Models of computation

Instructor's Guide

introduction objects web jamming architecture summary, Q/A, literature
The Programming the Web workshop was intended to focus on concepts and requirements for high-level APIs suitable for developing Web-aware applications. The papers that were submitted, which are available on the CDROM accompanying this book, covered a wide range of interests, including computation models, applications and user requirements, software architectures and libraries, as well as heuristics and guidelines for API developers.


Complaints


Observations


Requirements


Behavior


Answers

slide: Requirements for APIs
 The kickoff for the workshop was given by Luca Cardelli, who raised the question `What is the Web's model of computation?'. This question appeared to be of critical importance for understanding the requirements for APIs and for evaluating possible solutions. In summary, we may observe that there is some notion of global computation for the Web, but that computation on the Web is fraught with many obstacles, such as the lack of referential integrity (e.g. dead links), unreliable services (both in availability and quality), failures (due to servers or network congestion). What we need, in conclusion, is some (formal) model of computation that captures these properties. In addition, we need to be able to deal with such properties in our Web programs, for example we may wish to anticipate on the possible unavailability of a Web server, and provide an alternative in that case.

In slide api-requirements, an overview is given of the complaints about the functionality of the Web, observations concerning its `nature', general requirements for open systems development, a wish-list of desired behavioral characteristics and potential (technological) answers.


Actions


Perspectives


Interests

slide: Dimensions of APIs
 Not surprisingly, there did not seem to be a canonical approach to the definition and development of APIs and Web applications, perhaps not in the last place because the demarcation between computation models, languages and APIs is not clear-cut.

Nevertheless, as summarized in slide api, it seemed clear that we need to define a suitable computation model as well as (abstract) object models that capture the requirements for resources and application domains (such as for example e-commerce). In addition we must distinguish between client and server perspectives, with autonomous (intelligent) agents as a possible third perspective. And, naturally, our own (technological) interests play a role as well, to the extent that it may determine possible solutions.

Considering the basic needs for the development of Web-aware applications, as expressed by the workshop's participants, which ranged over resolving URLs, billing and payment facilities, and quality of service constraints, we may observe that facilities for Web programming are nowadays as a standard provided (as extensions) by languages such as Tcl, Perl, Python and Java. More domain-specific facilities are being developed in a CORBA context, or for frameworks such as San Francisco.

Document Object Model

Client-side scripting has been popularized by Dynamic HTML (DHTML) as originally introduced by Netscape and Microsoft. Nevertheless, scripting facilities are not standard accross the various browsers. To remedy this situation, the W3C has developed a recommendation for a Document Object Model (DOM), that provides a standard application programmer interface to access the structure and content of HTML and XML Web pages. The DOM allows XML and HTML pages to be treated in an object-oriented way, providing facilities for access, navigation and manipulation.
slide: Hierarchical structure of DOM
 Since XML is increasingly being used for other applications, such as Electronic Data Interchange (EDI), the DOM may in effect provide a foundation for developing Web applications.

The W3C DOM Recommendation provides interfaces, described in a language and platform-independent way in IDL, for the structural components that may be used in XML and HTML documents, as indicated in slide dom-structure. These interfaces have been refined independently for both XML and HTML, to allow programmers to access XML and HTML-specific features. In addition to the IDL interfaces, a language-binding is specified for ECMAscript, which may serve as an example for similar bindings for Javascript and other languages, such as Java.