"Anytime IBM gets into something, it usually calms people down as something safe to use in their business," said Michael Coté, an analyst with Redmonk last week.

Generally speaking, mash-ups are applications that combine data from a number of sources and present them through a rich user interface. They are meant to be built quickly and easily, and in many cases accessed over the Web. The concept first took hold with consumer-oriented applications that mashed publicly available data sources, such as Google Maps and event listings.

Today, there are a plethora of APIs (application programming interfaces) available, and an array of companies -- including large players like BEA and TIBCO -- are selling enterprise-focused tools.

The market is even showing subdivisions. Some vendors, such as SnapLogic, are focusing on data integration. Many others, often lumped into the RIA (rich Internet application) category, deal more with the presentation layer.

IBM's entry centers on two products: Mash-up Centre, a development environment aimed at non-technical business users that entered beta on April 15, and WebSphere sMash, a more robust toolset for developers that enables the use of dynamic scripting languages. It is based on the previously announced Project Zero effort. IBM will release the commercial product later this year.

While the products may not offer anything fundamentally new and follow years of work by other companies, the fact that a vendor of IBM's weight is pushing a broad set of mash-up tools could be proof the space has truly arrived, company officials and observers suggested.

"There comes a time when you have to place your bets," said Larry Bowden, vice president of portals at IBM. "With the force of our announcement, we're making a bet here."

"IBM's announcement substantiates the market," said John Crupi, chief technology officer of JackBe, an enterprise mash-up vendor that is launching the 2.0 version of its Presto platform this week.

But JackBe is well-prepared to compete, he said. "We are built as a pluggable part of the stack. We're very lightweight and very agile and can play with heterogeneous vendors."

JackBe has managed to get some heavy-duty customers to use its tools. The U.S. Defense Intelligence Agency created Project "Overwatch," a dashboard-like application that integrates and correlates various intelligence-related data.

Bob Gourley was chief technology officer at the US Defense Intelligence Agency during the project's creation. The DIA was drawn to mash-up -style development due to one of the concept's central promises -- speed.

"For the end-user, it is speed in getting data. We don't want to wait hours for data to end up on someone's desktop," said Gourley, who is now principal of Crucial Point LLC, a consultancy.

"Another component of speed is the speed to deliver something," he added. "In the military and intelligence community, we have been pioneers in displaying information over terrain and maps and letting people interact with it. But that was all hard-coded -- that would take years to build something like that."

While the DIA found a place for mash-ups within its IT strategy, enterprises shouldn't buy first and develop later, said Jason Bloomberg, an analyst with Zapthink, a consultancy focused on SOA (service-oriented architecture).

"[Mash-ups are] really a solution looking for a problem," he said. "It's great to be able to put your pizza places on a Google map, but what good is it?"

One clear-cut role for mas-ups has emerged, he said. "Mas-ups are becoming killer-use cases for SOA," Bloomberg said. "You can show a mas-hup -based solution to an executive and they'll get it."

Yet, any serious conversation about SOA inevitably turns to governance. "You can't just let anybody mash up anything ... all that has to fit into the governance framework an organisation has," Bloomberg said.

IBM says its tools have governance and security built-in. JackBe also offers governance components, which include a connector to Hewlett-Packard's SOA Systinet, a management platform for SOA.

The ever-present spectre of vendor lock-in is another concern, Gourley warned. "The challenge is that each of the big players have inherent biases to build things that only work best with their solutions," he said. "This is never their stated intent, but the internal pressures to do that are something that all CTOs should watch closely."

Meanwhile, smaller companies like JackBe are "absolutely forced to work well with everyone," he said.

Then there's the question of what constitutes a mash-up developer. While mash-up vendors broadly proclaim their tools' ease of use for nonprogrammers, some level of development skill is realistically required, according to Coté.

"I think what you can do in [Microsoft] Excel is the bar," he said. "If a user can create macros and pivots in Excel, then mas-ups should be conceptually within their grasp. ... [But] you're always going to need programmers."

Gourley said "the ideal mash-up developer is no developer at all."

"A mash-up platform should be so easy to use that in-house developer talent can succeed with it, and it should also be so easy to use that end-users can do the mashing," he added. "When large integrator companies claim that they field mash-ups, I start asking them hard questions like, 'Did you really field a mash-up or did you hard-code the interface?'"

"But ... generally some development is required, and when a mash-up platform is selected it should be one that lets internal developers field solutions fast, in days versus in months," he added.

Even before choosing a vendor's offering, companies should lay the groundwork for their mash-up strategy, Gourley said.

A first step is to determine the initial data sources to be used. "You can add more later, but it helps to have an initial list so you can ensure the data can be securely consumed." Second, companies should go with a browser-based product, he argued. "You have many options to chose from, including thick applications, but browser-based solutions are more maintainable."

Echoing Bloomberg, Gourley said the most important thing is to "begin with a vision for what needs to be accomplished. Without a vision, the odds of success will be much lower."