The World Wide Web and emerging Internet resource discovery standards for scholarly literature - Networked Scholarly Publishing

Library Trends, Spring, 1995 by Stuart L. Weibel

One can imagine a future in which scholarly documents contain dynamic data objects that, through the click of a button, will launch virtual experiments or demonstrations based on the published data and the models, all mediated through a universal browser that serves as a familiar entry point to all the user's information resources.

Current Limitations of the World Wide Web

The problems that currently exist with the Web as a document delivery medium fall out into two broad categories: (1) representation and rendering of complex text, and (2) the stateless model of Web transactions.

Representation and Rendering Problems

There are three major impediments to the representation of complex scholarly material in current Web browsers:

1. Mathematical Representation: Mathematical notation is among the most challenging aspects of any typographic system. HTML does not currently support mathematical notation at all; all such objects must be represented as rendered bit images (simple graphics).

2. Formal Models for Tables: Tables are notoriously complex as typographic objects. In an electronic delivery medium, the normal problems of expressing tabular material in a coherent notation are compounded by the need to express the underlying logical structure of tables in such a manner that they can be parsed by software that can read tables for the visually impaired (or convert them to braille). Currently, HTML supports neither objective, though work is underway to substantially improve the table model so that more complex information will be represented effectively, both logically and for display.

The recent submission of the HTML 2.0 standard for formal standardization under the Internet Engineering Task Force includes markup that will support the International Committee for Accessible Document Design (ICADD) recommendations for marking up electronic text for the visually handicapped. ICADD includes representatives from standards bodies, disabilities organizations, governments, and vendors of software and hardware. The ICADD text transformation process description forms an informational annex to ISO 12083, the SGML application for books and journals (ICADD, 1994). This is a first and important step toward making the Web accessible to those with visual impairments.

3. Character Sets: Currently there are no generalizable methods for displaying alternative character sets in HTML documents, thereby severely limiting the expression of any but a standard character set. At this time, the only characters that implementers can be sure will be present on all platforms for all browsers is the lower 128 characters of standard ASCII. Characters not found in this basic set must be rendered as small bitmapped images. This state of affairs will be rectified in future versions of the HTML standard, but until then, implementors must resort to the contorted work-arounds such as those described earlier in OCLC's SGML to HTML conversion.

Statelessness and Document Retrieval

The Web is an example of a stateless protocol. Each time a browser requests a document from a server, it makes and breaks a network connection, incurring a certain amount of overhead. This keeps the HTTP protocol simple, but when a server and client carry on extended transactions, it is inefficient (imagine hanging up and redialing the phone each time you completed a sentence in a phone call, and the person on the other end forgetting what you said in the previous call!). This is particularly problematic for an information service that benefits from (or requires) a session-based interaction, as is generally the case with reference databases and document retrieval systems. Users may modify a search strategy successively, reducing the size of a retrieval set until it is manageable. Such interactions require maintaining information about the state of a transaction.