Re: the IN club: Oracle "unpublished" information

On 3/15/07, Niall Litchfield <niall.litchfield_at_gmail.com> wrote:
> In addition, Oracle is not an OpenSource company, they do have internal
> intellectual property to protect.

But nowadays you don't license Oracle because it has "the most clever algorithms" - you license it because those clever algorithms are efficiently implemented AND because Oracle is very scalable AND because the company is so huge that it won't go away in the next 20 years AND because it's the market leader AND ...

Publishing the algorithms will simply enable (as you said in the parts I've chopped away) professionals to better design and tune systems, so increasing the market share. And keen enthusiastic professionals may even point out problems (bugs), suggest refinements and enhancements FOR FREE. An example of "peer review" in essence.

If I were a big competitor, and wanted to copy Oracle algorithms, I'd have all the resources necessary to reverse-engineer the CBO, or simply, I could hire a former CBO developer and have him (secretly!) tell me what the algorithms are. Only to discover they are patented maybe ..

After all, the multiversion read consistency algorithm has been publicly known for years ... and still it took MS more than a decade to "parrot" it.

So my point is - publishing the algorithms would be a very good move for Oracle, with no real downsides at all. I concur with Charles 100%.

-- 
Alberto Dell'Era
"Per aspera ad astra"
--
http://www.freelists.org/webpage/oracle-l