is able to use the 'playbackOrder' feature when they create 
documents that are not designed to be rewound (seeking backwards). For 
instance heavily scripted content such as most web apps, games etc... 
should probably set playbackOrder to forwardOnly. Some complex long 
running animations may allow the user to 'rewind', seek backwards and 
play content again, hence playbackOrder will be unspecified or 'all'. 
Other complex animations may use the 'discard' element to reduce the 
memory required by the UA, these presentations are designed to play in 
the forward direction only, hence the indication to the user agent by 
specifying 'forwardOnly' that disables controls allowing backwards 
navigation of the timeline.

The 'all' value indicates that no specific restrictions are placed on 
document timeline seeking. This is the same behaviour found in SVG 1.1.

Are you suggesting a new 'auto' value that allows the UA to determine 
whether to allow seeking backwards? If this is what you mean, I think 
this is probably not a good idea.

Or are you suggesting the current 'all' value (default behaviour from 
SVG 1.1 no restrictions) should be renamed 'auto'?

>The draft further states "Similarly the UA should disable any controls
>it may provide in the UI for seeking backwards."
Yes, we think the UA should do this if playbackOrder = "forwardOnly". 
What is the problem with this statement?

> Please change the draft
>to include a note that implementations may offer facilities to allow
>effectively reloading the document fragment which is unaffected by this
Yes, I agree that the specification should clarify that the document can 
be reloaded regardless of the value for playbackOrder.


Received on Tuesday, 26 April 2005 05:45:56 UTC