Since a new year beckons, and this is a blog, it follows that predictions for the new year must follow. Hey, rules are rules. And so, with no further adieu, here are my predictions for 2007:

Easy Calls:

1) Microsoft Vista comes, new machines get it, nobody on XP upgrades. Microsoft will be content to turn over their installed base with new machines sold and old machines obsoleted. Otherwise, upgrades will be slower than currently forecast.

2) Apple delivers Leopard, new machines get it, but upgrades are slower than currently forecast. Tiger is sufficiently good, and external packages (e.g VirtueDesktops) simulate a lot of the new functionality. Apple has a big decision to make, between adding functionality for the IPod/digital home, and functionality for business and enterprise solutions.

3) Linux will continue plodding progress in '07, and boom in '08. Linux has a lot to recommend it - including good security, scalability, and cost. Its openness makes it impervious to horrible, enterprise-inappropriate additions (DRM), and has a huge leg up on 64-bit solutions. If Ubuntu can continue advancement of a plausible desktop system, Linux will start taking Mac users in '07 and MS users in '08.

4) Dynamic languages will keep making inroads on "traditional" (.Net and J2EE) development. Compared to compiled C++, Ruby is one slow language in execution, but there are *so* many available processing cycles that development savings dominate. It won't run Google or the would-be Googles, but it's fast enough for everybody else. The second wave of RoR apps will appear.

5) Ajax is here to stay, and full page refreshes will disappear in 2007. Good packages (such as Prototype), and the ease of integration in new environments (Ruby on Rails) make ajaxification easy for new applications. Existing apps will ajaxify or be replaced by ajax'd apps.

6) Google backlash begins. Past a certain size, it gets really hard to not "be evil." Google has passed that size, and the leading edge isn't willing to give them a pass anymore. Look for one new search engine to break on top of the pack.

Wild Swags:

7) Oracle buys SAP. Alexander the Great "wept, because he had no more worlds to conquer." Larry Ellison does not weep, and SAP is all that is left. The current permissive antitrust environment won't last forever, so Oracle goes Exxon-Mobil(e) in '07.

8) Indian IT giants start buying second- and third-tier US software providers. Outsourced development from the US to India is now commonplace; the next logical step is Indian vertical integration into the American market. Wilder swag: Wipro buys Lotus assets from IBM.

9) Google buys Ebay. See 7) above. "Froogle" has gone nowhere, Ebay got spat out of China. Google has innovative IT at scale, Ebay needs innovative IT at scale. Match made in heaven, and these kinds of deals won't find a favorable approval environment forever.

Final Swag:

10) Innovators rule. Vista is great looking, but is it a compelling upgrade if DRM restrictions give new users less control than they have now? Apple is also wading into DRM waters that make sense for IPods but not for enterprise-anything. Oracle has its hands full with Fusion, and if you like the JD Edwards, Peoplesoft or Siebel software you're running you might not want to Fuse in '07. As Tony Curtis once said: "In confusion there is profit." '07 looks like a great year for disruptive technologies and solutions.

At some point (probably between the 2000 Super Bowl halftime show, and when all the companies that advertised in that show went bust) after the first Internet boom, it became clear that the Internet didn't change everything, and that some old things (revenue, profitability) still mattered.

And so, gradually "monetization" crept into the equation, and the conversation on monetization continues to this day.

I came into the game at about this point, recruited back to Oracle to run operations for one of Larry Ellison's pet ventures - a hosted applications service called "Oracle Business OnLine." (BOL for short, and later Oracle On Demand). I came in as part of the second or third team tasked with getting the initiative moving, and as a favorite venture we (my boss Chris Russell, and me at Chris' invitation) met with Larry Ellison every week for about 18 months from 1998 - 2000.

Here's some of what we learned in two years before the Ellison mast:

1) Just because you build it doesn't mean they'll come. We had a great sales effort and support from the Greater Oracle, but getting people to move their core systems outside their control was a tough sell.

2) Nobody knew what these systems should cost. We sold hosted seats, all-in for six-ninety-five per user per month. I can remember one of our prospective customers asking "Is that $695 per month, or $6.95?" On hearing it was the former figure, the prospect just shrugged. Who knew if that was a good price or not?

3) If it's on the Internet, it's assumed to be free. Prospects weighed $695 and $0, and liked $0 better.

4) If your customer's first price assumption is roughly $0, then you better have an architecture capable of running profitably at something near $0.

We at Oracle weren't alone in coming to these conclusions. Other folks who came to the same conclusion, and the architecture we all came to in the next installment.

"Life must be understood backward. But...it must be lived forward." Kierkegaard

Credit Bill Gates and Microsoft with being the first to realize that there was a mass market for software "on every desk and in every home." Software Economics 1.0 said 1) software had high fixed costs of development, 2) software had essentially zero marginal costs of delivery, and thus 3) the first to mass scale would rule all, and throw off incredible free cash flow in doing so.

Software 1.0 was sold by license, delivered on media, where new license sales are good, ongoing support revenue is ok (a necessary evil), and addition services (given that they deliver lower margins than license sales) are bad.

This model drove Microsoft (and to a lesser extent Oracle, SAP and other giants, who were more willing to flirt with independent P/L services businesses, for themselves or their partners). Microsoft still operates essentially this way (waiting for that Vista upgrade license revenue stream) today.

Still, software had no sooner left the garage than two giant holes began to appear in the model. Firstly, upgrading software is painful, and the incentives for the buyer to take and install software shrink as the software grows to meet the business need. Given that Windows XP is pretty good, will businesses really want to absorb the deployment/support costs to adopt Vista?

The second problem with the Software 1.0 model is that fixed costs become so dominant that only the leading players can stay in the game. Prior to all the acquisitions (PeopleSoft/JD Edwards, Siebel, etc.), Oracle needed to maintain a staff of 5,000+ application programmers, not to chase down SAP, but merely to stay credible the applications software business.

Leave it to the Children of the Corn to find a new way...

A number of commenters have speculated about what Microsoft would create after Vista - with the idea being that the market has moved beyond large, monolithic systems deliveries. Google, and specifically Google Labs, is the new delivery model, and in the next couple of posts I'm going to speculate on how the industry evolved to this point, and what it implies for software delivery going forward.

I was a VP at Oracle at one of the specific points when the industry turned, and there are a lot of interesting things to draw from the evolution, and Larry Ellison's (among others) vision for the future. Larry has (cultivates, maybe) an image as a maverick at least, maybe even a crazy man, but in the evolution of software he's been crazy like a fox, and his vision has been spot-on. Next up: The Year Zero: 1994-1995.