Post by Priya Lakshmi MuthuswamyHi,
Kindly review the fix for
https://bugs.openjdk.java.net/browse/JDK-8184205
webrev : http://cr.openjdk.java.net/~pmuthuswamy/8184205/webrev.00/
Thanks,
Priya
Priya,
I've spent a while looking at this, both the code, and the output
generated by
a build of JDK including your patch.
Although I now understand -what- you've done, I don't understand -why- you
have done it this way: you seem to have done both more than I would have
expected and less that I might have expected, meaning that the design and
the code seems to be in a somewhat mixed/confused state.
This is definitely more than a minimal solution, because you've changed the
element structure on all tables, and not just on the tables that maybe
needed
it. I'm not saying that's inherently bad/wrong, but it does seem strange at
first sight to have <figure> nodes just containing a single <table> node --
at least for all the tables that do not have active tabs.
Conversely, you have started into the ARIA world, but you are not applying
some of the possibly obvious ARIA attributes. The most obvious one I've
come across so far is "aria-controls". The attributes for tabs and
tabpanels
look enticing but I'm not sure they apply to our tabs and tables.
For reference, I'm looking at https://www.w3.org/TR/wai-aria-1.1/
In terms of a mixed design, although you have changed it so that you
always embed tables in figures, it's now the case that if its a singleton
tab, it appears as an explicit <caption> (as before) but if there are
multiple
tables, they appear as a row of boxes inside the figure but outside the
table.
Is that really the best way to organize this information?
Given that the summary tables are such a fundamental part of the
presentation of the API, I think it would be good to see a discussion
and/or design document on the alternatives and their relative merits,
with a possible focus on where we would like to end up, even if we
don't get there right away.
-- Jon