'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

O repertório musical português no ensino da tuba

Relatório de Estágio para a obtenção do grau de Mestre em Ensino da Música

Trajetória profissional de professoras do ensino médio : sentidos e usos de tecnologias digitais na sala de aula

Tese (doutorado)—Universidade de Brasília, Instituto de Psicologia, Programa de Pós-Graduação em Processos de Desenvolvimento Humano e Saúde, 2016.

Cognitive Assessment System (CAS): Psychometric studies with Portuguese children from 7 to 15 years.

The Cognitive Assessment System (CAS) is a new measure of cognitive abilities based on the Planning, Attention, Simultaneous and Successive (PASS) Theory. This theory is derived from research in neuropsychological and cognitive Psychology with particular emphasis on the work of Luria (1973). According to Naglieri (1999) and Naglieri and Das (1997), the PASS cognitive processes are the basic building blocks of human intellectual functioning. Planning processes provide cognitive control, utilization of processes and knowledge, intentionality, and self-regulation to achieve a desired goal; Attention processes provide focused, selective cognitive activity and resistance to distraction; and, Simultaneous and Successive processes are the two forms of operating on information. The PASS theory has had a strong empirical base prior to the publication of the CAS (see Das, Naglieri & Kirby, 1994), and its research foundation remains strong (see Naglieri, 1999; Naglieri & Das, 1997). The four basic psychological processes can be used to (1) gain an understanding of how well a child thinks; (2) discover the child’s strengths and needs, which can then be used for effective differential diagnosis; (3) conduct fair assessment; and (4) select or design appropriate interventions. Compared to the traditional intelligence tests, including IQ tests, the Cognitive Assessment System (CAS) has the great advantage of relying on a modern theory of cognitive functioning, linking theory with practice.

Show

Script for non-verbal performance. Research Component: Silent Treatment: Creating Non-verbal Performance Works for Children The research field of theatre for young people draws on theories of child development and popular culture. SHOW explored personal and social development, friendship and creative play through the lens of the experience of girls aged 8-12. This project consolidated and refined innovative approaches to creating non-verbal theatre performance, and addressed challenges inherent in the creation of a performance by adults for young audiences. A significant finding of the project was the unanticipated convergence of creative practice and research into child behaviour and development: the congruence of content (Female bullying) and theatrical form (non-verbal performance: “Within the hidden culture of aggression, girls fight with body language and relationships instead of fists and knives. In this world, friendship is a weapon, and the sting of a shout pales in comparison to a day of someone’s silence. There is no gesture more devastating than the back turning away Simmons, Rachel (2002:3) Odd Girl Out: The Hidden Culture Of Aggression In Girls Schwartz Books The creative development and drafting process focussed on negotiating the conceptual design and practical constraints of incorporating diegetic music and video sources into the narrative. The authorial (and production) challenges of creating a script that could facilitate the re-mount a non-verbal work for a company specialising in text-based theatre . Show was commissioned by the Queensland Theatre Company in 2003, toured into Queensland Schools by the Queensland Arts Council and in 2004 was performed at the Sydney Opera House.

Systemic racism: the hidden barrier to educational success for Indigenous school students

Explanations for poor educational experiences and results for Australian Indigenous school students have, to a great extent, focused on intended or conscious acts or omissions. This paper adopts an analysis based on the legislation prohibiting indirect racial discrimination. Using the elements of the legislation and case law it argues that apparently benign and race-neutral policies and practices may unwittingly be having an adverse impact on Indigenous students' education. These practices or policies include the building blocks of learning, a Eurocentric school culture. Standard English as the language of assessment, legislation to limit schools' legal liability, and teachers' promotions.

In the mouth of the imagination : positioning children as co-researchers and co-artists to create a professional children's theatre production

In recent decades a number of Australian artists and teacher/artists have given serious attention to the creation of performance forms and performance engagement models that respect children’s intelligence, engage with themes of relevance, avoid the cliche´s of children’s theatre whilst connecting both sincerely and playfully with current understandings of the way in which young children develop and engage with the world. Historically a majority of performing arts companies touring Australian schools or companies seeking schools to view a performance in a dedicated performance venue engage with their audiences in what can be called a ‘drop-in drop-out’ model. A six-month practice-led research project (The Tashi Project) which challenged the tenets of the ‘drop-in drop-out’ model has been recently undertaken by Sandra Gattenhof and Mark Radvan in conjunction with early childhood students from three Brisbane primary school classrooms who were positioned as co-researchers and co-artists. The children, researchers and performers worked in a complimentary relationship in both the artistic process and the development of product.

Applying Keller’s brand equity model in a B2B context : limitations and an empirical test

It has been recognised that brands play a role in industrial markets, but to date a comprehensive model of business-to-business (B2B) branding does not exist, nor has there been an empirical study of the applicability of a full brand equity model in a B2B context. This paper is the first to begin to address these issues. The paper introduces the Customer- Based Brand Equity (CBBE) model by Kevin Keller (1993; 2001; 2003), and empirically tests its applicability in the market of electronic tracking systems for waste management. While Keller claims that the CBBE pyramid can be applied in a B2B context, this research highlights challenges of such an application, and suggests changes to the model are required. Assessing the equity of manufacturers’ brand names is more appropriate than measuring the equity of individual product brands as suggested by Keller. Secondly, the building blocks of Keller’s model appear useful in an organisational context, although differences in the subdimensions are required. Brand feelings appear to lack relevance in the industrial market investigated, and the pinnacle of Keller’s pyramid, resonance, needs serious modifications. Finally, company representatives play a role in building brand equity, indicating a need for this human element to be recognised in a B2B model.

An Improved Synthetic Route to the Potent Angiogenesis Inhibitor Benzyl Manα(1→3)-Manα(1→3)-Manα(1→3)-Manα(1→2)-Man Hexadecasulfate

An improved synthetic route to α(1→3)/α(1→2)-linked mannooligosaccharides has been developed and applied to a more efficient preparation of the potent anti-angiogenic sulfated pentasaccharide, benzyl Manα(1→3)-Manα(1→3)-Manα(1→3)-Manα(1→2)-Man hexadecasulfate, using only two monosaccharide building blocks. Of particular note are improvements in the preparation of both building blocks and a simpler, final deprotection strategy. The route also provides common intermediates for the introduction of aglycones other than benzyl, either at the building block stage or after oligosaccharide assembly. The anti-angiogenic activity of the synthesized target compound was confirmed via the rat aortic assay.

"Theatre of the dancing language" : new possibilities in contemporary Australian playwrighting

This study focuses on trends in contemporary Australian playwrighting, discussing recent investigations into the playwrighting process. The study analyses the current state of this country’s playwrighting industry, with a particular focus on programming trends since 1998. It seeks to explore the implications of this current theatrical climate, in particular the types of work most commonly being favoured for production. It argues that Australian plays are under-represented (compared to non-Australian plays) on ‘mainstream’ stages and that audiences might benefit from more challenging modes of writing than the popular three-act realist play models. The thesis argues that ‘New Lyricism’ might fill this position of offering an innovative Australian playwrighting mode. New Lyricism is characterised by a set of common aesthetics, including a non-linear narrative structure, a poetic use of language and magic realism. Several Australian playwrights who have adopted this mode of writing are identified and their works examined. The author’s play Floodlands is presented as a case study and the author’s creative process is examined in light of the published critical discussions about experimental playwriting work.

Remediating vernacular creativity : photography and cultural citizenship in the Flickr photosharing network

Read through a focus on the remediation of personal photography in the Flickr photosharing website, in this essay I treat vernacular creativity as a field of cultural practice; one that that does not operate inside the institutions or cultural value systems of high culture or the commercial popular media, and yet draws on and is periodically appropriated by these other systems in dynamic and productive ways. Because of its porosity to commercial culture and art practice, this conceptual model of ‘vernacular creativity’ implies a historicised account of ‘ordinary’ or everyday creative practice that accounts for both continuity and change and avoids creating a nostalgic desire for the recuperation of an authentic folk culture. Moving beyond individual creative practice, the essay concludes by considering the unintended consequences of vernacular creativity practiced in online social networks: in particular, the idea of cultural citizenship.

A news story as big as a doctoral thesis? : deploying journalistic methodology in academic research

The purpose of this paper is to identify and recommend the emergence of an academic research methodology for Journalism the academic discipline, through reviewing various journalistic methods of research – those making up a key element in such methodology. Its focus is on journalistic styles of work employed in academic contexts especially research on mass media issues. It proposes that channelling such activity into disciplined academic forms will enhance both: allowing the former to provide more durable and deeper outcomes, injecting additional energy and intensity of purpose into the latter. It will briefly consider characteristics of research methodologies and methods, generally; characteristics of the Journalism discipline, and its relationship with mass media industries and professions. The model of journalism used here is the Western liberal stream. A proposition is made, that teaching and research in universities focused on professional preparation of journalists, has developed so that it is a mature academic discipline. Its adherents are for the most part academics with background in journalistic practice, and so able to deploy intellectual skills of journalists, while also accredited with Higher Degrees principally in humanities. Research produced in this discipline area stands to show two characteristics: (a) it employs practices used generally in academic research, e.g. qualitative research methods such as ethnographic studies or participant observation, or review of documents including archived media products, and (b) within such contexts it may use more specifically journalistic techniques, e.g. interviewing styles, reflection on practice of journalism, and in creative practice research, journalistic forms of writing – highlighting journalistic / practitioner capabilities of the author. So the Journalism discipline, as a discipline closely allied to a working profession, is described as one where individual professional skills and background preparation for media work will be applicable to academic research. In this connection the core modus operandi will be the directly research-related practices of: insistent establishment of facts, adept crafting of reportage, and economising well with time. Prospective fields for continuing research are described:- work in new media; closer investigation of relations among media producers and audiences; journalism as creative practice, and general publishing by journalists, e.g. writing histories.