In the OpenACS 4.0 Guidelines, there are a couple of aD questions, including

• what's with the disabled constraints, followed by index creation, followed by constraint reenablement?

Please note...I'm NOT an aD employee (nor have I ever been, though I almost was...), but just looking at that, I can guess. Sounds like the SQL was generated a DDL creation tool. I use ERwin and SQL*Navigator at work, and get those kinds of disabled/mucked with/re- enabled constraints sometimes. It could be one of those, or even a perl script for all I know (there's supposedly a pretty cool DDL:Oracle module that I haven't tried, for example). But that's what it looks like to me...

I suppose someone may've thought it would be useful for cases where
you copy in tons of data directly into the objects table.  In this
case, though, the cost of dropping and rebuilding the index isn't
really that big a deal.

The code certainly doesn't hurt, but unless the ACS were to start
disabling and reenabling the unique constraint on a frequent basis it
doesn't help, either.

It may actually slow things down a bit.  When a unique index is used
to enforce the constraint, Oracle can throw a constraint error based
on whether or not the insertion into the index succeeds.  Using a
non-unique index to enforce the constraint means that the check has to
be done on the index after insertion.  Of course, the insertion code
may flag dupe entries anyway specifically to optimize this case.  But
then again it may not.