Sources say that Microsoft may be on the verge of introducing a new pricing model for Office, called "union," which does not differentiate between cloud and on-premises installations. (It's already told some of its partners.) With union, customers would pay the same price for Exchange and Sharepoint subscriptions whether services are hosted in Microsoft's cloud or installed on premises.

I'm unclear about whether union would involve one of Microsoft's existing subscription-based hosted services, such as Business Productivity Online Suite and its "deskless worker" option, or if Microsoft will allow customers to develop and run Office applications on Azure and pay the same amount as on-premises. Windows Azure is open for business today.

Further, I am told that Microsoft has classified two types of business users: enterprise and light users. Enterprise users follow the client/server model, but light users might opt to run Exchange and SharePoint in the cloud. However, many of Microsoft's enterprise customers would follow a hybrid model, according to Tim O'Brien, director of the platform strategy group at Microsoft.

When I had breakfast with O'Brien last Monday (Microsoft paid), he acknowledged that some workflows (or parts of a workflow) should not be hosted in the cloud. That is true, and Microsoft has long been consistent about that.

Microsoft realizes that many of its customers' cloud strategies will follow a hybrid model, O'Brien said. It learned that lesson partly from experience with its hosted servics business. Microsoft intially believed that customers, including McDonalds and Kelley Blue Book, would consolidate Exchange and SharePoint online. Instead, they partitioned what ran on and off premises, O'Brien said.

Union would make sense for large enterprise customers like McDonalds, but it would also work for smaller customers that just need more capacity. My biggest question is how aggressively Microsoft intends to push subscription pricing, and whether Office is going to be available under a subscrption.

A week ago, I let ya'll know that the core PHP team had been brought to Facebook's main campus. That team were forced to sign NDA's, and taken to a very quiet, secluded meeting room where some cool new Facebook-backed open source project was described.

Well, I was able to put all the pieces together on this one, finally, and I now understand exactly what is up: Facebook has rewritten the PHP runtime from scratch. This coming Tuesday, they will make a big announcement around this project, and will make it available as open source software. I'm not really sure of any of the details of the project, but I do know that Facebook hired someone two years ago to do this, and I'm relatively sure this was a one-man project during that entire time.

So, why has Facebook rewritten the PHP runtime? Because PHP is obviously too slow for their tastes. A few years ago, I had a coffee meeting with some of the folks from Zend. When they asked what I had been hearing about PHP in the market from my sources, I hemmed and hawed, then told them that I had heard people complaining about how slow PHP was. Now, I don't personally consider PHP slow: it is simply not a language designed for the sorts of workloads that Java and .NET are.

But that still doesn't change the fact that PHP can be a tad pokey on the server. Well, when I said this to the Zend folks, their immediate reaction was similar to that of a gestapo officer looking for a spy: "What? Who said that? Tell us their name!"

Clearly, Zend does not think there is a problem. But Facebook did. Not enough of a problem to support more than one paycheck, but then, considering how many users they have, even a 1 percent performance gain would be a massive help.

This Tuesday, salvation should arrive. I would imagine this new project will push a lot of the weight in the PHP community into Facebook's corner of the world. It will be nice to see what they can do with all that interest, since Yahoo!, in the same position 6 years ago, largely squandered their opportunity to mold PHP into a more robust platform and language. 

UPDATE: After sifting through the comments here and elsewhere, I'm inclined to agree with the folks who are saying that Facebook will be introducing some sort of compiler for PHP. This sounds highly plausible, and fits into what I've heard. Obviously, I don't have absolute specifics. Thanks for the extra info, readers.

For those of you in the Java world, my colleague Alex Handy has confirmed that there will be a JavaOne this year. Alex reports that it will be co-located this year with Oracle OpenWorld, which is scheduled for Sept. 19-23 at the Moscone Convention Center in San Francisco, home to JavaOne for many years now. Neither the OpenWorld Web site nor last year's JavaOne site make mention of this, and it will be interesting to see if the same members of the Java ecosystem support the event now that it appears Oracle will own Sun well before then.

I spent the day touring Microsoft's Cambridge office to learn more about the nature of its relationship with Novell. In the coming days I'll be writing about a new project that the companies have jointly developed, and will examine how they work together in their interoperability labs. Expect a few surprises.

A friend in the valley clued me into a juicy bit of gossip that I have been unable to confirm or yet uncover. Perhaps y'all can help. I've heard that a whole gaggle of PHP core developers were invited to Facebook's offices today to discuss some grand new open-source project from Facebook. The company has already opened a number of projects, including Hive, a data-access layer for Hadoop.

But I'm as yet unable to uncover the actual nature of this new project. Considering the PHP developers being called to the scene, I'd imagine this has much more to do with the Facebook presentation layer than with its back end. We already know the back end is producing a ton of data, and that Facebook's biggest challenge is to make heads or tails of all that information. Hence their use and development of Hadoop-based solutions to the problem.

PHP is certainly the underpinnings of Facebook's presentation side. Does this mean we're about to see Facebook open its platform?

Embedded systems modeling tools provider Artisan Software has merged with embedded Java and Ada tools provider Aonix to form Atego, which will focus on safety-critical systems and software, the company announced today. James B. Gambrell, the former CEO of Artisan, will become executive chairman of Atego, with responsibility for the new company's strategic direction and future acquisition opportunities. Pierre Cesarini, the former CEO of Aonix, will serve in that role at Atego and will be responsible for worldwide operations.

Artisan Studio remains the company's flaship software for modeling, with support for UML, SysML and architectural frameworks, while Artisan Workbench remains the company's development framework. Aonix brings in the PERC product line for application development in Java and Ada. Terms of the transaction were not disclosed.

Over the weekend, a fairly large shift in the balance of power in rich Internet applications occured. As we all know by now, Adobe's Flash is the most popular browser plug-in and the most common way to play games in your browser. It's also the plug-in used to bring us video, audio and animations resplendent with silly characters and non-sensical music.

For some time now, the lack of a Flash player for the iPhone has been a big problem for mobile users. They've complained that the iPhone uses a full, real browser, but does not have access to whole swaths of the Web due to lack of Flash support. Certainly, Flash shouldn't be used as a front mend to data for this very reason: It may make data pretty, but it offers no text search or little accessibility support, and it confounds low-tech users who prefer Lynx to Firefox.

But this weekend, German developer Tobey Schneider may have shifted the balance of control on Flash away from Adobe. You see, Tobey has written a Flash runtime in JavaScript. The project is called "Gordon," and it looks to be in very early stages thus far. But I would imagine Mr. Schneider will have an awful lot of people interested in helping him flesh out this effort. Gordon relies on HTML 5 and the SVG standards, so this isn't going to take over for Flash anytime soon. But it may just help to remove the need for Adobe and Macromedia's love child once HTML 5 becomes widely adopted.

Frankly, having dealt with an obnoxious phone directory listed in Flash with no search function, and having been forced out of websites because I wasn't using a graphical browser, I can't wait.

Microsoft has pledged $1.25M to relief to NetHope, a humanitarian organization, to help victims of this week's Haitian earthquake. Here are some helpful links if you would like to support relief efforts:

American Red Cross: http://www.redcross.org/

Care: http://www.care.org/index.asp?&p=http

NetHope: http://www.nethope.org/

MercyCorps: http://www.mercycorps.org/

Save the Children: http://www.savethechildren.org/

Unicef: http://www.unicef.org

World Vision International: http://wvi.org/wvi/wviweb.nsf

Yele Haiti: http://www.yele.org/

Rob Caron, lead product manager of developer content strategy at Microsoft, announced on his blog this morning that Visual Studio 2010 will ship on April 12. Microsoft has previously delayed the launch due to poor performance and virtual memory usuage. Pricing information is available here.

Agner Fog is a computer science professor at the University of Copenhagen's college of engineering. As he puts it, “I have done research on microprocessors and optimized code for more than 12 years. My motivation is to make code compatible, especially when it pretends to be.”

Fog has written a number of blog entries about Intel's compilers and how they treat competing processors. In November, AMD and Intel settled, and Fog has written up a magnificent analysis of the agreement.

If you have any interest in compilers, and in Intel's compilers, you should definitely read his paragraph-by-paragraph read through.

Fog broke it down for me:

The machine code is actively testing for vendor ID before testing for any CPU models or instruction sets. When the vendor ID is not "GenuineIntel" then it chooses what Intel calls the "generic path", which is the least advanced among the possible paths, using the oldest instruction set for compatibility with old processors.

This happens in code generated by the Intel compiler if you allow CPU dispatching. It also happens in many Intel function libraries that are called from code compiled with the Intel compiler or with any other compiler, even if you are not explicitly asking for CPU dispatching. — Agner Fog

The big question now is: Will Intel remove barriers to its compilers working on other processors? And thus far, no one seems to know the answer.

But Fog has written some code to solve the problem. You know. Just in case. He says it's "very easy. Remove the check for vendor ID and check for supported instruction sets only. I have made a small function library as a showcase to show how to do this and how to support all x86 platforms. http://www.agner.org/optimize/#asmlib"