November 2007

Oracle extends the scope of its PL/SQL compiler optimisation. November 2007

I just came back from a few days at OOW, and it was fun. I shared a biggish apartment with Anjo Kolk, Krister (Sweden) and Oliver (Danish CSC), and it was beautiful to see the beer bottles (good beers, mind you!) gradually filling up the kitchen table allocated for that purpose.

On Sunday, November 11, I was invited to a seven-hour briefing for Oracle ACE Directors (I am such a thing). It was mostly about the Fusion Middle Ware (MW) and in the end a bit about the 11g database.

I have three observations:

1. During the MW presentations I saw more acronyms than in my entire military career.

2. 'Oracle' was the only word with less than seven letters in all those slides.

3. The best thing that can happen to any product is to be bought by Oracle. Turns out, that the purchase itself will transfer the product overnight from being worth-, use- and hopeless to being an absolutely state-of-the-art, best-of-beer product.

Couple of years ago an interesting fact floated up in Oracle-L – a regular user with only SELECT privilege on a table can successfully execute a SELECT FOR UPDATE against it, locking all rows and even lock the whole table using LOCK TABLE command. Locking a table in exclusive mode would stall all changes and selects against that table – effectively hanging all applications using that table. Pete Finnigan wrote a review of the issue in this blog entry.

Couple of years ago an interesting fact floated up in Oracle-L – a regular user with only SELECT privilege on a table can successfully execute a SELECT FOR UPDATE against it, locking all rows and even lock the whole table using LOCK TABLE command. Locking a table in exclusive mode would stall all changes and selects against that table – effectively hanging all applications using that table. Pete Finnigan wrote a review of the issue in this blog entry.