How much positioning control for captions etc.?

This relates to the issue

    3.PR#271: Checkpoint 4.7: Change to P2 since arbitrary repositioning 
not a requirement.
      http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#271

See the telecon minutes for the following points:

Action:

AL: Al will post to the list a discussion of this topic to show that we 
still need to work on wording

JG:
1)this is priority 1
2) we do need to allow user to choose whether info is obscured or not
3) functional requirements - how much configuration is needed.

Summary:

The limits on magnification have to do with the legibility of the text.
The caption is treated here as useless if you magnify to the point beyond
where one word fits on the screen.  But that may be a lot of magnification,
so the relative placement of text and picture needs to be pretty fine.  Not
clear we can lay in a requirement which is better for implementers than "to
screen resolution."

Discussion:

Note: "Not to obscure" does not describe the critical threshold of
feasibility for P1 analysis.  The user should be able to trade some
obscuring to make caption and video simultaneously visible even into a
placement which produces some degradation / partial blockage.  Concurrent
presentation even of partial video and caption has a higher priority than
complete non-interference between the different media modes.

The scenario that I see as disclosing the critical functional threshold is
a user who needs magnification such that only one word fits but there is
room to see some of the video at the same time.  The user need the ability
to position the text relative to its video partner in screen vertical and
horizontal dimensions with a resolution on the order of a few letters in
the text or some small-integer fraction of a word.  One character is safe.
Maybe we should say one EM of the actual font size after application of
user font controls.  Note that the user may wish to view the video very
microscopically, so this gets down to a few pixels on election of small
font and large magnification.  

So maybe I am convincing myself that arbitrary positioning of the caption
layout region relative to its cognitive partners in the concurrent display
is required within the framework of enabling this scenario.

The maximum usable magnification / zoom of the video content is
document-specific.  There is no way for the User Agent to know this in
advance per media class i.e. video.

It is possible that people working with low-vision customers may know a
rule of thumb for a magnification beyond which the technique is generally
useless.  I don't know any such value.

So there isn't much relief on the "minimum" front because it sounds like
the capability is useful up to the following extreme capability:

One high-contrast styling [in the presentation object for streaming text]
is available, scalable, with clear background as an option; and this is
placeable relative to the other content at an arbitrary X and Y location
and at a Z-index closer to the user than the other content.  The
positioning does not necessarily have to be all the way down to screen (1
pixel) resolution, but it does have to get down to the order of the 'ex'
unit of the smallest available font scaling.  Thus it _is_ nearly that
fine; and there may be no acceptable resolution less than screen resolution
that makes the implementation meaningfully easier.


If the player doesn't provide this much control, it seems that a tough
situation is converted to an impossible situation.  So that sounds like
near-arbitrary X,Y positioning is required.  Not necessarily in conjunction
with arbitrary restyling.

Al

Received on Thursday, 20 April 2000 17:23:44 UTC