Oracle Observations

December 10, 2012

ukoug conference report: Dave Ensor

Filed under: Uncategorized — bigdaveroberts @ 9:53 pm

I only became aware of Mr Ensor when I read his chapter in the book: Oracle insights – tales of the Oak table.(


The chapter lists and explains the significance of the major new features in each release of Oracle.

I was at the time working on the support of two oracle based applications, which at the time presented me with baffling architectural eccentricities.

One was a document management system where the client application didn’t access the database directly, but rather made calls to an API implemented on the middle tier that had several daemon processes that would execute the required SQL on your behalf. I suspected that the application had been designed in that manner to better facilitate working with multiple database back ends.

The other held large chunks of PL/SQL in Long columns in the database. It then included an OCI*C daemon process that was passed messages via a DBMS_PIPE would then extract the appropriate PL/SQL code, perform parameter substitution, execute the PL/SQL and return the results via a table. This was again baffling to me, why not use stored procedures?

Both of these architectural mysteries were resolved by Dave Ensors chapter. While both applications ran on Oracle8i, it was apparent that their architecture was much older.

In Oracle 5, due to memory address limitations it was only possible to have 40 concurrent contentions to the database. Thus to support more users, the user SQL had to be multiplexed across a limited number of connections.

Similarly in Oracle 6 PL/SQL was introduced, but without a standard way of storing PL/SQL in the database. (Procedures and functions wouldn’t come till Oracle 7.)

Thus Mr Ensor explained to me the architecture of the applications I was supporting.

Dave Ensor attended last years Oak table Sunday at the UKOUG conference and participated in a panel discussion there. He also visited All bar One afterwords, but only briefly, and I missed my opportunity to thank him for the insight he had given me.

This year he didn’t attend either the conference or the Pre-conference OAK table Sunday event.

However he did visit All bar One and I had my opportunity to thank him for his contribution to TFTOT.

From his recollection, I was the first person to ever thank him for that specific piece of work, he had been asked to produce 20 pages, produced 120 and was then negotiated down to 70ish.

I suppose writing about older versions of Oracle isn’t necessarily sexy and on that basis this commensurate piece of work hasn’t received due recognition however if you support an application with a long history, you may do well by getting hold of a copy of this book and reading his chapter!

NB. the book also includes other excellent contributions and chapters!


Create a free website or blog at