“Compatible” products are never perceived as zero-risk, zero-harm. So there’s a lot of FUD about them, led by the most knowledgeable internal experts.

A few web-centric companies may and do just say “Ixnay on the Oracle pricing model, period, and partial compatibility is better than none at all.” But to my knowledge, “few” is the operative word.

Existing Oracle customers were not interested, I think because having made everything work with Oracle, the promise of lower prices wasn’t enough to get them to try to switch to something new and unknown. Oracle is “sticky”. DBA’s know it. Oracle DBA’s heavily influence the purchase decision, and obviously have a huge interest in keeping Oracle, for their own job security as well as their own technical comfort.

For someone who does not already run Oracle, being Oracle-compatible is not that much of a benefit.

Does that sound right?

I agree that data mart outsourcers are a market one sells to directly, even if vendors conceive of them as an indirect channel.

Dataupia sold directly – see Subex and ITIS press releases.

1. Those that explicitly imitate a market-leading platform product.

2. Those whose main selling point is that it runs over multiple market-leading platforms.

In the 1980s, there were some huge successes selling software TO HARDWARE VENDORS that ran on multiple hardware products — Microsoft and Oracle rose to prominence that way, and Ingres was in the game too. But end-users have rarely cared much.

As for straight emulation — I can’t think of a significant success since Amdahl emulated IBM’s mainframes. Just recently, EnterpriseDB was the “Oracle-compatible database company” — now it’s licensing off its Oracle compatibility technology and competing on a different basis. ParAccel’s “Amigo” mode hasn’t amounted to much. And so on.

Could you elaborate a bit on your first point? I’m not sure exactly what you mean with “transparency/emulation features.”

I wholeheartedly agree with your second point. It’s very consistent with our experience.

-EE