Search
This issue includes advance information on the CHEST deal for the two image processing packages ERDAS from Remote Sensing Services and VISILOG from Data Cell Ltd, reported by Steve Morgan. He also provides some further information on the CHEST deal for the PC drawing software MicrograftDesigner.
There is also an article on NORGRAF - the Northern Area Graphics Meetings. One issue that was raised at the NORGRAF meeting on 21 December was the need for Computer Centres to think about HCI aspects of GUIs. Computer Centres are increasingly providing tools and application builders for general use, with virtually no guidelines on recommended styles or interfaces. This issue has also surfaced in the context of the Universities Funding Council (UFC) Teaching and Learning through Technology Programme (TLTP) where the development of CBL/Multimedia teaching materials needs to be standardised in some sense, both for ease of use and portability of software to other platforms and/or application areas and also other institutions. If anyone has any thoughts on this issue, please let us know!
What will future computer interfaces look like? Read the report on a recent M1T Media Laboratory Symposium on Interface Agents to get a glimpse over the horizon!
AGOCG have held a number of events recently which have attracted good numbers and have been interesting events. The workshops are written up in Technical Reports which can be obtained by E-mailing Joanne Barradell (J.T.Barradell@uk.ac.lut) or through me.
The workshop discussed the topic of Managing the Graphics Common Ground. It was clear that there was a lot of common ground that was recognised by both service departments who were represented at the day which attracted over 40 people. The discussions looked at the technical advances which led to this overlap in work - the use of video, output onto 35mm slides, the advent of multimedia systems and so on. It was apparent that many of the problems had been faced by one service or another before - the file format issue was a good example of this from the computing service point of view. The need to get together again and exchange views was agreed as were a number of other recommendations including: the need for good practice guidelines for "amateurs"; information on who could offer services on what kit; evaluation of authoring systems. There is a summary in this Newsletter and the workshop report is available as AGOCG Technical Report 18.
One of the things we really did decide was how to spell the m word. We also decided it was an adjective, i.e. multimedia WHAT? This was thanks to Wendy Hall from Southampton who spoke about the Microcosm project. John Dyer from the JNT described the SuperJANET pilot projects which will be looking to use multimedia applications such as video conferencing across the network. John stressed the need to think distributed and to think about standards. Standards came up again as we looked at video technology and compression issues. It was clear that this is a very wide area of discussion and that there is a need for good advice and for common directions. AGOCG are going to try to get together some papers addressing the issues and are looking to hold an event in the summer to bring people together again. The workshop is written up in AGOCG Technical Report 19.
There were 53 people at Loughborough on 14/15 December getting away from it all (ie phone and E-mail) at Burleigh Court A variety of topics were addressed: update on CHEST deals; X technical issues; Graphic Design; time management. Staff training is really important and this event proved that people needed to get away from the desk to look at new technologies and to exchange views with other support staff. We hope to run this one next year.
This course on Graphics and Visualization being held at the University of Leeds in early January has been heavily oversubscribed. We will be considering whether this course should be run again. If you are interested in ensuring that you receive details on any future events then please send me a postal address and/or E-mail (preferably both!).
It seems to me that many of the things which were once guarded by each of us and considered our own domain have got less clear over the last few years. Suddenly though I have found myself needing to know about lots of network tools - what are these things called Archie, Gopher, WAIS, Z39.50? Are they acronyms I should drop in conversation at parties? The worst of it is that it seems that I really ought to know about them and I have not been sure where to find an overview.
At the recent Supporters Course for Graphics Support Staff, Bob McGonigle from the University of Edinburgh presented an overview of some of the books he had found useful on X and also brought along a new book on the Internet.
This book was The Whole Internet by Kroll, published by O'Reilly and marketed by Addison Wesley. This looks to be a really useful overview. I have done a deal with Addison Wesley and have some copies available. Contact me for details of getting a complimentary copy for your site. The book covers a range of topics including the tools noted above and also: what the Internet is and how it works; basic Internet utilities; electronic mail and USENET; what resources are available.
The ISBN is 1-56592-025-2 and the book costs £18.95
It is very gratifying, having reported progress on the following items in several previous issues of this Newsletter, to be able to say that the deals for image processing software for UNIX workstations, and a PC drawing package are now available.
The following information gives a little background to the deals. I should like to acknowledge the significant input provided by Fred Hopper (NERC) to the image processing part and by David Le Quesne (University of Central Lancashire) and Dave Cunningham (Bath University) to the PC drawing part of this article.
The recommendation of the evaluation team was for two packages, one to cover remote sensing applications and one to cover non-remote sensing applications (medical, vision, astronomy, microscopy, radar). The two packages chosen were: ERDAS from Remote Sensing Services and VISILOG from Data Cell Ltd.
These recommendations are subject to final funding approval by the ISC and subsequent finalization of contracts.
ERDAS is the market leading remote sensing package developed in the USA and used by both industry and academia. Importantly the software provides comprehensive facilities to read satellite (and radar) data formats and to register imagery maps of various projection. The software also provides for:- Classification, Geometric correction, Radiometric correction, Image mathematics (Principal components analysis), Comparison, Area extraction, Feature extraction, Regions of interest, User co-ordinate systems, Spectral processing and Digital terrain modelling. The package also has close links with Arc/info and provision for moving data in and out of Arc/Info format. Of particular attraction is the look and feel of the GUI, windowed man machine interface (MMI) provided for the display and manipulation of imagery on a workstation and the tools provided to generate high quality, annotated hard copy output.
VISILOG is also a well-known, market leading package for handling microscopy and medical imagery data. Data input in this environment is frequently real time, e.g. TV imagery, and if facilities for data capture are not provided with a particular workstation, users will need to acquire interface boards to attach cameras etc. VISILOG have expertise in this area and are able to provide such boards at very competitive rates. Once acquired, the VISILOG software provides a rich suite of tools for analysis of this imagery, including: Input control (control of an automatic microscope stage etc), illumination correction, geometry correction, Composite frame integration and registration, Density slice, Thresholding, Texture analysis, Identification and measurement, User-defined measurement parameters and scale factors, Erode, Dilate, Open, Close, Skeletonization etc. The software also allows for the 3D reconstruction of imagery taken in slices through a body, however until the Silicon Graphics (SG) Open GL library is supported by other vendors an SG machine is required to DISPLAY the reconstructed body. Again the package has a superb Gill windowed MMI.
In short both packages satisfied the selection criteria very well in most areas. They both run on the main UNIX workstation platforms and support the de facto standards used in the CHEST community. As a further bonus both vendors have versions of their software which will run on a DOS PC (386/486) with VGA. Licences for this are included in the CHEST deal but users may have to purchase a dongle for PC implementations. Obviously a PC environment cannot provide the facilities or power available on a UNIX workstation but, especially in the Microscopy area, a PC is perhaps the best platform on which to do the data acquisition.
Apart from the rich set of processing techniques, both packages also provide interpretive language and programming interfaces to enable users to add facilities and techniques to the basic software. These features will make them extremely useful in both teaching and research environments. It should also be pointed out that both packages will come with context sensitive on-line help modules and comprehensive documentation.
It should be pointed out that whilst the evaluation of these packages was being carried out both companies were actively updating their products. These developments will be of great benefit to the community but have the effect that the original target platform (UNIX workstation with minimum 12Mb memory) as specified in the original RFI, has become a little underpowered. Minimum required memory sizes were typically increased to 24Mb with 32Mb recommended. This, apparently, is the price to be paid for the modem graphical user interface and is an effect observed recently for several other large software packages.
The full details of the CHEST deal are not available at the time of writing but will consist of the usual five-year contract with an annual licence fee covering all platforms on a site-licence basis. Documentation will, as for other major CHEST software deals, be available at low cost from Manchester Computer Centre. All this should mean that image processing facilities will become available which have hitherto been out of reach for many research groups and should also mean that teaching of image processing techniques using powerful and state of the art software will become possible.
A report containing details of how the evaluation of the above products was carried out, notes on all the packages considered and also 'positioning' information on related software packages such as AVS, GRASS and PV-WAVE is available as an AGOCG/IGWP Technical Report 17 from Anne Mumford.
Details of the deal for Micrografx Designer have been circulated to all CHEST contacts. The deal provides for site licences at a very reasonable cost. Since the gap between the end of the evaluation exercise and the finalising of the deal has been rather long (approximately one year) it is felt that the following information may help to clarify some of the issues.
The Designer deal was struck as a result of an evaluation which was completed in late 1991 with a report compiled by February (1992). The evaluation concentrated on Corel Draw and Designer since a survey carried out over the community clearly showed that these were the packages already in use. No other package had any significant mention. Both Corel Draw and Designer had excellent facilities in their area. The packages were the market leaders and therefore restricting the evaluation appeared to be well-justified. The evaluation carried out did not ignore other products but concentrated on comparing Corel Draw and Designer. The conclusion was that both packages were excellent and had different pluses and minuses. The recommendation was therefore that CHEST should talk to the suppliers to see what deal or deals could be struck.
The contractual negotiations were delayed for several reasons and eventually, after much effort by CHEST, Corel Draw was withdrawn. This meant that the deal finally become available approximately a year after the evaluation finished. However the evaluation team still feel that Designer is an excellent product and, given that no deal could be obtained for Corel Draw, still think that it is the best choice for the community. It was by no means clear which of the two packages was the better since they both had their good and bad points.
In general Corel Draw was better suited to illustration and had better 3D capability and better font support. It did not allow editing in preview mode which meant that editing had to be done in monochrome and then previewed separately. Designer was easier to learn, and had better facilities for technical drawing (scale drawings and layers) and a good on-line Help system. [please remember that this was the assessment in late 1991, Corel Draw 3 now has on-line Help]. Designer comes with Adobe Type Manager and Adobe Type Align which are useful in their own right (with any Windows package).
Not another acronym - I hear you say! Or is it something to do with the use of graphics for logical design? (NOR-GRAF). Not so. Just a forum for meeting those with common interests in graphics, CAD, visualization, and related matters - in the North! North is a vague and relative term - some people in the south think it means north of Watford. For us it means the Yorkshire area and beyond!
NORGRAF was set up in December 1989 with an Inaugural Meeting at the University of Leeds. It arose originally from a question that was raised at a Yorkshire Computer Centres Directors' Meeting - itself an informal forum for discussing matters of mutual interest which has been in existence for many years. The matter raised was the take-up or non take-up of the CHEST deal for UNIRAS and the implications for graphics support at the Yorkshire sites. This matter was referred to the graphics experts for further discussion.
Rob Fletcher and Rae Earnshaw, the two founders of NORGRAF, felt at the time that there would be considerable benefit in setting up occasional meetings to discuss matters such as this. We also felt it would be very useful to extend the agenda to include other items of mutual interest, e.g. the plans each site had for 'going distributed', and also to widen the attendance to include the whole of me North of England, and anyone who had an interest in attending.
NORGRAF has met twice a year since 1989, once just before Christmas and once in July. Although people say there is never really any free lunch even if it's paid for by someone else we are very grateful for the hospitality of the Computer Centres who have hosted us over the past three years. Typically 15-20 people attend each meeting. The Christmas lunches are particularly appreciated!
An early feature of the meetings was the mid-morning coffee followed by a seminar on some aspect of current interest, usually provided by an expert from the host site. This lasted some 45 minutes, before we began the Agenda for the meeting.
Here is a summary of the presentations we have received to date:
The Agendas for the meetings include topics such as UNIRAS matters, general graphics applications software, PC graphics, graphics hardware, graphics standards, distributed graphics, super graphics, (i.e. super graphics workstations!), visualization, image processing, multimedia, training materials and documentation. Minutes of the meetings are produced and are available from the undersigned on E-mail or paper.
Much of interest usually comes to the surface during the meeting. In particular, we learn of initiatives and developments at other sites in various areas - all of which can be useful. We learn of software, documentation, training materials, and E-mail lists that we did not know about before - which we can take back to our own institutions.
In an age where graphics permeates every aspect of the desktop, and almost every aspect of applications software - from visualization to GIS, and from local PC to X-Windowed terminal - we have come to the conclusion that no single person can be an expert at everything! Thus having a forum where expertise and experience can be shared has proved very useful to all participants.
It all goes to show that a good meeting with lots of items of interest can be as useful as getting 100 messages by E-mail, many of which may not be relevant and take more time to process than they're worth. A good question for NORGRAF to consider is how to make E-mail more useful!
Perhaps other areas of the country might be interested to consider the NORGRAF model, with a view to setting up their own forum of interest and activity? Or perhaps you have already done so, and we don't yet know about it? If so, please send an article to the Editor summarising what you have done! If you would like to join the group please send your name, address and E-mail to the undersigned. You would be most welcome.
I am based at The University of Otago in Dunedin but have lectured and/or given workshops on 3-D visualization at five of the other universities and also at Wellington Poly during a tour of New Zealand Universities have good students and are very well equipped, with courses of high quality and Kiwis are quick to accept, and adapt to new ideas. Many of the suggestions that I made during my previous visit to Otago have been adopted and have been expanded enormously, so it is an ego trip for me. 3-D visualization is now part of the computer graphics syllabus; the Otago Museum now has a separate Handson-Science section Discovery World; and the programming contest now has 54 entries with another success at the World finals in the USA.
New Zealand has a small population, three and a half million, but there is high awareness and usage of technology even in small companies. For example Wellington Council has made E-mail available for everyone in its area, and the banking system is a model, providing facilities which could well be copied in the UK. There are a vast number of computing companies in New Zealand but I will concentrate on one nearby company, Animation Research Ltd (ARL).
ARL is a small innovative graphics company formed at Otago, when Geoff Wyvill gathered together a group of highly talented, committed and creative programmers. Many of the readers will know his work in computer graphics and animation so will not be surprised at the technical competence of the company. (I will bring back a video of some of their work.) It was set up as a joint venture between Television New Zealand and The University of Otago in 1989 and, in 1991, was formed into a separate company, to provide computer solutions for the broadcast media, in conjunction with an existing company, Taylormade Productions. The initial emphasis was placed on providing high quality, animated, 3-D computer imagery for station and programme IDs, and has developed many programs and animations including commercials such as a superb United Airlines advertisement (wait and see the video).
Their in-house hardware includes: Silicon Graphics Iris; DECStation 5000/200; DECStation 3100; and HP 9000-720 workstation. In addition the unit is attached to a full TV production company with access to C-Format 1", Sony Betacam SP and Panasonic MII VTRs.
The connection with TVNZ resulted in the unit being awarded the contract to produce the high profile imagery for the New Zealand Americas Cup Challenge. As well as the opening 3D sequence they also produced a 3-D, real-time, animated graphics package which was used in live television coverage of the Americas Cup. This has been their most significant realtime project undertaken to date so I will concentrate on this.
The System, known as Winged Keel, delivered a broadcaster's rather than a computer programmer's solution and could be driven in two modes, simulation or telemetry.
In simulation mode commentators and/or yachting experts control 3-D models of the two yachts on the actual course for the day. They were able to simulate the behaviour of the yachts in the actual wind conditions at the time, and used the system to demonstrate rules and to create what if scenarios.
In the telemetry mode, data from the yachts was fed back to the computer base and a 3-D real-time graphic display showed what was happening on the water, 7 to 10 seconds after the actual event on the water.
The interface was designed to look and perform as a vision mix panel to:
During the development of "Winged Keel", two ARL programmers worked at the Silicon Graphics HQ in Mountain View, California for two months, and stayed, supporting the package, in San Diego for the three months of racing.
If you want further information contact ARL: their address is PO Box 5580 Dunedin, New Zealand; E-mail to Paul Sharp, boskie@otago.ac.nz
If you are thinking of coming to NZ, perhaps I should make one or two points.
Even at the current exchange rate, NZ is cheap but salaries are low by UK standards: NZ$30,000 (£10,000) is a good starting salary for a graduate; Lecturers are paid between NZ$37,000 and 50,000, but this is commensurate with cost of living - although it does not help with foreign subscriptions!
New Zealanders are said to be very laid back and unfashionable, particularly in women's dress styles (maybe due to the occasional sudden cold wind), but the upside is an honesty which is rare to find these days. Everyone assumes that you are honest, so Motels, for example, though cheap, (£25 for lounge kitchen 2 bedrooms etc, with food in the fridge) are extremely well equipped. No one steals the gear, and you present your credit card when you leave. Unfortunately the high unemployment rate (14%) is beginning to raise the crime rate too.
The country is breathtakingly beautiful with an interesting history - spectacular mountains, lakes, rivers, and ocean views - though, like USA, no pretty English villages. Walking, skiing, sailing and of course rugby are everyday sports. The public transport system is not good so many people walk to work, in Dunedin for me, that means 20 minutes down and 50 minutes back. I seldom walk back up! Most of the main roads are sealed, but you can find yourself driving along a dirt road with a long drop to one side and no protective barriers (hence their poor driving record) - good views though. We took 1300 photographs in our last seven month stay and nearly as many this time.
Many people attended this Symposium - to hear the latest plans and developments in interfaces. All those with interests in any or all of the following areas should find something exciting and relevant for the future computer graphics, HCI, information systems, multimedia and learning systems.
Around 800 people attended this symposium, including over 500 industrialists comprising of presidents and research directors who were invited to attend. The topic under discussion was future user interfaces in the form of user interface agents, semi-autonomous computer programs that mediate between the user and the computer and which are capable in some sense of adapting to the user's needs, preferences, idiosyncrasies, or as Prof. Underpants, Director of the Media Lab, put it some 20 years earlier - The personalized computer should be like a well-trained, long-standing English butler - someone intimately aware of your idiosyncrasies, your habits, your friends, your goals, and who you deal with.
A spell checker transformed Negroponte into Underpants without the proof-reader noticing ;-)
The proceedings consisted of 11 presentations by pioneers and leaders in the fields of computers and human computer interaction, and were fittingly introduced and chaired by Jeeves, the classic English butler in white tuxedo, black dinner jacket and tails, and impeccable English accent.
Apple Fellow Alan Kay, designer of the Macintosh style of user interface in the 1970's, began by outlining some of the key developments in the areas of computing and graphics from 194592, including Sketchpad by Ivan Sutherland (1965), the Virtual Reality Helmet also by Ivan Sutherland (1968), Dynabook in 1968, and the Architecture Machine by Nick Negroponte (1970) - an early example of 'ubiquitous computing' in the sense of matching what humans are good at to what computers are good at (human computer symbiosis!). Significant major trends include Main Frame to Desk Top to Anywhere, which coincides with corresponding shifts in control from Procedural (and Data Structures) to Object-Oriented to Agent-Oriented. Recent trends were evidenced by pronounced shifts in the market place, for example IBM had lost 2.8 billion dollars in the last quarter, and were followed by DEC and others. Only those companies with the right products will survive.
When considering multimedia agents for the 21st century it is essential to have a human computer symbiosis where humans interact kinaesthetically with the machine, or can build agents themselves, rather than just treating the computer as a mechanical slave and - like the Roman Empire of old - eventually come to find it has no future. Thus agents with value systems are indispensable to the network systems of the future, and it is of no real consequence whether such agents are anthropomorphic or just metaphors.
Prof Nick Negroponte, Co-Founder and Co-Director of the Media Laboratory at MIT, outlined two trends accompanying the movement of agents, a taxonomy of a world of agents, direct manipulation and agents, and direct interfaces to interface agents.
In 1994 he expected that 80% of the human computer interface would be an RF communicating device in the pocket, and the important component would be the RF. The second trend was to observe that the desk top metaphor was becoming rapidly out of date and no longer relevant. Desk tops were already full and busy people were not often at their desks anyway. The same will soon apply to computer desk tops.
There is already a world of various kinds of agents - from domestic to electronic. The central question is What lives in the network and what lives on the periphery? Currently thin wire TV could provide information by compression, transmission, and decompression. The world of agents was the opposite of this. A wide band network will carry vast amounts of information; what is required at the user end is a filter or agent to select the portions required - whether it be personalised TV, tailored newspapers, or technical data. As the amount of information grows, it becomes increasingly difficult for a user to filter this personally. We all see the tip of the ice-berg at the moment in longer and longer E-mail queues - and those who respond quickly soon get known as efficient, which in turn generates longer E-mail queues!
The central issue is one of direct manipulation versus delegation. 90% of business at MIT was currently run by E-mail. This generated vast amounts of electronic folders to contain all the information. How can a particular item be readily accessed? A simple request like Fetch the last letter from Alan Kay needs to be processed. We should be able to delegate this in a simple way to the computer. Just as a butler would know the relationship between his employer and his colleagues, and arrange things likely to be required accordingly, so an agent should do this for computer-based information. The central idea is shared knowledge between user and machine.
Prof Michael Dertouzos, Director, MIT Laboratory for Computer Science, presented Modem Interfaces for Ancient Humans and advocated systems that are easy to use and easy to learn, the so called Gentle Slope System(GSS). A recent DARPA study of Excel, Word, Filemaker and Powerpoint had showed that 70% of the commands were similar yet used a different syntax. It was difficult to migrate between systems - yet most of the functionality was duplicated. What is needed is a base system with a richer set of functions, on top of which more advanced facilities can be added if required. This would bring a reduction in user effort required, a reduction in programming time, and it would blur the current distinction between languages and operating systems.
Dr Seymour Papert, LEGO Professor of Learning Research, MIT Media Lab, outlined the relationship between thinker and thought. There are many books on How to Teach, but not so many on How to Learn. Best ways of learning concentrate on the interactive, syntonic, and paradox.
Prof Patti Maes, MIT Media Lab, presented the results of some early experiments with Learning Interface Agents in the context of Meeting Scheduling agents and Electronic Mail agents (known as a Personal Apprentice) and news retrieving agents (known as an Evolving Population). In the first category, the regularities and trends are noted by the agent in the way the user currently handles meetings and E-mail. Its competence level increases as the knowledge base is built up, until the agent can process new requests automatically or request the user for more guidance or information.
Evolving Population agents are useful for handling Usenet information, by building up a picture of what the user is really interested in, and also noting any changes of interest on the part of the user. As the user's interests evolve or change, the agent follows this change.
Dr Irene Grief, Director of Workgroup Technologies, Lotus, outlined the current trends in Computer Systems for Collaborative Working (CSCW) in agents working together. A system called 'Chronicle' will be made available in Lotus 1-2-3 for group activities such as group spreadsheets and group communications.
Dr Hirotada Veda, Deputy Director of the FRIEND21 Research Center in Japan, described the multimedia project set up by the Ministry of Trade and Industry in Japan.
Dr Mark Weiser, Head of Computer Science Lab, Xerox PARC, described his experiments with ubiquitous computing and interface agents. Ubiquitous computing was somewhat orthogonal to the current mainframe, PC, and laptop genre and started with the premise that the most powerful technologies are invisible and integrated the virtual and physical worlds. Phase 1 consisted of tabs palm-top computers for post-it notes, pads - notebook sized computers, and boards - wall displays. New technology is needed in the areas of networks and operating systems.
He outlined the work on activity-based information retrieval being done at Xerox Europark, Cambridge, UK, by William Newman and Mick Lamming - a system which would do retrieval in an analogous manner to physical retrieval, viz, model the filing assistant; use events, times, context, and names; and track things by badge and video shape. Such a system just did indexing; it was not an agent.
He proposed four myths which were current in interface thinking:
Thus the following characteristics of interfaces were proposed as suspect because they consisted of one or more of these myths.
The limitations of interface agents were that they did not go far enough, they kept the computer in the foreground, they were still within the old paradigm, and they fascinated obsessively - like all human-like machines!
Meyer Billmers, DEC and MIT AI Lab, described methods of providing personalised information from the vast databanks currently being accessed via network connections to newsgroups, bulletin boards, wire service newsfeeds, and E-mail lists. Generally the headers in all these kinds of information were structured, but that was about all. Greater structure in the body of the message is required, so that machine processing can pick up items of information from key words of interest to the user. However, something smarter than just string matching is required, since often different people use slightly different words for the same concept. He proposed that just as networking was the key to the information revolution of the 1980's, information management would be the key to the information explosion of the 1990's.
The final event of the day was a Panel consisting of all speakers and chaired by Mitchell Kapor, President of Electronic Frontier Foundation. Alan Kay pursued the concept of value systems for agents and suggested we might either be robbed by agents or be found to be using tools in a stupid way. He defined an agent as an entity that can take on some part of the goal structure and pursue this autonomously. An agent could therefore also be a human, a dog, or a computer! Anthropomorphism is a red herring. The important thing is not artificial intelligence but whether the goal has been transferred, as perceived by the originator. Pervasive networks would mean the production and transfer of large amounts of information - too much for the ordinary user to handle - autonomous agents would be required. He also maintained that if a user interface system used an agent it should make it clear to the user that it was a system and not a human, to protect the user and their expectations. He also felt that summaries of information keywords etc - should contain extra information, so that the user could retrieve additional items of interest within this 'halo' of information, in an analogous way to picking up a related book from a stack in a library. He raised the question of how one should cater for prejudices of the user against other users, institutions, or topics?
This symposium was visionary, eclectic, inspiring, and controversial. Many points of current concern with computers, their interfaces, and current failures to deal with info-glut were well articulated. As network infrastructures increase in bandwidth, these concerns will become more acute. Hopefully all the vendors present either have useful products in the pipeline, or will return to their development laboratories to remedy the deficiencies!
About 40 people attended this meeting, representing about the same number of Higher Education Institutions. Computing Services and Audio Visual Services were equally represented.
Dr. David Mack (Director AN Service, Loughborough) began the Workshop by addressing the area of "Managing AN". Changes in technology have blurred the traditional dividing lines between Audio Visual Services and Computing Services. They can no longer be easily divided into separate autonomous areas. Formerly the staff of the two departments had been speaking a different language, but it was now imperative to collaborate in areas of mutual interest. In some institutions with forward plans for IT developments, this is already being brought about by the need to implement policies in the following areas:
Murray Weston, Director of the British Universities Film and Video Council, discussed the use of Media for Higher Education and advocated better management strategies and exploitation of economies of scale by avoiding duplication. Top-slicing can yield substantial benefits. 'DIY' was a very British syndrome, especially amongst academics. What needs to be appreciated is that graphics design needs specialist training, and there are definite benefits in delegating work to central services.
Chris Osland, Rutherford Appleton Laboratory, outlined how to connect computers and video systems. There are three options:
For real-time images there are four possibilities:
It was emphasised that:
Dr Brian Negus, Loughborough University, discussed the role that Computing and Audio Visual Services should be playing in the development and support of multimedia facilities. In the past there have been isolated success stories (e.g. Doomsday project) but these had been limited either by dedicated hardware not being general available, or by being discipline-specific. With current trends in technology it is now possible to produce multimedia teaching materials and presentations which can operate on relatively low cost equipment. Thus the end-user level is fairly straightforward. What requires more thought is the complex process of producing multimedia materials and how they integrate into the overall instructional and educational processes. This includes the areas of educational design, graphics design, and computer systems design.
Some questions to consider are: What is the best way to disseminate multimedia instructional materials on campus? Copies of CD's for students? (this has implications for the provision of CD readers and security of the readers). Should the materials be held centrally and be accessed via the campus network? Most sites are bandwidth limited in this area. What kinds of information compression might help in this area?
Dr Anne Mumford, AGOCG Coordinator, analysed the issues of file formats and exchanging information between different systems. The overall requirements of users could be summarised under the following headings:
The most successful formats for interchange of information were as follows:
Problems in the latter area are often caused by limited implementations of agreed specifications. For example, graphics packages often use the full facilities of the standard whereas DTP software only supports a subset. Thus the user could end up in a situation where a file generated in a graphics package cannot be successfully imported into a DTP package. All this can be very frustrating and has caused many users to give up in disgust and revert to good old-fashioned manual cut and paste!
It was agreed there was a need for a list of conversion utilities currently available for file formats and what they would do.
The meeting then split into Discussion Groups to consider the questions raised. Of particular interest were those relating to policy and ensuring good communication, collaboration and cooperation between Computing and NY Services. Having a well-formulated IT strategy in the institution, with associated Development Plans should assist in this coordination. It was felt that institutions would need guidelines from the Information Systems Committee (ISC) on how they could best allocate their ISC monies when they were no longer earmarked for computing service provision in the University. It was noted that Computing Services tended to promote technical people. But what was also needed was much better management of IT resources in the institution and better development plans for the future to take account of advances in technology and the associated benefits to libraries, photographic units, language laboratories, printing units, design units, audio visual centres, and TV services.
Further information is contained in the workshop report - AGOCG Technical Report No 18.
This Workshop was concerned with identifying the current developments and issues in the area of graphics and multimedia. The latter is becoming an increasingly significant area due to current hardware and software developments, the interfaces from graphics to video, the developments in the area of faster networking (e.g. SuperJANET), and the computer-based teaching initiatives currently being set up to utilise this technology. This Workshop brought together people with interests and experience in these areas.
Multimedia applications are to be used as some of the projects for the evaluation of SuperJANET in 1993. The Joint Network Team has strong interests in multimedia in the context of developments in SuperJANET, for the areas of distributed functionality, interworking of standards, and the implications for network loadings. Some of the pilot applications expected on SuperJANET by March 93 are in the areas of computer visualization, supercomputer interconnection, medical scanning, and remote teaching. User applications are expected in the areas of distributed hypermedia systems, video conferencing, supercomputer support, and the use of X400 mail with video clips. The latter is likely to overload even a fast network!
Current generation hypermedia systems such as Hypercard and Toolbook tend to be closed systems with proprietary formats. Applications are limited in size because of the authoring effort required. They are difficult to maintain, update, and extend and are often limited to the original author. There is a lack of integration with other applications. Thus there is a need for more open systems, to utilise source material from sources such as CD-ROM, networks, databases. A database of links is needed. Microcosm at Southampton has been designed to fulfil these objectives.
The Microsoft products expected in 1993 are likely to have a major impact in the area of multimedia. Vendor initiatives and developments can become de facto standards if sufficient people find them easy to use and widely available and supported (e.g. just as Postscript has become the standard for document output).
HyTime has been adopted by Apple, IBM, OWL. Current users are US DoD, the Davenport Group, and the Esprit Multimedia Information Presentation System (MIPS) project. The ISC Courseware Development Working Party are exploring how such tools can enable personalised materials to be produced, i.e. customised for particular lecturers requirements and contexts.
Compression is essential because the lowest resolution (VGA) and lowest frame rate saturates an FDDI or high speed line. Compression of still images can be done by non-image compression (e.g. Unix LZW), run-length encoding (can give 20:1 reduction), JPEG, or fractal methods (using area decomposition).
For JPEG, there are lossy and loss-free methods, and software and hardware implementations, with several modes of operation:
JPEG is fast and predictable in hardware.
Compression of moving images can use H.261 teleconferencing standard, real time JPEG, MPEG, fractal methods, and DVI (Intel).
Different application areas require different compression systems - no one is best for all. Network speeds are increasing at exactly the moment that hardware compression is being realised. It is a good opportunity for effective handling real-time images. JPEG and MPEG provide symmetric, predictable results for compression.
Digital video is expected to be the way forward. Vendors have large investments in digital formats - PhotoCD (Kodak), CD-I, DVI, IPEG, MPEG, Quicktime, Video for Windows. Developments are taking place in both domestic and technical markets.
Digital video is required so that video can be stored with electronic data, transmitted with electronic data, and manipulated by the computer along with electronic data. Video is expected to be a kind of data type. There is no clear standard as yet - DVI, MPEG, MPEG 11, Quicktime, Video for Windows.
It was agreed that recommendations were needed for the following areas:
A survey of user requirements is needed for this area. There are also significant possibilities for new users/ new areas using the current tools (e.g. Papyrologists looking for image enhancement of papyrii). The question of standards is becoming important It was felt that more user-friendly systems were needed to handle emerging standards such as SGML and HyTime, if they were to be of use to ordinary users.
The major publishing houses are all looking at the potential of multimedia, but they are currently very expensive to set up (£100K - £200K cost for production of one title). Thus publishing houses needed to see a well defined and guaranteed market to go down this route.
Information and advice on how to select tools to be used was needed. It would be very useful to obtain State of the Art reports and circulate them. It would be useful to define the current market leaders on each of the platforms. The time-line of developments needs to be observed that decide when is the optimum time to invest in a product and its expected lifetime.
Sound and video need to be supported across Wide Area Networks. Some of the issues for the teaching context included:
There is a need for a State of the Art Report based on a longer Workshop - as was done for Scientific Visualization with the Scientific Visualization book (Brodlie et al) and the popular book by Earnshaw / Wiseman. A National Coordinator is required for the area.
CTISS file No 14, Oct 92, is devoted to the theme of Multimedia. A detailed report on this Workshop is available in AGOCG Technical Report No 19.
