There’s Money In The Middle

In 1997, when WAP was unveiled to the world, the proposed information flow chain neatly stated that content would be provided in wireless markup language (WML), converted to binary WML, sloshed through a WAP Gateway, blown out on cellular networks like GSM, and finally sucked into and displayed on mobile telephone handsets.

Customers who were even able to get the first WAP phones (many models were late in rollout) complained bitterly of slow speeds, caused not just by the service but also by the devices themselves. The over-hyping of WAP, especially in Q1 2000 and Q2 2000, and subsequent disappointing offerings nearly put the nail in WAP’s coffin, from a marketing standpoint.

More significant than the slowness, however, is the fact that with the wireless Internet there are heaps of different devices to format for, and WAP-oriented content providers have the not insignificant task of managing two content formats, one in HTML and one in WML.

Problems aside, WAP probably isn’t going anywhere, at least for the next few years, simply because of device penetration: millions of WAP handsets are already in the hands of users, and new GPRS (general packet radio system) or 3G-enabled terminals will need time to run their product lifecycle from early adopter high-fallutin’ business people, through to the kids in the discos to, well, my mother.

New solutions So as mobile data delivery moves from phone handsets to “terminals”, competing browser protocols and devices will come and go in the coming years. Getting content to all the different devices is still the challenge and there are lots of ways to do it.

Take a straight “delivery system” such as AvantGo, which is purely infrastructure: companies use it to extend their content or applications to a mobile device, by compressing image size and format and optimizing layout for the device requesting the information. It also manages offline versus online content, letting devices with always-on connections browse at will but caching entire sites locally for people with dial-up connections.

That’s a straight compression solution and many in the industry say that “trans-coding” (conversion) of one form or another will be the way to go in the future. Because legacy content isn’t just HTML (it’s often in the form of Word, Quark XPress, flatfiles and PDFs) software that trancodes or converts from old formats to new ones is hot these days, with dozens of startups saying they can do it better than anyone else. Those companies will undoubtedly get shaken out, and some clear winners will emerge in the next year or so. More interesting than them, however, is the coding method and the process used.

As we have seen, the darling of the “do-it-all-code” pack has been XML (extensible markup language). While HTML, the markup language of the Internet, allows control over the appearance of content, such as for bold (the command for a bold typeface), XML allows markup that describes the content itself, such as Le Grove.

The beauty of XML, and XLST, the stylesheets that control how XML can be presented on a page, is that they create a single source of uniformly-formatted data from existing content, which can in turn be squeezed out into whatever flavor you want – HTML, WML, nML and so on.

A new data chain So the new chain of data goes from legacy content to content conversion; to the generic, XML-ized content; to a content gateway, which takes the XML and converts it to both device and code-specific content based on the type of device requesting the data; to the protocol gateway, which negotiates multitudinous device protocols such as WAP, and iMode; to the network and finally to the wireless devices.

You could see how this type of thing would be of compelling interest to Roger Barnes, a consultant for the Rough Guides series of travel guidebooks, which sits on a heap of content in QuarkXPress.

Barnes was approached by AuthorOnce, a company that claimed that they could “actually do it now: take our content, put it through a GUI, and put it out to any platform we wanted,” says Barnes. As we went to press, Barnes had seen and been impressed with a small demo, the success of which had led him to schedule a meeting in New York with the AuthorOnce team and Rough Guides’ senior management.

AuthorOnce is one of several companies offering what may be looked upon as complete middleware solutions – from one end of the chain to the other, and then back again. The company, which has received friends and family backing to the tune of $750,000 and is currently fishing for a first round of funding, claims that what sets it apart from companies like AvantGo and Everypath, is its method of getting data from the legacy system into XML in the first place.

“We’ve got travel books, but we’ve also got guides to music,” says Barnes, “Converting text to XML is one thing, but we’ve got pictures, maps, headlines. The company’s “rule engine” system learns about the way we publish our books every time we work on one. So preparing the new Rough Guide to New York, it knows what we did last time.”

That’s a different added value from offerings from other companies, like AvantGo and Everypath, that simply take content, pull it up into XML, and send it out to a Web or WAP interface. Those companies say that their products are perhaps the most effective way of getting legacy information out to a world of different device formats.

AuthorOnce might disagree, saying that the hardest part of the chain isn’t delivery to the devices, it’s XML-ing it in the first place, and doing it in a way that allows you to control the flow of data and create rules for future conversions of like-formatted but different texts.

Taking one end of the chain
“Well, if you’re in the business of from n to XML, of course you want to view this as the problem,” said Rikard Kjellberg, CEO of Ellipsus Systems, a company in Stockholm that provides the Protocol and Content Gateways. “There are lots of excellent tools that offer the mechanics of going from the database to XML – I’d bet even Oracle would have tools for that.”

Kjellberg’s Ellipsus concentrates on what happens after the content is in XML, and how to best transmit the data to the jungle of devices out there. Its Sargasso Mobile Internet Server gives an open software platform that lets legacy content connect, through any IP bearer (CSD, GPRS, etc) to client devices. It consists of a pull and a push proxy gateway, a directory interface, a manager interface, a security pack and a “gatekeeper” firewall, allowing access control for the Web as well as RMI, CORBA, SOAP and other objects.

That is the unique selling point; Ellipsus allows developers to introduce CORBA (and, for example, Enterprise Java Beans) all the way to the device, letting them make a more dynamic interface to legacy systems than would be available with traditional HTTP.

What it’s doing is creating a virtual thin client within the Ellipsus system, which end users access via nML from their phones. The phone doesn’t need to support CORBA, it just needs to communicate with Ellipsus, from where the object communicates with the legacy content or application. The menu the user sees on the phone doesn’t change, it’s just got a different back end: where a menu would have behind it a URL, like , the object-access menu has an address like .

Ignoring the problem
And then there are those who would ignore the problem completely, saying that they’re focusing on the problems created by having multiple systems in the first place. Companies such as mi4e, a Stockholm-based company that makes a plug-in for web servers that acts as a WAP protocol gateway on existing Microsoft IIS or Apache webservers.

There are also service developers, like France-based Selfswitch or Stockholm-based Expedio, which is producing unified messaging systems that let operators offer customers one central repository from which they can stay connected to voicemail, email, faxes, and a synchronized schedule; or Port42, which makes application portfolio packages that operators can buy in bulk, branding entire suites of applications to offer their customers instant application packages.

Similarly, there is Stockholm-based ZoomOn, which designs and implements vector-based graphics (VBG), and operates on the assumption that WAP – which does not support VBG – isn’t here to stay.

These companies are in effect saying that it’s too early to dedicate a company to bringing content to users via existing platforms or procedures, but that when the platform is agreed upon, they’ll be there selling the stuff that will make people want to burn up those airtime minutes.

In fact, unified messager Xpedio is going one step further, developing a platform for that time, about three years from now, when Britney Spears or whoever is then Britney Spears decides to become a “Virtual Operator.” Britney’s going to give away a SIM card with every CD that lets her teeny-bopping buyers get 10 minutes of phone time, 300 SMS messages and a Kiss Britney game.

“The platform they’re working on lets you, say, if you’re a U2 fan get a U2 subscription whether you’re in Ireland or Sweden,” said Port42’s CEO, Johan Rosenlind. “That’s a great idea but it’s still a couple of years away.”

Not so fast; they, and all other platform vendors will confront significant resistance in the form of iPlanet (the Sun/Netscape alliance), Oracle (Portal-to-go, ASWE9i), the Icelandic entry WAPalizer and Microsoft MIS. Basically, all these tools do much the same thing. How Xpedio will stand up in a fight against the portal-mongers is left to be seen.