![]() ![]() |
![]() |
|||||
![]() ![]() |
||||||
![]()
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
![]()
Web design diary: push technologyJoanna Bawa Say goodbye to waiting ages to download data from the Web: push technology could soon be sending you updates on just the kind of information you're interested in.
Push is conceptually similar to television broadcasting. You can subscribe to different news services in the same way that you can choose between channels, viewing (or hearing or reading) only the information you actually want. Pushed information is arguably more flexible than a TV news broadcast, though, because there's no need to sit through an entire programme to access what you want. You can also store information that's pushed at you on your PC's hard disk, which means you can study it whenever you like, and not just at the moment that it's broadcast. Set up correctly, some push products will work unattended, retrieving new information and downloading it to your hard disk at regular intervals even when you're elsewhere. Effective push technology requires 'push capability' to exist across servers and networks, as well as on the receiver's PC. At the server end, push content--content that's aware of and responsive to push products--is created and grouped according to broad headlines, such as news, sport, corporate information and weather. A user, equipped with a compatible (and currently proprietary) push client can subscribe to any of these content groups (or 'channels') and define their own information profile--the precise nature and frequency of the content they want to receive.
Going to the pollsThe mechanism whereby information flows from channels to your desktop is much more variable, but two main forms of it exist. The push client might poll the server at regularly defined intervals and retrieve whatever is available at each particular moment. Alternatively, the server might send out all its information to all its clients whenever it arrives, so it's up to the client to determine whether the broadcast fits the specific profile. If it does, then the broadcast will be passed on to the user's desktop. This works across corporate intranets, but it can also be effective across the Internet--content on any given site is preconfigured for a particular push client, but users of that client don't have to contact that site directly to retrieve the information that they want. Instead, an intermediate push server caches relevant information and then passes it on in one go to users whenever they access the Internet. Despite its coherent concept, the ways in which push products work are surprisingly diverse. In many ways, the term 'push' is a misnomer, because although users have information being pushed at them, what's actually going on is quite different. Key products, such as Marimba's Castanet or PointCast's I-Serve (see Figure 1), work by polling the push server at regular intervals and then pulling relevant information off it--so a more accurate name would be 'client poll' products.
Pushing applications
One of the most innovative push products is Castanet (www.marimba.com), shown in Figure 2. This isn't because of the way it pulls or pushes, but how it manages data. Castanet is geared to pushing and maintaining entire applications across a corporate intranet, rather than sending news or information. Once an application has been pushed to a user's desktop, Castanet updates it by sending only the information that's changed, minimising network traffic. This is perhaps one of the most useful applications of push technology so far, although it's currently restricted to network-based Java applications. No doubt this will change in the future. 'Traditional' push technology is based around the idea of distributing messages and Web pages. The most established product here is PointCast's package of push channels and its push product, I-Server(www.pointcast.com). Subscribers to PointCast gain access to an array of content feed, such as news, weather, stock information and world headlines, right through to personal horoscopes. If several users subscribe from the same corporation, however, this can lead to serious bandwidth bulges as packets are sent at the same time. The amount of space available to customised corporate channels is also limited--only one true channel can exist, although it can contain a number of groups which must be individually requested by users (making criteria for delivery more complex, and content harder to control). The other drawback of PointCast is commercial. Its own servers are a haven for advertisers, whose graphics and messages are pushed, along with the requested information, across busy networks to corporate desktops. The most genuine push product is Wayfarer's Incisa, which is based around a multicast approach. Instead of requiring clients to check for new content, Incisa sends a single message that gets routed to all recipients. It's one of the few to be marketed as a corporate more than a consumer tool, and its multicast approach makes it more efficient and more secure. It also means that what it delivers is quite restricted--usually little more than links to Web pages.
Consumer benefitsWhat's the use of push technology? It might not be as clear cut as it appears, with many corporations uneasy at the prospect of employees constantly distracted by notifications of information, much of which will be irrelevant to the task on which they're working. As a result, it's generally touted as a consumer benefit, bringing news, share dealing information, weather, TV and movie listings to users. This hasn't been lost on the browser warriors, whose attitude is to embrace and incorporate push technology rather than attempt to compete with it. Both Microsoft and Netscape plan to build in push capability, with Microsoft promising that Internet Explorer 4.0 will include the ability to connect to PointCast channels. Netscape prefers the Castanet option and is building its capabilities into NetCaster, an add-on to Netscape Communicator, which is currently in beta testing. Both vendors want to merge the Web and the desktop, and each wants to dominate the resulting interface. Push technology-enabled browsers will deliver content via channels of information sorted into logical groupings, and ultimately something already dubbed 'metacontent'. This is information about the channels themselves, which will allow overloaded users to focus more accurately on what's useful to them. The limiting factor today is the small number of Web pages that are sufficiently active to justify anything being pushed to users. As dynamic HTML becomes more widely used, pages will themselves become more dynamic as multimedia and back-end databases become more involved. A likely increase in Java applets and ActiveX controls will also increase the extent to which pages become genuinely live and dynamic.
Presenting carefully defined information to workers is seen as a natural application for push technology, but that doesn't mean everyone wants it. Designing push to work within organisations is difficult and demanding, and established technologies mean that the niche for push is surprisingly small. The primary benefit of push technology to the corporate desktop is time-sensitive information that's directly relevant to the business. For most, this means financial and share-related information, sent to a small number of professionals who genuinely need to know. Even so, live television feeds already exist, so the Web itself must be offering better quality and better presented information before this becomes a viable proposition. The easiest way of selling push to some companies is by stressing its role as an advertising medium when content is pushed to other companies. Unlike most Web sites, push servers know exactly who they're pushing information to, when and why. They also guarantee a readership with an active interest in what they're receiving. This idea works for advertisers, but can be a pain for network administrators, whose bandwidth is rapidly soaked up by extraneous advertising information. For corporates, notification is the best and worst aspect of push technology. Users get the information they need the instant it's available, but often that means constant interruption. And there's a range of interrupts: email messages, system beeps, alerts, animated icons, scrolling tickers and flashing screens. It's distracting, and the criteria for selection aren't sophisticated enough to eliminate useless information. Not only is this a problem for managers, it's a potentially enormous problem for users. Information overload is already an issue for many people, and the pressure imposed by a constant feed of new information might just tip them over the edge.
Upgrading applicationsThe most useful business application for push today is the distribution of new and upgraded applications around an organisation, which Castanet does best at the moment. Another powerful justification is to improve customer service: some high-tech companies, such as Epson, push information to customers to notify them of new products and driver updates. Push technology is a long way from the information nirvana that many of its backers seek. Although it's a technically robust and rapidly improving technology, it doesn't yet deliver a compelling solution, particularly within the corporate environment. There are very few people who absolutely must have regular injections of new information throughout the day, and fewer still can constructively use the information they do receive. As a consumer tool, it needs to be better at working unattended, and filtering must be refined if it's to compete successfully with conventional listings. Nonetheless, niche applications are emerging for push technology, and as it improves these will certainly grow. |
|||||
Copyright © 1997 ZDNet UK. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of ZDNet is prohibited. ZDNet and the ZDNet logo are trademarks of Ziff-Davis Publishing Company. |