Thursday, 1 July 2010

Set the ADF whale free! Um no, why ADF should be a "cost option" to truly save the whale

I had a chance at this year's ODTUG conference to throw an idea out there. It goes against the grain, but sometimes opposite ideas stir useful discussion.

From the ADF EMG, the OTN forums, user group conferences, and more, we often hear the statement "set the ADF whale free!" (okay, "whale" is my word, but it has a certain ring to it) The going idea is if the whole ADF stack was free, or at least there was a free "nobbled" version, it would assist the adoption of ADF because developers like free stuff and they would naturally start using it.

Fair enough, it's a good idea, free is good, at least, it's good for the consumer. No argument there on my part.

By the way, for the uninitiated, the Oracle JDeveloper IDE is free to use, while ADF the major JEE web framework supplied with JDeveloper, is (to coin the right phrase) "a no cost option" on OAS and the contemporary WLS platform. "No cost option" means you don't pay for it directly, instead you pay for it through your OAS/WLS license as a "bonus". This is something I've discussed before.

Anyway, what developers want is for ADF to be free. Agreed?

Yet, would that necessarily help the uptake of ADF?

Huh, what? We're talking about ADF being free aren't we?

Um, well, I've tricked you, you already agree with me. Look back and see what I said. To quote with my emphasis: "if the whole ADF stack was free ..... it would ASSIST THE ADOPTION of ADF".

Okay, so we're also talking about ADF adoption as well, not just about ADF being free.

So for sake of discussion, let's take this idea of "ADF adoption" on board, and let's put "make ADF free" aside for the moment. Further to this, let's discuss something I've noticed waaaaaaaaaay in the outback of Australia, far away from Oracle HQ, that may change your mind on how to drive ADF adoption.

When you work away from OTN, the forums, the user group conferences, and you start working with enterprise customers using Oracle, the customers' main interface to Oracle and technology is their management talking to Oracle salesmen in your local region (Australia being my case), the salesmen trying to push Oracle technologies. (And I should also comment that most of us know the emphasis Oracle puts on salesmen via Larry's recent announcements of a huge Sun hiring spree for new salesmen, while sacking technical staff)

So the 2 parts of the equation on Oracle software adoption are "management", and "Oracle salesmen".

As we know management has a large influence on the adoption of a specific technology into most enterprises. We can't discount developers in this equation, but managers are surely the biggest part.

And as we know an Oracle salesmen's job is to "sell" Oracle technology to management though all sorts of techniques. And what's the driver for Oracle sales to "SELL" something? Kind of obvious isn't it. Something that has monetary value. And we also know that most salesmen don't work for wages alone, they work for wages and commission. (okay, that's a lot of "ands", we're running with logical A-B-C progression here)

So maybe at this point our question on ADF adoption has changed. Will setting ADF free give any reason for an Oracle salesmen to "->SELL<-" ADF?

A: The going answer has to be no. They have no real reason as it's a "no cost option" on OAS/WLS that makes them no direct commission.

And before you scoff at this idea, can you really? Think about most salesmen you know. What's priority number 1 to them? Making money for themselves or doing the best for their products/employer/customers/whatever.

You know the answer to that.

So, how do we get Oracle salesmen to sell ADF.

You know it.

Oracle needs to charge directly for ADF (or maybe potentially the IDE as part and parcel).

Then, and only then will the salesmen start making commission.

And then they'll start pushing ADF along with that $1mill Exadata server to customers.

And then management will wonder what they've bought, and start pushing ADF onto their developers.

And then, well, you get it.

Free is good.

Money is better.

Maybe ADF should be a "cost option".


Sten Vesterli said...

I appreciate you trying to think outside of the box, but I think you're wrong for two reasons:

a) An Oracle salesperson can't be bothered to spent time selling a $5,800 ADF license when he could be selling a $125,000 WebCenter license. So unless you make ADF really expensive, you won't get any support from Oracle sales in pushing it

b) The purpose of making ADF free is not so that enterprise customer won't have to pay - these would want to pay for a support contract anyway. The purpose is to make ADF available to people who can't pay for a license: Students who need apps for their fraternity or sports club, developers from India, China, Indonesia and other places outside the industrialized world, etc. Having all of these people on board would grow the ADF community hugely, And that would create a talent pool for enterprise customers with support contracts to draw from.

Right now, choosing ADF involves a skills risk - you might not be able to find the required number of ADF developers. Free ADF would get more people on board.

Best regards

Sten Vesterli

robert.nix said...

I gotta disagree with you.

Oracle houses will use ADF because it's the technology for integrating with Oracle products.

The idea isn't to force people to use ADF because they've paid for it. Developers don't just want free, they want a pleasant and familiar environment in which to work. JDev isn't it. It's just different enough from Eclipse (and very different than NetBeans) and buggy enough to make it a pain to learn and use.

To get adoption, Oracle should produce free, full-featured, robust (i.e. fewer bugs than JDev) plug-ins for Eclipse and NetBeans. Full ADF design support in Eclipse alone would probably give ADF a huge boost. Developers could try it simply by downloading the plugin from the Eclipse Marketplace.

Chris Muir said...

Of course you're both right.... well at least I appreciate your point of view(s) :-) Thanks for taking time to out to post a reply.