|Table of Contents|
|Table of Contents|
The intended receivers (or `target audience') of scientific articles are peers, i.e. scientists who have approximately the same background as the author. Nevertheless, as we already stated in section 2.1.3, scientific communication via articles is heterogeneous: not all members of the target audience are equally familiar with the subject. The receivers' `degree of familiarity' ranges from the least-informed to the most-informed receiver.2.39 The most informed receiver is the expert. The question as to who is the least informed receiver, i.e. what is the minimum of knowledge on a specified subject that the receivers are presumed to have, depends on the domain and the level of specialism of the article. The `degree of familiarity' with the subject affects the receivers' goals and their requirements for effective and efficient communication.
Receivers acquiring documents: searchers and browsers
The receiver can have different specific goals in the stage of information acquisition: 1) to find specific information, 2) to look for new ideas, and 3) to keep informed.2.40 To achieve these goals, the scientist engages in different activities.
To locate the desired type of information in activity (11), the receiver needs the means to differentiate between information sources, with respect to their nature and quality. Therefore, articles are required to be certified, as an outcome of the quality control (6).
Also, the following technical requirements must be satisfied: the article representing the relevant information has to be encoded (by the author in (4), and in particular by the publisher in (8)) in such a way that it can be archived efficiently (in activity (10)). And the receiver needs adequate searching and browsing tools at his disposal.
To see what is available, the scientist browses. In other words, he performs a semi-directed or an undirected search, skimming articles or tables of contents of journals, or wandering through larger repositories of information (e.g. the shelves in the library or an electronic documentbase ). The browser's goal is to encounter interesting information.2.41 Informed browsers mostly monitor the information system to keep informed [Ellis et al., 1993]. Scientists who are less familiar with the subject rather look for new ideas, in what is called in [Chang, 1993] `serendipity browsing'.
Especially for the less-informed browser, both the article itself and the larger depository of documents (e.g. the library) are required to be structured clearly and relevantly. This implies that the structure needs to be compatible with what the receiver expect and that it has to be made explicit. Therefore, both the documents and the connections between them must be characterised properly, so that the browser can decide quickly whether a particular document is worth inspecting or not. Making the structure explicit in a `road map' would greatly assist the browser.
In order to obtain a clear structure, it has to be properly conceived on the conceptual level in (2) and properly realised in the writing activity (3). The archives have to be clearly structured in (10).
The informed browser mainly tries to keep abreast of the latest developments in his field. Therefore, he wants to be alerted when potentially relevant information is published (9) . The time delay in the publication process has to be minimised, in order to ensure the actuality of the information. For the sake of efficiency, the informed browser wants to see immediately what is the novel information (as opposed to the background information) that warranted the publication of the article.
In order to find specific information, the scientist performs a directed search. For example, he can search for a number or a definition, for advice to help him solve a problem, for a review on a specific subject, or for the precise address of an author. The more informed the searcher is about the subject, the better he is able to perform such a search, either with or without the support of an information specialist assisting in the technical aspects of the manipulation of the information system.
The basic requirements for searching information are completeness and findability: if the particular information unit the receiver needs exists, 1) it has to be available, and 2) the receiver has to be able to locate it.
For completeness, all potentially relevant information has to be taken into account in the author's organisation of the information at the conceptual level (2). A fundamental problem concerning the completeness is that the author does not know what information could be relevant in other contexts, even if he strives for the completeness of what he thinks essential. Then, all potentially relevant information is required to be expressed in the article in (3) (the actual writing of the article) and subsequently encoded in (4). Relevant information can be lost for technical reasons in these activities.
Completeness also implies that all articles containing relevant information must be made available. They risk being lost altogether, when they are not included in some appropriate collection (in activity (5)). In the quality control (6), relevant information can get lost, if it is rejected along with incorrect information. Furthermore, the information should not stay in the `publication pipeline' too long, but it has to be published (9) quickly. Once articles have been made available, the searcher requires that they remain available, i.e. that the articles are properly archived (10). Access to the archives and the security of the articles have to be guaranteed permanently.
Supposing that relevant information is available, the next step is to find it. For that purpose, a correct characterisation is indispensable. The information has to be adequately characterised, and that characterisation has to be represented at the symbolic level in a set of index terms, which then must be associated to the signal encoding the document.
The characterisation must be correct (i.e. the index terms must correspond to the information they describe), clear and relevant (i.e. they should describe the information in a way that fits the receivers' expectations and needs). It is also required to be complete (i.e. all information must be characterised), to reduce the risk of the receiver missing the target altogether. In terms of Information Retrieval, this point is mostly concerned with the `recall' of the search result.2.42 For the enhancement of the `precision' of the search result, the characterisation must also be sufficiently precise: it should be narrow (i.e. the combined index terms should not be more general, for example, than the information) and fine-grained, in the sense that the representation of each different type of information that the receiver may be looking for should carry its own index terms.
The information that the searcher is trying to locate is not necessarily what is represented in an entire article. Therefore, the precision of the search not only implies that no irrelevant documents should be retrieved, but also that the retrieved documents that contain relevant information do not contain any irrelevant information, or that the representation of the relevant information is highlighted within the document. Therefore, smaller publication units than the article have to be explicitly labelled in the case that an article is composed of parts representing different types of information. In the next chapter, we shall see that in a modular article, each module carries a unique characterisation.
Then, the receiver wants to retrieve the article quickly, easily and cheaply in activity (12). Therefore, there should not be insurmountable technical or financial barriers blocking the archives.
Receivers assimilating the information contained in documents: comprehensive and selective readers, and users
For the transmission (and in particular storage) of the document, the symbolic representation of the information may have been translated into some computer language. In that case, the receiver has to decode the signal (13) in order to obtain the article itself. For example, if information represented in terms of a picture has been transmitted as a computer file in a picture format, it has to be translated back to the pictorial language that the (human) receiver can interpret. For this purpose, the receiver requires appropriate `tools'. In cases of exotic or obsolete encoding formats such tools may be unavailable.
The receiver aims to learn from the article and to use the information represented in it in his own research. When consulting a scientific article, he therefore has the following specific goals, which reflect the author's interactional goals: 1) to get informed, 2) to evaluate the reliability of the information, and 3) to evaluate the relevance of the information in the light of his own needs. For this purpose, the receiver can read the article comprehensively or selectively in activity (14) of the communication process.
In the case of comprehensive reading, the receiver consults the article as a whole, following the discourse linearly from the beginning to the end. Traditionally, articles are written to be read in this way. According to [Bazerman, 1985], most comprehensive reading is for self-instruction in areas beyond the intimate knowledge of the receiver. Thus, the receiver studies the article as a tutorial [Dillon et al., 1989]. He has to be sufficiently informed to be able to read an article written for peers comprehensively without additional background information. The most informed receivers may read an article comprehensively, in order to judge its quality. However, they are not always interested in reading the entire article, being already familiar with most of its contents.
In the case of selective reading, the scientists do not consult the article as a whole. Instead, they extract only the specific parts they need. Bazerman has interviewed informed physicists, to analyse their reading strategies. He found that the majority of the interviewees read the larger part of the articles selectively [Bazerman, 1985, p.11]. Receivers do not consult the article sequentially either. Instead, they start with the interesting part and then jump to additional information. Rather than studying the article like a tutorial, the `selective receiver' consults the article more like a reference book. A special type of `selective reader' consults the meta-information of an article, rather than its scientific content.2.43 This role is played by non-scientists for administrative purposes, but also by scientists who want, for instance, to contact an author or to find a specific reference.
To allow the receiver to assimilate the information represented in an article, it has to be structured and expressed in a clear and readable way, so that the receiver can absorb (15) and evaluate (16) the information easily. In particular, the coherence of the information has to be expressed. In this context, the problem of incompleteness, already mentioned for the acquisition of information, also affects the receiver: explanations, extensive calculations and interesting proofs cannot be replaced by a statement like ``you can easily show that''.
We assume that, for the sake of clarity, the distinction between old and new work has to be made explicit. We also assume that the receiver wants the main claim of the article to be highlighted and that the receiver wants to be able to asses the contents of the article and evaluate its potential interest quickly. Therefore, it also has to be clear how the article fits in its context.
For comprehensive reading, readability implies that the receiver is guided smoothly from the beginning of the article to the end, while in the structure the main issues are separated from side-issues. For the purpose of selective reading, the article has to be organised in small coherent units, so that the relevant specific information is not encumbered by the rest. For the selective consultation of the meta-information, these data in particular are required to be grouped and clearly structured. Thus, comprehensive readers and selective readers have conflicting requirements. If receivers must be able to read same article in two different ways, the article will have to be structured in a particularly clear and flexible way.
Furthermore, the article has to be produced (8), published (9) and stored (10) in such a way that its (physical or electronic) `technical presentation' is either directly user-friendly, or easily re-formatted as such. The reader has to be able to transport and store the article for (re)use at a later time, to compare information and annotate the article.
If the receiver judges the information to be sufficiently reliable and relevant, he can incorporate it in his own work. For instance, he can utilise it directly as input in a calculation, as a guideline for the fine-tuning of his own experimental set-up, or for comparison with his own results. This implies, from a technical point of view, that the information has to be presented and stored in a manipulable form.
The information can only be used wisely if it is complete, in particular with respect to its restrictions. The more familiar the scientist is with the subject, the more natural it is for him to use information from the article, without first reading the whole story. If the receiver does not accept the information, he can contribute to science by explicitly refuting that information.
In the context of scientific articles, this implies that the information has to be controllable. Therefore, the source of cited information must be made explicit and the acquisition of the cited material has to be easy, so that the information can easily be checked. Furthermore, the receiver requires assistance in judging the article, in particular if he is not sufficiently well-informed to recognise unreliable information as such, or if he wants to use the information directly. Therefore, certification plays a crucial role in this stage, informing the receiver about the reliability of the information.
Related to the requirement of scientific quality is that of the integrity of the information. In the archives, the article has to remain `intact'. However, the archives must be flexible enough to allow for the addition of new versions, comments and explicit corrections, to keep the information up to date and ensure that obsolete information can be recognised as such. The original version of a published article must not be changed.
Summarising, we have discussed the following requirements of the interactants in the process of scientific communication.
As a prerequisite, the communication system in general has to fulfil the following requirements, including dissemination requirements. It should:
Concerning the article, the author has the following requirements
The scientific article itself has to meet the following criteria to satisfy the receiver: