As a followup to my post starting
> The scenario where I need to use HTML 5 Database storage requires
storing a tree of data....
My particular need can be met by generating identifiers in software,
rather than having the database do so. This also eliminates the need
to chain database calls.
However, there are related scenarios where that strategy doesn't work,
including most scenarios where what gets written to the database
depends in more complicated ways on what has been written.
And even when there's no need to chain database calls, if one has
several outstanding calls to the database, it would be very helpful to
know which call triggered a callback. One example is issuing several
database calls for data to update several different regions on a page
-- when the data is returned, it will not indicate _which_ region on
the page is to be updated (unless extra effort is made, such as
storing that info in the database, which would pollute it.
So, I still think there is a need to pass a variable from the code
calling the database, to the callback. JavaScript has closures, which
would do the job in many cases, but other languages don't.
