Re: level of detail in alt, for a "self-describing" SVG

The golden rule to follow, whenever crafting an alt text for an image, is
that it must convey the SAME information to a blind person that the image
conveys to a sighted person. Or as the WCAG SC1.1.1 puts it, it must be "a
text alternative that serves the equivalent purpose" as the image.

Creating an alt text for a complex diagram like this is not simply a matter
of stating the bits of text seen in it and just assuming that's all blind
people need to know. The diagram has a much larger purpose, and you need to
make that purpose clear.

Looking at this image, a quick glance immediately tells us that it is in
two parts, one component at the top is about a "credential", and the
component underneath seems to be about a "signature", and they are linked
(this is very important) by "proof". So you have to capture those three
things - credential, signature, and proof - in your alt text. Neither of
the proposed alternatives do that. Without them the names in the different
text items, and certainly their values, mean little.

 If this were truly a flowchart, then it would not be sufficient to state
the data items (with or without their values). You would have to describe
the flow as well, or the whole thing becomes meaningless.

However this is not a flow chart at all. It is much more like an entity
relationship diagram (not something I know much about, but the diagram
explains it). However, the same principle applies - a flowchart shows a
process, not just a list of unrelated items, and likewise this diagram
shows two components each with certain properties, and a relationship
between them.

Actually the original of this diagram (from W3C at
https://www.w3.org/TR/vc-data-model/) already has a perfectly reasonable
alt text, one that starts off by conveying the fact that there is 'a
Credential component on top connected via a proof to a Proof component on
the bottom'. Why don't you like their alt text? I agree it's somewhat long
winded, mainly due to including all the data values which, as others have
already said, is perhaps a little unnecessary in an example. And they use
the word "graph" in a sense that nobody else on Planet Earth uses it. But
Hey, this is W3C we're talking about, of course its long winded, and of
course they redefine the dictionary. What else have we come to expect!

But if you want to replace their alt text with a different one, you'ld
better improve on it. Just dumbing it down and getting rid of all the
important information, structure and relationships is not the answer. Blind
people are not stupid, so don't treat them as such. And this is a highly
technical diagram about a highly technical subject, so we can expect it to
be complex - and therefore the alt text will necessarily be complex as well
or it won't convey the information. Personally I think the author has done
rather well in the alt text for a difficult image (which is more than I can
say for the rest of the article which I would need ten years leisure time
to understand!).

There are plenty of blind people in highly technical roles, including many
on this forum. If a blind person is reading a complex document about a
highly  technical subject they work in, they don't want all the diagrams to
be replaced by a few meaningless data names that convey nothing on their
own! So start with the alt text this image already has, and go on from
there - make it clearer and more concise if you can. But don't just throw
away the important information.