Automated linguistic judgement aids, such as spelling and grammar checkers, make probabilistic inferences about the likelihood of linguistic constructs being correct or ideal in context. Although these systems might internally estimate their own reliability, or "certainty", current interfaces in e.g., text editing software do not communicate this information to users. Across two online experiments (total N = 268), we examined whether users could interpret and benefit from this reliability information, presented as likelihood estimates accompanying automated suggestions for rewrites. Using an aided judgement task and an analysis method based on Signal Detection Theory, we manipulated both the correctness and the communicated reliability of a notional aid’s advice. In the first experiment, more reliable advice led to more correct responses and improved decision accuracy. Nevertheless, participants systematically underused even highly reliable advice, and they were generally overconfident in their own decisions. Despite the absence of feedback, confidence ratings were higher for correct responses and for trials with more reliable advice, indicating a degree of metacognitive sensitivity. The second experiment, which used a more reliable aid, replicated these patterns and provided clearer evidence of persistent underuse of reliable advice. These findings show that linguistic judgement aids can meaningfully communicate useful reliability information. However, results also reveal limits in users’ propensity to rely on an automated aid, even when its advice is highly reliable. Further, we show how our research contributes to the methodology of analysis of acceptance of automated advice, and we discuss potential for the design of more transparent writing support tools.

• Malignant stroke is a life-threatening condition, with mortality rates reaching up to 80% among patients managed conservatively. • We developed a novel computational approach that combines advanced biomechanical modelling with deep neural network (DNN) for predicting brain swelling following stroke. • Using in-silico simulations of 3,000 stroke cases, our DNN model demonstrated robust learning of anatomical features, achieving minimal errors in brain swelling prediction. • Without the need for retraining or fine-tuning, our model achieved clinically comparable performance for predicting 3-month stroke outcomes. • To the best of our knowledge, this presents the one of few pioneering computational workflows specifically designed for brain swelling prediction. Malignant stroke is a life-threatening condition, with mortality rates reaching up to 80% among patients managed conservatively. Brain swelling volume and midline shift are pivotal clinical markers for predicting stroke outcomes. However, brain oedema typically peaks two to five days post-stroke onset, which significantly delays the implementation of timely interventions. Early prediction of these markers is therefore critical for enhancing treatment strategies and improving patient outcomes. Predicting brain swelling is inherently challenging due to its biomechanical complexity, governed by factors such as lesion size and location. In this study, we proposed a novel physics-informed computational approach that combines advanced biomechanical modelling with deep neural networks (DNNs) for predicting brain swelling following stroke. Using in-silico simulations of 3,000 stroke cases generated from a poroelasticity-based mathematical model, our DNN model demonstrated robust learning of anatomical features, achieving minimal errors in brain swelling predictions for hold-out test cases. Furthermore, we externally validated the trained model using clinical imaging data from 60 stroke patients; without the need for retraining or fine-tuning, the model achieved clinically comparable outcomes, with an area under the curve (AUC) of approximately 0.7 for predicting 3-month stroke outcomes. Our findings underscore the transformative potential of this physics informed neural networks to accelerate clinical decision-making and improve the management of malignant stroke. By enabling earlier and more precise predictions of critical imaging markers, this pipeline provides a significant step forward in reducing the mortality and disability associated with this devastating condition.

We report an experimental study in which participants make a series of linguistic judgements with help from a fictional text-editing aid. The study illuminates how acceptance of a software aid's advice might be affected by, and in turn affects, users’ trust and confidence.We measure the propensity to accept automated advice by adapting Signal Detection Theory (SDT), and particularly the construct of bias. Our study demonstrates a significant difference in bias between groups who interact with aids that perform at different levels of reliability. Participants’ pre-task trust in similar systems also affected bias towards the aid, but we found no evidence for such an effect of prior perceived self-efficacy.Users show an awareness of their own performance, demonstrated by higher self-reported confidence in correct responses than in incorrect ones. Further, participants using a more reliable aid are more confident when accepting than when rejecting the aid's advice. Overall, participants were overconfident in their judgements.Our research demonstrates opportunities of using an experimental paradigm, analysed with SDT, to study aspects of performance, trust, and confidence in knowledge-based cognitive tasks aided by an artificially intelligent aid.

Over the past few decades, osmotherapy has commonly been employed to reduce intracranial pressure in post-stroke oedema. However, evaluating the effectiveness of osmotherapy has been challenging due to the difficulties in clinical intracranial pressure measurement. As a result, there are no established guidelines regarding the selection of administration protocol parameters. Considering that the infusion of osmotic agents can also give rise to various side effects, the effectiveness of osmotherapy has remained a subject of debate. In previous studies, we proposed the first mathematical model for the investigation of osmotherapy and validated the model with clinical intracranial pressure data. The physiological parameters vary among patients and such variations can result in the failure of osmotherapy. Here, we propose an AI-assisted in silico trial for further investigation of the optimisation of administration protocols. The proposed deep neural network predicts intracranial pressure evolution over osmotherapy episodes. The effects of the parameters and the choice of dose of osmotic agents are investigated using the model. In addition, clinical stratifications of patients are related to a brain model for the first time for the optimisation of treatment of different patient groups. This provides an alternative approach to tackle clinical challenges with in silico trials supported by both mathematical/physical laws and patient-specific biomedical information.

In ischaemic stroke, a large reduction in blood supply can lead to the breakdown of the blood-brain barrier and to cerebral oedema after reperfusion therapy. The resulting fluid accumulation in the brain may contribute to a significant rise in intracranial pressure (ICP) and tissue deformation. Changes in the level of ICP are essential for clinical decision-making and therapeutic strategies. However, the measurement of ICP is constrained by clinical techniques and obtaining the exact values of the ICP has proven challenging. In this study, we propose the first computational model for the simulation of cerebral oedema following acute ischaemic stroke for the investigation of ICP and midline shift (MLS) relationship. The model consists of three components for the simulation of healthy blood flow, occluded blood flow and oedema, respectively. The healthy and occluded blood flow components are utilized to obtain oedema core geometry and then imported into the oedema model for the simulation of oedema growth. The simulation results of the model are compared with clinical data from 97 traumatic brain injury patients for the validation of major model parameters. Midline shift has been widely used for the diagnosis, clinical decision-making, and prognosis of oedema patients. Therefore, we focus on quantifying the relationship between ICP and midline shift (MLS) and identify the factors that can affect the ICP-MLS relationship. Three major factors are investigated, including the brain geometry, blood-brain barrier damage severity and the types of oedema (including rare types of oedema). Meanwhile, the two major types (stress and tension/compression) of mechanical brain damage are also presented and the differences in the stress, tension, and compression between the intraparenchymal and periventricular regions are discussed. This work helps to predict ICP precisely and therefore provides improved clinical guidance for the treatment of brain oedema.

The microvasculature plays a key role in oxygen transport in the mammalian brain. Despite the close coupling between cerebral vascular geometry and local oxygen demand, recent experiments have reported that microvascular occlusions can lead to unexpected distant tissue hypoxia and infarction. To better understand the spatial correlation between the hypoxic regions and the occlusion sites, we used both in vivo experiments and in silico simulations to investigate the effects of occlusions in cerebral penetrating arteriole trees on tissue hypoxia. In a rat model of microembolisation, 25 μm microspheres were injected through the carotid artery to occlude penetrating arterioles. In representative models of human cortical columns, the penetrating arterioles were occluded by simulating the transport of microspheres of the same size and the oxygen transport was simulated using a Green's function method. The locations of microspheres and hypoxic regions were segmented, and two novel distance analyses were implemented to study their spatial correlation. The distant hypoxic regions were found to be present in both experiments and simulations, and mainly due to the hypoperfusion in the region downstream of the occlusion site. Furthermore, a reasonable agreement for the spatial correlation between hypoxic regions and occlusion sites is shown between experiments and simulations, which indicates the good applicability of in silico models in understanding the response of cerebral blood flow and oxygen transport to microemboli.

The inspired sinewave technique (IST) is a non-invasive nonionising method which can measure accurately lung parameters. Six anaesthetised pigs were studied after surfactant depletion by saline-lavage. Paired measurements of lung heterogeneity were taken with the IST and computed tomography (CT). Lung heterogeneity measured by the IST was calculated by the ratio of tracer gas outputs measured at 180s and 60s. In CT images, lungs were manually segmented and divided into different regions according to voxel density. The IST and CT show strong agreement on lung heterogeneity measurements. In the IST ventilation heterogeneity value, the linear correlation in total of six pigs are weaker than individual animal (R2=0.33 vs R2=0.67 on average). The IST perfusion heterogeneity agreements are lower than IST ventilation values due to the IST technical setup. IST has the potential to measure the lung heterogeneity in ARDS animals.

Many ischaemic stroke patients who have a mechanical removal of their clot (thrombectomy) do not get reperfusion of tissue despite the thrombus being removed. One hypothesis for this 'no-reperfusion' phenomenon is micro-emboli fragmenting off the large clot during thrombectomy and occluding smaller blood vessels downstream of the clot location. This is impossible to observe in-vivo and so we here develop an in-silico model based on in-vitro experiments to model the effect of micro-emboli on brain tissue. Through in-vitro experiments we obtain, under a variety of clot consistencies and thrombectomy techniques, micro-emboli distributions post-thrombectomy. Blood flow through the microcirculation is modelled for statistically accurate voxels of brain microvasculature including penetrating arterioles and capillary beds. A novel micro-emboli algorithm, informed by the experimental data, is used to simulate the impact of micro-emboli successively entering the penetrating arterioles and the capillary bed. Scaled-up blood flow parameters-permeability and coupling coefficients-are calculated under various conditions. We find that capillary beds are more susceptible to occlusions than the penetrating arterioles with a 4x greater drop in permeability per volume of vessel occluded. Individual microvascular geometries determine robustness to micro-emboli. Hard clot fragmentation leads to larger micro-emboli and larger drops in blood flow for a given number of micro-emboli. Thrombectomy technique has a large impact on clot fragmentation and hence occlusions in the microvasculature. As such, in-silico modelling of mechanical thrombectomy predicts that clot specific factors, interventional technique, and microvascular geometry strongly influence reperfusion of the brain. Micro-emboli are likely contributory to the phenomenon of no-reperfusion following successful removal of a major clot.

We report a laboratory study (N=53) in which participants browsed their own Facebook news feeds for 10-15 minutes, choosing exactly when to quit, and later rated the overall emotional utility of the episode before attempting to recall threads. Finally, the emotional utility of each encountered thread was rated while looking over a recording of the interaction. We report that Facebook browsing was, overall, an emotionally positive experience; that recall of threads exhibited classic primacy and recency serial order effects; that recalled threads were both more positive and more valenced (less neutral) on average, than forgotten threads; and that overall emotional valence judgments were predicted, statistically, by the peak and end thread judgments. We find no evidence that local quit decisions were driven by the emotional utility of threads. In the light of these findings, we discuss the suggestion that emotional utility might partly explain the attractiveness of reading the news feed, and that an emotional memory bias might further increase the attractiveness of the newsfeed in prospect.

In this paper we consider how visualizations might support judgements about fairness in collaborative work. We present a qualitative investigation of meters, existing artefacts that enable awareness of contributions in the online game World of Warcraft. Through in-depth interviews with game players, we draw attention to the value of meters as tools for self-reflection and group evaluation. Yet we also describe problematic consequences that arise as a result of meters, distinguishing their usage as in-the-moment awareness tools from instruments used after the fact to apportion credit and blame. We argue that representations like meters may be enough to provoke initial questions about fairness, but are likely to undermine the efforts of collaborators if they fail to combine a set of indices that reflect important aspects of individual work in the context of team activity. We consider broader lessons for the way in which future designs might aim to support fairness in collaborative systems, pointing to multidimensional mechanisms adapted specifically to real-time or retrospective usage.

Over the course of the next 10 years, the Internet of Things (IoT) is set to have a transformational effect on the everyday technologies which surround us. Access to the data produced by these devices opens an interesting space to practice discovery based learning. This paper outlines a participatory design approach taken to develop an IoTbased ecosystem which was deployed in 8 schools across England. In particular, we describe how we designed and developed the system and reflect on some of the early experiences of students and teachers. We found that schools were willing to adopt the IoT technology within certain bounds and we outline best practices uncovered when introducing technologies to schools.

Search engines typically return so many results that choosing from the list might be predicted to suffer from the effects of "choice overload". Preliminary work has reported just such an effect [12]. In this paper a series of three experiments was conducted to investigate the choice overload effect in search engine use. Participants were given search tasks and presented with either six or twenty-four returns to choose from. The results revealed that the choice behaviour was strongly influenced by the ranking of returns, and that choice satisfaction was affected by the number of options and the decision time. The main results, from the third experiment, showed that large sets of options yielded a positive effect on participants' satisfaction when they made a decision without time limit. When time was more strongly constrained, choices from small sets led to relatively higher satisfaction. Our studies show how user satisfaction with found information can be affected by processing strategies that are influenced by search engine design features.

Recent research efforts have led to the creation of a number of systems that provide specialised support for collaborative web search. However, the use of these tools has not been studied outside of the laboratory, and as collaborative search becomes increasingly commonplace in everyday life, there is a need to understand whether the support provided by collaborative search systems fits with real-world information seeking practices. In the present study, we deployed two collaborative search tools to pairs of searchers with genuine information needs. We report findings from in-depth interviews conducted after searchers had used their assigned system for an extended period of time. Our findings show how system features were used and appropriated in pursuit of collaboration, throwing light on the way in which collaborative search is conducted in quotidian settings. Theoretical and practical implications of the results are discussed.

For large, spatially and temporally distributed engineering projects, e-mail is a central means for the discussion of engineering work and sharing of digital assets that define the product and its production process. The importance of communication and the value of its content for resolving issues post facto are universally accepted. More recently, the potential value of its content to predict events, issues and states a priori has been explored with some success. However, while in the former context (post facto) trends and patterns can be established through iteration and refinement over time, for prediction, heuristics need to be established in advance and closer to real-time analysis becomes necessary due to the critical and very often short timescales. It is this challenge of making predictions from the content of e-mail that is considered in this paper. In particular, the paper deals with engineering e-mail and the ability to automatically predict its purpose from its content rather than relying solely on the subject line. The work builds upon previous studies by the authors concerning the characterisation of the content of e-mail: what they are about, why they were sent and how the content is expressed. The paper summarises the previous work and looks at the potential of identifying the purpose of e-mail through the use of Naive Bayes and an adapted Latent Semantic Analysis approach. While the techniques have only been applied to an initial exploratory study of 98 e-mails, the results suggest the potential for automated real-time categorisation of engineering e-mails through achieving an accuracy of 66%. Such a capability would both support prioritisation of e-mail for engineers and macro level characterisation of project e-mail dynamics. The latter provides the opportunity for real-time analysis of an engineering projects status and correspondingly, modes of management intervention.

We report a study in which dyads use Instant Messaging to agree a preference among a set of three apartments. The information given to participants is partially overlapping, and contains a “hidden profile” (HP), such that a single apartment emerges as the best according to an unweighted sum of feature values only if dyad members pool information that is presented to only one of them. When dyads were additionally provided with a shared online spreadsheet, their decision strategy was more likely to be compensatory and relatively exhaustive, even if the distribution of importance among the cues in which the apartments vary meant that a “fast and frugal” heuristic such as take-the-best would be a rational strategy. This study shows the potential of classic experimental tasks, the HP task in particular, for understanding technological constraints on group decision making and signals the importance of understanding decision-making strategies, and the potential of fast and frugal heuristics, for informing the design of decision support systems.

Arterial spin labelling (ASL) MRI offers a non-invasive means to create blood-borne contrast in vivo for dynamic angiographic imaging. By spatial modulation of the ASL process it is possible to uniquely label individual arteries over a series of measurements, allowing each to be separately identified in the resulting angiographic images. This separation requires appropriate analysis for which a general Bayesian framework has previously been proposed. Here this framework is adapted for clinical dynamic angiographic imaging. This specifically addresses the issues of computational speed of the algorithm and the robustness required to deal with real patient data. An algorithm is proposed that can incorporate planning information about the arteries being imaged whilst adapting for subsequent patient movement. A fast maximum a posteriori solution is adopted and shown to be only marginally less accurate than Monte Carlo sampling under simulation. The final algorithm is demonstrated on in vivo data with analysis on a time scale of the order of 10min, from both a healthy control and a patient with a vertebro-basilar occlusion.

Readers on the Web often skim through text to cope with the volume of available information. In a previous study, Duggan and Payne [11] tracked readers' eye movements as they skimmed through expository text under time pressure. This article presents novel analyses of these eye-movement data. Results indicated that readers were able to explicitly direct attention to the most important information in the text and that this improved performance on a subsequent test of memory for the meaning of text. We suggest readers achieve this by satisficing reading through text until the rate of information gain drops below threshold and then skipping to the next section of text. Further analyses of gaze patterns for paragraphs and pages supported this explanation. Combining satisficing with some form of scanning or sampling behaviour could explain patterns of reading found on the Web. A greater understanding of the way that text is read on the Web would assist many producers of online content.

There are a number of mobile shopping aids and recommender systems available, but none can be easily used for a weekly shop at a local supermarket. We present a minimal, mobile and fully functional lambent display that clips onto any shopping trolley handle, intended to nudge people when choosing what to buy. It provides salient information about the food miles for various scanned food items represented by varying lengths of lit LEDs on the handle and a changing emoticon comparing the average miles of all the products in the trolley against a social norm. When evaluated in situ, the lambent handle display nudged people to choose products with fewer food miles than the items they selected using their ordinary shopping strategies. People also felt guilty when the average mileage of the contents of their entire shopping trolley was above the social norm. The findings are discussed in terms of how to provide different kinds of product information that people care about, using simple lambent displays.

The importance of background knowledge for effective searching on the Web is not well understood. Participants were given trivia questions on two topics and asked to answer them first using background knowledge and second by searching on the Web. Knowledge of a topic predicted search performance on that topic for all questions and, more importantly, for questions for which participants did not already know the answer. In terms of process, greater topic knowledge led to less time being spent on each Webpage, faster decisions to give up a line of inquiry and shorter queries being entered into the search engine. A more complete theory-led understanding of these effects would assist workers in a whole range of Web-related professions.

We report two studies investigating readers' ability to allocate limited time adaptively across online texts of varying difficulty. In both studies participants were asked to learn about the human heart and were free to allocate time to 4 separate on-line texts about the heart but did not have enough time to read them all thoroughly. Of particular interest was whether readers attempted to select the best text for them (by sampling the texts before reading) or to monitor texts while reading them and continue reading any text judged good enough (a satisficing strategy). We argue that both strategies can be considered adaptive, depending on properties of readers, texts, and tasks. Experiment 1 tested readers with a range of background knowledge and allowed them either 7 or 15 min study time. It showed that participants were adaptive in how they allocated their time in that more knowledgeable readers spent more time reading more difficult texts. Satisficing was a much more common strategy than sampling. Experiment 2 showed that providing outline overviews of each text dramatically increased the number of participants using a sampling strategy so that it became the modal strategy. However, this change in strategy had no effect on learning. Outline overviews presumably changed readers' perception of the ease with which relevant dimensions of text quality can be judged.

Two experiments aimed to explore the general usability of filenames that were either chosen for personal use or designed for another user, within a document-filing scenario. Both experiments revealed a strong benefit for using self-chosen names over using names that were chosen by another for their personal use. However, both experiments also showed that participants could to some extent adapt their chosen names for use by a wider population. In experiment one, users could recall these ‘designed’ names more accurately than names that had been chosen for personal use by another (although this difference was not evident in a recognition test). Experiment two supported this effect; names designed for general use were more accurately recalled by another than the names designed for self-use after two sessions of use. Names designed for general use employed higher frequency words than did the self-chosen names, which supports the view that the benefit for self-chosen names partly lies in their exploitation of idiosyncratic associations.

A Bayesian method is proposed for estimating model parameters from noisy data sets. The method is based on maximizing the posterior kernel, which enables priors on the model parameters to be incorporated. The posterior kernel is found by specifying hyperpriors and integrating the priors out, due to the use of conjugate priors. The use of probability models enables simultaneous data streams to be used to maximize the posterior kernel. The solution is found using an iterative scheme. The algorithm's performance is briefly illustrated using a real data set, demonstrating rapid convergence.

The gene activities in T lymphocytes that regulate immune responses are influenced by Ca2+ ([Ca2+]i). The intracellular calcium signals are highly heterogeneous and vitally important in determining the immune outcome. The signals in individual cells can be measured using fluorescence microscopy but to group the cells into classes with similar signal kinetics is currently laborious. Here, we demonstrate a method for the automated classification of the responses into four categories formerly identified by an expert's inspection. This method comprises characterising the response by a second-order model, performing frequency analysis, and using derived features as inputs to two multilayer perceptron neural networks (NNs). We compare the algorithm's performance on an example data set against the human classification: it was found to classify identically more than 70% of the data, despite small sample sizes in two categories and significant overlap between the other two classes. The group characterized by an oscillating signal showed the presence of a number of frequencies, which may be important in determining gene activation. A classification threshold enables the automatic identification of patterns with a low-classification certainty. Future refinement of the algorithm may allow the identification of more classes, which may be important in different immune responses associated with disease.

Two important phenomena in human - computer interaction (HCI) are considered: the reliance on external information rather than memory, and the interleaving of planning and action. These phenomena are important, it is argued, because they challenge some particular cognitive models. However, we reject those views, influential in the HCI literature, that phenomena like these require radically new conceptions of cognition or behaviour. It is shown that the phenomena are not universal laws of behaviour, but that instead people decide how much to remember and how much to plan according to a consideration of the costs and benefits of different strategies. Thus the classical cognitive conception of humans as adaptive decision makers is vital for a deep understanding of HCI.

An important practical question is: how should instruction for computer skills be designed to facilitate effective learning? The reported study examines the instructional efficacy of animated demonstrations within active and passive learning contexts of teaching basic spreadsheet skills. Four content-matched instructional regimes were compared: the commercially available tutorial (a ‘scenario machine’), an animated demonstration of this tutorial being used, and problem solving supported by either the user manual or a set of task-specific demonstrations. Acquired spreadsheet skills were then tested on a standard task. Results indicate a clear learning advantage of problem solving, over prompted interaction (the scenario machine). The study suggests two distinctive roles that animations could exploit within computer instruction. Simply watching an animated demonstration can provide a useful introduction to complex interfaces; additionally, animations can be an effective ‘example following’ resource for more active problem-solving.

Four experiments were performed to test the relationship between instructionally derived knowledge and practice in the use of a simple device. Using a derivative of Kieras and Bovair's (1984) device, we show that different subjects can be given instructions that convey equivalent information but that lead to crossovers in the time to perform different tasks (i.e., one task is easier with one set of instructions, a second task is easier with other instructions). Experiment 1 shows that the performance crossover between question types perseveres when subjects relinquish the instructions, after they have been committed to memory. Experiment 2 shows that the performance crossover perseveres over considerable experience using the device. Experiment 3 shows that the crossover can disappear if sufficient practice is given with the particular question types. Experiments 2 and 3 taken together suggest that subjects may only be able to overcome the computational disadvantages of their initial instructional material by adopting task-specific strategies. Experiment 4 shows that when new problems are introduced after the point at which the crossover disappears then a new crossover appears, implying that, even with extended practice of operating the device and solving problems on the device, some features of the initial instructional device description are preserved and continue to determine the users' behavior. We argue that a definition of internalization coupled with Anderson's (1983, 1987) ACT* theory of skill acquisition provides a good account of these results.

This article is an interview study of calendar use and a cognitive analysis of the interactions between the design of calendars and the task of prospective remembering. The study and analysis are coordinated to present a general critique of current electronic calendar designs and to note opportunities for future designs. The interview data reveal continued use of paper calendars in a highly computerized setting. A key conclusion is that paper calendars support prospective remembering by promoting browsing of existing appointments during subsequent calendar keeping but that this advantage is compromised in current electronic designs. Other aspects of the interviews and the analyses address the representational limitations of both paper and electronic calendars. This research explores a methodology in which design implications of qualitative empirical data are understood and systematized through theoretical analyses of existing artifacts.

This article advances a heuristic for conceptual instruction, based on the yoked‐state space (YSS) hypothesis about the contents of users’ device models (Payne, 1987a; Payne, Squibb, & Howes, 1990). The YSS hypothesis suggests that users of a system need to understand the system's representation of the task domain. Accordingly, conceptual instructions should express the mapping from device objects onto goal‐space objects, especially those aspects that are not visible at the user interface. Such instructions are developed for a menu‐driven computer system based on the RATES system for performing remote diagnosis of telephone lines. A simple comparative experiment shows that novices who receive these instructions learn to use RATES more quickly than novices who receive only background instruction and a brief procedural manual. These results increase empirical support for the YSS hypothesis, and, at the same time, suggest a heuristic for the design of conceptual instructions.

In an animated demonstration the device's display behaves as if the device were in use. We investigate the instructional potential of a ‘pure’ version of animated demonstrations, in which there is no commentary or supporting documentation for the animation. Such animations are heavily used in video games (such as PacMan), but are rarely seen in computer-based office systems, although modern designs allow them to be readily and meaningfully implemented. We report two experiments that test the efficacy of animated demonstrations as an aid to exploratory learning of the MacDraw graphics editor. The animated demonstration is simulated by playing a short, uncommented, silent video recording of the screen-in-use. Experiment 1 shows that this technique offers large learning advantages over a no-instruction condition, and is, in our limited experiment, as effective as text-based instruction. Experiment 2 replicates the main effect of Experiment 1, again revealing a significant positive effect of a short animated demonstration on first-time exploratory learners. A notable feature of uncommented animated demonstrations is that they do not offer a complete method specification for the performance of any tasks. In the face of this limitation, their success as instructions can perhaps best be understood in terms of psychological models of performance that do not demand complete cognitive encoding of plans or methods.

This paper addresses the problem of how a novice computer user, engaged in exploratory learning, accounts for the behaviour of the device. Exploratory learning is the norm for many users who encounter computers in the work place. Exploratory learners must acquire methods from a suboptimal stream of task directed behaviour and its observable effects.

To construct a conceptual model of a device, the user must conceptualize the device's representation of the task domain. This knowledge can be represented by three components: a device-based problem space, which specifies the ontology of the device in terms of the objects that can be manipulated and their interrelations, plus the operators that perform the manipulations; a goal space, which represents the objects in terms of which user's goals are expressed; and a semantic mapping, which determines how goal space objects are represented in the device space. The yoked state space (YSS) model allows an important distinction concerning the mental representation of procedures. If a step in a procedure specifies a transformation of the user's device space, then it has an autonomous meaning for the user, independent of its role in the sequence or method. The device space provides a figurative account of the operator. However, some operators do not affect the minimal device space, and their only meaning for the user derives from their role in a method: The method affords an operational account of the operator. Figurative accounts can be constructed from operational accounts only by elaborating the device space with new concepts. The YSS is illustrated through a simple description of a device model for a cut-and-paste text editor. Three experiments addressed the claims of this model. The first experiment used a sorting paradigm to show that users do acquire the novel device space concept of a string of adjacent characters (including space and return). The second and third experiments asked novices to make inferences about text editor behavior on the basis of simple demonstrations. They showed that (a) the availability of the string concept is critically dependent on the details of interface design, (b) figurative accounts of the copy operation afford more efficient methods and may be promoted by appropriate names for procedure steps, and (c) a conceptual model may transfer from one device to another. Together, the three experiments supported the YSS hypothesis.

A formal model of the mental representation of task languages is presented. The model is a metalanguage for defining task-action grammars (TAG): generative grammars that rewrite simple tasks into action specifications. Important features of the model are (a) Identification of the "simple-tasks" that users can perform routinely and that require no control structure; (b) Representation of simple-tasks by collections of semantic components reflecting a categorization of the task world; (c) Marking of tokens in rewrite rules with the semantic features of the task world to supply selection restrictions on the rewriting of simple-tasks into action specifications. This device allows the representation of family resemblances between individual task-action mappings. Simple complexity metrics over task-action grammars make predictions about the relative learnability of different task language designs. Some empirical support for these predictions is derived from the existing empirical literature on command language learning, and from two unreported experiments. Task-action grammars also provide designers with an analytic tool for exposing the configural properties of task languages.

Users perceive consistency and inconsistency in syntax, and family resemblances among syntactic constructions. These factors are not captured in conventional BNF-like grammars. We argue that a generalised form of a two-level grammar is a better model of the user's perceptions, and show how the model relates to current psychological notions of organisation in recall and language learning. The model provides a unified interpretation of many previous results in HCI: we analyse here findings by Reisner (1981) and Barnard et al. (1981). Two preliminary experimental tests supporting the model are described.