|
IwC Journal Special Issue Active Web home | special issue | part 1 | part 2 | conference | tutorial |
The massive interest in the Web as an active and interactive experience is reflected in the fact that this special issue needed to come in two parts. However, the pace of change over the last year or two is rapidly turning Active Web into a tautology. It is increasingly hard to find web sites that are not in some way ‘active’ whether it is dynamically delivered content, personalised pages, site search engines, community areas and bulletin boards, or simply annoying animations!
With this change comes new problems. Cross platform compatibility used to simply be a matter of bad line breaks, but increasingly one visits sites that simply do not work. One of the editors recently attempted to log in to the business services section of a major high street bank, only to see a blank screen. A bit of investigation revealed that the page consisted purely of Javascript that clearly failed to work properly on the particular browser and platform. On other sites we find applets that crash the browser or lock the machine, web sites that deliver pages listing internal CGI errors, or, when they work at all, long delays and unusable interfaces.
It's not surprising that some sites are moving to simpler interfaces. In the search engine area Google pioneered this with its direct ‘type here and go’ front page, but now previously cluttered portals such as Yahoo! and AltaVista are adopting cleaner, more task directed pages.
So will the Active Web die in its own complexity? In fact the opposite seems to be the case. The examples above of simple interfaces are simple, but still interfaces. Indeed, it is often when appropriately personalised that the most appropriate information can be delivered most easily (and, of course, the most lucrative banner advertisements). Furthermore, the rise of non-PC use of the Internet: interactive TV, WAP, PDAs, means that we will see more dynamic creation, not less, as sites modify themselves dynamically for different types of display device. Keeping the presentation clearly separated from the content will also be key here, with technologies such as XML/XSL being particularly relevant.
If Active Web technology is going to increase, we need Active Web interfaces that are usable and robust. The papers in this issue address both the design of web interfaces and software mechanisms to support deployment.
Between the two parts of this special issue, three of the nine papers deal with educational applications. This is probably not surprising given academics perhaps choose their own workplace as a challenge. However, it also reflects the increasing demand for flexible, location independent, lifelong learning.
In "Automatic Generation of Instructional Hypermedia with APHID", Thomson, Greer and Cooke look at the design of learning materials by using design patterns and automatic content production. In APHID the producer of learning materials does not construct a whole web site, but instead creates fragments of information with dependencies and other high-level information. The system then uses this combined with heuristics derived from design patterns to weave a web of learning.
Isenhour, Rosson and Carroll's paper "Supporting Interactive Collaboration on the Web with CORK" also involves educational examples, but their focus is on technology to make it easy to create collaborative web-delivered applications. CORK is a Java toolkit that allows the easy management of shared, replicated data. Amongst the challenges faced in this work, the systems constructed using CORK allow fluid movement between synchronous and asynchronous interaction.
In these two education domain papers, we see a contrast in design vs. technology and also between an author-reader model in the case of APHID and a collaboration/co-production model (sharing!) in the case of CORK.
Ciancarini, Rossi and Vitali’s paper "Designing a document-centric coodination application over the Internet" also considers shared applications constructed using PageSpace, a framework for Java agents on different networked machines to communicate and coordinate their activities via the Linda tuple space model (which has been very influential in other areas of computing). Their central example is the management of paper refereeing for conferences close to the heart for many academics.
Although there are many commercial and research tools to aid web page authoring and technology to allow dynamic page creation from databases, there is less support for the production of web-based shared applications. Both PageSpace and CORK, although taking different approaches, show that this is achievable.
Kohrs and Merialdo in "Creating user-adapted web sites by the use of collaborative filtering" look at a different form of collaborative activity helping a user to navigate a web museum (that is a museum delivered via the web, not a museum about the web!) by using the explicit and implicit preferences of other users. The museum in question contains only images and so conventional searching based on words is ineffective and so collaborative filtering offers a strong alternative. Unfortunately, like many collaborative systems, collaborative filtering suffers from critical mass problems until lots of people are using the system and offering recommendations, there is little benefit in using the system. The solution Kohrs and Merialdo adopt is use content information (colour and texture) to create ‘artificial users’ which can then be used to ‘seed’ the system.
Finally, Nielson, Morris, Charlton and Little look at the future for web-delivered applications in "Interactively on the WWW is the ‘WWW shell’ sufficient?". As a way of eliciting and explicating the issues and problems of web-based interaction, they consider three case studies including a web-based email system. Their conclusion is that the future of web delivery lies in ultra-thin web clients that deliver the ‘presentation’ and rapid feedback part of the Seeheim model whilst leaving the underlying business logic (dialogue) and application back-end to the server side.
The web email example is interesting as it typifies a movement towards web-centric applications. Until a few years ago the largest volume of Internet traffic was email. This changed as the web began to drown all other traffic. Now we see the web delivering the email! However, this is just one example of a general trend to move applications to be delivered over the web (e.g. Star Office). If we interpret the web more widely to encompass other forms of global networking (WAP, wireless etc.), perhaps the Active Web will become a tautology, not just because all web sites are active, but because all interactive applications are delivered via web interfaces.
For more about the Active Web including links to papers from the Active
Web conference in 1999 and related material see http://www.hiraeth.com/activeweb
Dave ClarkeEditorial - Interfaces for the Active Web (Part 2)
Alan Dix
| Active Web home | special issue | part 1 | part 2 | conference | tutorial | |
| Guest Editors Dave Clarke and Alan Dix | pages hosted by Hiraeth Mixed Media and Visualize Software Ltd |