More SQL precompiler

["Jon Paris" <Jon.Paris@xxxxxxxxxxxxxx>]

<soapbox>

The basic problem with the current pre-compilers is that they require the
pre-compiler to have a knowledge of the host language.  The failure of the
pre-compiler to keep up fully with _any_ of the host languages is a direct
result of a lack of funding in this area.  What monies were available have
been poured into trying to update the database to the latest UDB levels.
That being a moving target, there is still quite a way to go, as Bruce
Hoffman for one will be only to happy to explain to anyone who wants to
know.

Recently funding was made available for some changes in the pre-compiler
technology - in particular to address the significant shortcomings in RPG IV
support.  Kent Milligan at IBM was a significant force in helping to make
this happen.  Sadly it took so long to get to this stage that the proposed
architecture (which would have free the pre-compiler from knowing anything
about the language) was no longer adequate.  The pre-compiler folks
(mistakenly) believed that adopting the new approach would have moved from
an "SQL must understand RPG" position to an "RPG must understand SQL"
position.  In fact this need not have been the case but it would have
required significant architectural work be done to ensure that the
components could all "talk" together.  The decision was made that that would
take too long and since there was no additional funding available ....... a
far more short-sighted (and in my opinion ultimately more expensive)
approach was adopted.

In essence - we have not as a group complained enough about the hopelessly
inadequate support for embedded SQL in the host languages.  You could add to
that that we haven't complained enough about manual examples only written in
C and Java, indecipherable documentation (even if you can find it) etc. etc.
I think we focused for too long on the "my SQL isn't as big as yours"
syndrome and did not concern ourselves sufficiently with the inadequacies of
the tools etc.  Rob Berendt, John Carr and Bruce Hoffman have all done a
great job raising DCRs, PMRs etc.  It is time the rest of us started making
a nuisance of ourselves where it matters - which frankly in not in the forum
but in official change requests to IBM.

If we don't tell IBM what we want in a fashion that is meaningful to them
(i.e. that affects their management performance ratings) then we are partly
to blame for the situation.  It is a bit like getting the politicians we
deserve.

Thus endeth the first lesson <grin>

</soapbox>

Jon Paris
Partner400