At 05:08 PM 12/1/00 -0700, Jason Wharton wrote:

>> That mechanism allowed the very clever (ambitious? overly
>> eager? utterly insane?) programmer to write UDF's that were
>> data type independent. There has been no great clamor
>> for datatype independence from the user base, but there
>> is frequent complaint about the absence of null handling.
>
>I recall some clamor and I would like to add to the clamor for UDF's
>becoming based upon descriptors.
>
>I have not bothered doing anything with UDF's because they are essentially
>worthless to me if I cannot consider nulls and have some flexibility in
>regard to data types.
>

I don't know the current state, but the original solution to this
problem (and lots of others) was the "missing value" -- an in-band
value designated to represent the "null" value. The "missing value"
was not alternative semantics for null, but a convenience for languages
in which in was awkward to carry about a separate null flag.

Using missing values, the trigger could check for the designed value
to detect a null and set a field to the designed value to indicate
a null. Simple, yes?

I'm sure the JRD code is still there -- <flame> Borland Marketing
Bozos who screwed up the product were too dim to remove the code.
</flame>. So unless some misguided sole went out of his way to
make a mess of things, probably all that is needed to hack the
missing value clause into SQL DDL.



Jim Starkey