|andrew.treloar.net||Andrew Treloar's personal website|
A range of navigation mechanisms are possible. This section will consider first the question of navigation granularity and then browsing and searching.
Navigation granularity can be defined as the smallest item that can be navigated to or cited. A range of increasingly fine granularities is possible, depending on how much information has been provided: Journal, Issue, Article, Section. Beyond the level of a section within an article, there are some difficulties.
The most common navigation mechanism in the print world is page numbers. This presupposes that the page cited will be the same in all versions of the work referred to. This is a safe assumption as long as editions and reprints are carefully flagged in the citation. E-journals in general have no direct equivalent to page numbers. Some first-generation e-journals, such as PACS-R introduced 'page numbers' every 60 lines into their ASCII articles. Unfortunately, these page numbers did not align with those in the parallel print incarnation of the journal. This meant that locating a particular cited page required the citation to include which of the versions of the article was intended. PDF-based e-journals usually inherit the page numbers of the print versions they replicate and so do not have this problem. HTML-based e-journals cannot have page numbers because the concept of page is meaningless in a user-reformattable document space. Changes in font size, browser window dimensions or paper used for printouts will all affect the size of pages. Some journals are experimenting with making the citation granularity more useful and introducing either paragraph numbers (long the practice in legal circles) or defining visible HTML anchors (which can then be linked to) for each section heading.
Chemical Journals Online (a commercial full-text service run by the American Chemical Society using STN network which includes all of the ACS journals as well as those of several other major publishers of chemistry) numbers the paragraphs of the online versions of articles. This allows users of its service to just look at those paragraphs referenced in a search set.
Browsing through print is actually a very sophisticated set of activities, many of them unconscious (as noted in [Olsen, 1994]). Users make use of tables of contents, highlighted articles on the front cover of a journal, known 'favourite sections' of a particular journal, flicking quickly through all the pages, skimming abstracts and authors names, consulting the index at the end of the issue (if it exists) and so on. These are skills acquired over years of use of these particular communication artefacts and are, in part, tied to the affordances of that artefact. The serendipitous nature of browsing (as opposed to searching) was an important factor for a number of Olsen's respondents.
Browsing in the electronic world is constrained by the inability to quickly 'flick' through content in the same way, and by the similarity in appearance of many online journals. Browsing mechanisms therefore have to use the affordances of the online media in different ways. Examples include use of images of front covers to select a particular issue, hyperlinked tables of contents, and hierarchical arrangements of content by year, then issue, then article.
Searching print on a small scale can be done manually. Skimming an article is partly a process of searching for words or phrases (although these may not have been decided on before starting). For larger amounts of text, the user needs to work with document surrogates such as indexing or abstracting services, usually in some computerised form.
The searchability of e-text is of course one area where it provides significant additional functionality over print. Because the entire content of an online journal is already machine-readable, it can be indexed relatively easily (depending on its format) allowing for free-text searching across issues or even sets of journals.
Last modified: Monday, 11-Dec-2017 14:38:54 AEDT
© Andrew Treloar, 2001. * http://andrew.treloar.net/ * firstname.lastname@example.org