A proposal for figures

Jonathan COHN

Gene has started a lengthly discussion whose main complaint appears to be with figures…

I just looked at a couple of articles on the New York Times.
Essentially, it looks like the pictures are now composed in the following way:
Picture with no alt description
Text description of picture

This is at least better than a year ago where the picture had the same information in its "alt" tag as the figure description text. The major problem with this is that as screen readers instead of
"graphic Los Angeles"
You get
Start Figure Los Angeles End Figure"
That is unless you are running MacOs VoiceOver with Grouping Navigation.
So for this simple case, where there is an unlabeled graphic contained in a Figure with a Figure caption, it would probably make sense to not use start and end tag reading.

Would this adversely effect performance?
What I am thinking is if
a) Graphic within figure does not have a label
B) Graphic with in figure has alt attribute that matches the Figure caption
Read Graphic with figure caption.
when encountering start figure marker Read it.

One thing I am not sure about off hand is if figure tags have labels or are just automatically named by the figure caption.