IDC greases slope of despondency

Web services are a decade away from fulfilling their promise, research organization IDC has declared in a press release issued yesterday. One wonders what makes this news at the end of October, when IDC first published the report in question in August? (We first made note of it here in early September). Perhaps the explanation lies in Gartner's hype cycle theory, which as we reported earlier this month, finds that web services hype peaked in the summer, and has now begun its long plunge into the 'trough of disillusionment'.

Gartner's finding is borne out by the alacrity with which detractors have seized on IDC's press release as vindication for their own scepticism over web services. Clearly IDC's marketing people are avid followers of Gartner's hype cycle predictions and recognize when the time is ripe to exploit negative sentiment to drum up sales.

Anyone who hands over the $1,500 list price in the expectation of having their gloomiest prophecies confirmed will find themselves sorely disappointed, however. Any realistic assessment of web services recognizes that the creation of the necessary standards, tools, methodologies and best practices is going to be a multi-year pilgrimage, and IDC's report in fact turns out to be an excellent guide to the major milestones along the way.

The report is definitely an antidote to all the mindless, ill-informed hype that surrounds web services, but it's also a valuable practical guide to the technology, business and cultural obstacles that web services adopters will need to negotiate over the coming decade. Even though IDC's publicity for the report has given the web services movement a sharp downhill shove towards disillusion (hey, thanks, guys), its substance will enlighten those who are already looking ahead to the plateau of productivity that web-based service-oriented architectures will deliver in future years.
posted by Phil Wainewright 4:06 AM (GMT) | comments | link

XDocs and rivals put users in charge

For most of its history, IT has concentrated on automating the firmly established 20% of business processes that do 80% of the work, seeking to grind down costs by maximising economies of scale. But it is the much larger number of highly dynamic, opportunistic business processes that make the most difference to a company's profits and competitive edge — they contribute the inspired 20% of the work that generates 80% of the value.

The challenge for IT is finding a way to provide effective automation of these unpredictable, fast-moving and often one-off processes. But the old approach of interviewing users and then returning weeks or even months later with an automated solution cannot work for such dynamic activities. Indeed, the attempt to do so — most notably when automating customer-facing processes — has often ended in unmitigated disaster, to the lasting discredit of IT's good name.

What is needed is a way of providing automation on demand, one that acknowledges that the people best placed to decide their day-to-day application needs are the experienced users and managers who run an enterprise. There can be no doubt that they are eager to gain more control over the computing at their disposal. The phenomenon of 'power users' — using advanced macros and scripting to add custom capabilities to spreadsheets, word processors and other productivity tools — demonstrates their willingness to learn new skills, if doing so enables them to see direct results from their efforts.

But the idea of allowing users to create and modify their own applications has always seemed an idle dream. Just a few weeks ago, a reader admonished me when I wrote "the role of developers and vendors is to give us tools that generate the script to do the things we need to do, on demand." Peter Herndon responded that, "short of identifying a very large percentage of 'the things we need to do', the tool you are describing is called a programmer."

Certainly, there will always be occasions when the services of a developer or programmer are indispensable, particularly when the aim is to find a faster, slicker way of executing a predefined business process. But just as developers are the experts at innovating software, why not acknowledge that business users are the experts at innovating business processes? If that's the case, then they have to be given the tools for doing the job, and one of the most important capabilities when creating a new process is to be able to test assumptions and fine-tune the automation along the way. That can only be achieved effectively by allowing users to prototype their own process automations and iteratively test and debug them. They don't need to be able write code, but they do need to be able to directly create, modify, dismantle and reassemble processes.

A broad range of vendors, from startups to leading corporations, are beginning to bring various products to market that will empower users to do just that.

On Monday, Silicon Valley based startup Movaris announced it has closed an $8m second round of funding led by well-known VC firm Mohr, Davidow Ventures. The company is an example of an emerging new breed of startup that is dedicated to putting process development in the hands of users. Its products provide a point-and-click interface for business and system analysts to design and deploy business process applications that cut across multiple applications, delivered as PDF-based forms modules which can then be managed by their users.

Microsoft has a similar objective in mind with its XDocs project, which as Mary Jo Foley wrote this week, "is also the company's trial balloon Web services client ... a smart client alternative to Office."

An advance cohort of developers will be getting early access beta test versions of XDocs and Office 11 this week, so doubtless there will be more information soon about XDocs. There can be no doubt however that Microsoft's debut of a user-centric forms-based process development tool will give the sector a strong boost, benefitting startups like Movaris who are shipping finished products rather than early betas.

Whether in the long run that will rebound to Microsoft's credit remains to be seen. It has a natural advantage in the corporate workgroup market, where Office is the trusted brand. But other brands also have credibility and confidence there, not least Lotus Notes/Domino and Novell. Meanwhile, startups may lack a credible track record but they will have a greenfield capability for innovation that Microsoft will be hard pressed to emulate while keeping faith with legacy users of Office.

Outside of the corporate workgroup, where Microsoft has less natural dominance, it is even more at risk from competitors. When collaboration and business processes have to cross corporate boundaries, the LAN server is replaced by a Web server, and here Linux is a real threat. Not only is it cheaper and better established as a high-volume hosting platform, it also supports the development of low-cost, easy-to-modify applications based on PHP and MySQL, with highly flexible forms development and XML web services integration capabilities. Novell's announcement this week that it will include MySQL with its NetWare 6 software is a big boost for this approach, and one that will do the vendor no harm at all in its quest to leverage its server heritage while repositioning its brand as a net-native leader.

That's why there was a serious edge to my tongue-in-cheek question last week, Is XDocs OS/2 for Microsoft?. Microsoft is clever enough not to fail in the face of these challenges, but they are going to be a tough test of its mettle.
posted by Phil Wainewright 8:03 AM (GMT) | comments | link

The semantic scandal of web services

Web services connect, but they don't communicate. Despite the hype, they solve only one part of the integration problem. This is the scandal described by Jeffrey Pollock in an article in August's EAI Journal, downloadable as a 4-page PDF (thanks to Doug Kaye for the link).

The article dovetails into the finding I mentioned yesterday by Sand Hill Group about the lack of semantic standards in web services. Jeffrey Pollock explains why this is important: "It's what's inside the envelope that matters ... With web services, there's an assumption that either every system using them will be speaking the exact same language and dialect or a translation service will have to be coded to enable systems to communicate well. Where's the value in that?"

The point here is that, while web services break down the barriers at an application infrastructure level, they contribute nothing at the business integration level. You can connect at the speed of thought, but you still don't know what the other side is thinking. Participants need some way of sharing the context that gives meaning to their communication — of achieving "semantic interoperability".

As VP of technology strategy at semantic mediation vendor Modulant Solutions, Pollock of course has a specific solution in mind. But his exposition of it makes for an interesting read. He argues for a move away from reliance on either the old-fashioned method of custom coding or the more recent innovation of standardized vocabularies such as ebXML and RossettaNet.

Instead, he says, we need to move to a "sophisticated, loosely coupled information infrastructure" that uses an ontology-based system of semantic mediation. This makes it possible to share meaning by mapping data to shareable information models, using techniques that are reminiscent of "old-school" artificial intelligence, along with the ideas embodied in the W3C's work on development of the Semantic Web.

Like most people, I'm afraid my eyes glaze over whenever anyone mentions the Semantic Web. Yet deep down, I know that these concepts are vitally important. Web services will provide a firm foundation for moving ahead, but if we really want to achieve effortless electronic interactions, we will have to find a way of automating the fuzzier elements of human communication. In the end, it's all about making the machines behave more like the people that use them. Pollock's conclusion neatly encapsulates this:

"All serious, difficult problems — including interoperability and integration — are human problems. Ultimately, the disjointed vocabularies of disparate computer systems are the results of differing cultures, ideas, business processes, folklore, and viewpoints. A semantic interoperability solution focuses on the needs of users and organizations that are at the core of any sophisticated information exchange."

posted by Phil Wainewright 2:10 AM (GMT) | comments | link

Progress buys eXcelon for Sonic boom

Today's acquisition of XML tools and object database vendor eXcelon by Progress Software is intended to boost the appeal of its web services messaging subsidiary, Sonic Software. According to a prompt piece of news analysis by Line56, eXcelon will add extensible information server and business process management components to Sonic's enterprise service bus for service-oriented integration, accelerating B2B transformations.

"The combination of our [Sonic] network-based infrastructure and their storage and processing capabilities is a very natural fit for what we think is the long-term requirement for integration," Gordon van Huizen, VP of product management at Sonic told Line56. Loosely Coupled has previously linked to a Van Huizen article on the advantages of a messaging-oriented architecture for web services, and in June noted that Progress was one of the overlooked vendors that could make unexpected headway in the web services sector.

The acquisition of eXcelon looks like a smart move to strengthen its offering as Sonic moves up the web services stack, but like most XML pureplays, the strength of eXcelon's technology assets has not yet given rise to a significant customer base. Progress can help make up the shortfall with its own extensive market presence and substantial cash reserves, but it will need to work hard to make the most of its latest buy.
posted by Phil Wainewright 3:17 PM (GMT) | comments | link

Sand Hill tips web services winners

A July report on web services adoption has become a popular read among tech industry cognoscenti, despite a hefty price tag. The Web Services Derby by Silicon Valley-based conference organizer and business consultancy Sand Hill Group retails for $3,300, but impecunious readers can get a good taste of its conclusions by accessing the Executive Summary as a free download (PDF, 44k).

With web services being seen as a major driver of software industry growth, it's hardly surprising that a report from such well-connected authors should have excited interest. Billed as the first of a series (which naturally begs the question of when the sequels are due), the 70-page report is based on analysis of 60 current web services projects at Global 2000 companies, as well as interviews with 117 executives at leading software companies and their Fortune 500 customers. Among the findings of interest:

• A desire to solve business problems was the primary driver behind 74% of web services projects, with 61% aimed at improving customer or channel relationships, and one third focussed outside of the firewall. The remaining 26% sought to lower IT total cost of ownership. The primary reasons cited by enterprise customers for considering web services were:
  
  • Easier, faster, cheaper integrations — 90%
  
  
  • Partner integrations — 81%
  
  
  • Push business process over firewall — 76%
  
  
  • Leverage existing IT investments — 71%


• Early adopters have a strong belief in ROI. "Participants said that although it cost more to build work-arounds for security and other shortcomings of web services technology, the business value justified the investment."


• Nearly half of the projects were for what the authors termed "EAI Lite" and "B2B Lite" projects — not as heavyweight as EDI or EAI, but requiring "more semantic integration than an enterprise portal." They tipped this "lite" approach to integration as a high-growth area that in the long term "will pose a significant threat to EAI vendors."


• They predict that web services-based interoperability will lead to greater emphasis on business solutions. "This implies that systems integrators, application vendors and other software makers with strong connections to line-of-business executives and CIOs will benefit."


• Software vendors said they are adopting web services standards to help lower integration costs, although the majority are merely adding web services support to existing software rather than rearchitecting applications to expose functionality as business services. None of them believe that adopting web services will make them more vulnerable to competition "because there are no semantic standards in place." Enterprise portals and CRM vendors are the most aggressive early adopters.

Sand Hill has stuck with the web services theme in its monthly email newsletter, which carried an interview in its September issue (PDF, 123k) with SAP board member and entrepreneur Shai Agassi, who is leading the vendor's xApps initiative. He has some interesting insights to share:

"This new type of application really involves weaving applications. It won't be an application that will sit directly on top of the database as today's applications do. It will be an application that sits on top of existing applications, abstracts services from the applications, and then aggregates the services to create new processes ...

"What you'll see above these layers is a simple solution that can be implemented in less than a month — I call it a "bite-size" solution. These bite-size solutions will not only get implemented but will also get aggregated within one month. That means they will leverage all the same repositories, all the same standards, the same integration stack and they will get integrated and aggregated over time to solve more business processes. And these processes will be weaved together."

Agassi's comments are an illuminating expansion on the theme of exposing business processes that the report identified as a key objective for software vendors. Sand Hill is evidently staying abreast of the developing issues in web services and their impact on software vendor strategies, so it will be interesting to see if further editions of the report emerge as originally promised.
posted by Phil Wainewright 7:18 AM (GMT) | comments | link