Search
Welcome to 1992! This issue contains two reports on Visualization '91 held in San Diego in October, one from a USA perspective and the other from a UK perspective. Graphics around the Country concentrates on Edinburgh in this issue and returns to the theme of X, with contributions on X windows: the PC X Server Vista eXceed, X Windows Training Materials, and a report on the 1991 European X User Group meeting. We thank all contributors and invite articles from readers for future issues!
Things we once felt were only reasonable to provide as a national service (eg computer graphics of any description!) are now commonplace at all institutions. It is inevitable in IT that new technologies (both hardware and software) are only viable for national or regional provision early in their lifetime and then, increasingly quickly, become something we all think that we want (or alternatively in some cases something we decide we never want). AGOCG are looking at provision for more specialised output and input devices than many of us provide at the moment. Such devices might include video, CD-ROM, 35 mm slides, microfilm (see the articles in the October issue of the Graphics Newsletter from John Gilbert and Barry Robertson), large colour plots, and 3D input. I would be very interested to know what requirements people have for such devices, what software and hardware interfaces they would like to see, what sort of turnround is acceptable etc. Any information would be useful.
The funding and support for graphics standards from the DTI and the BSI has been changing over the last year or so with the advent of DISC (Delivering IT Standards to Customers). Money is short and the current bottom line is that customers for standards are to have their say in what standardisation efforts get supported in the UK. This means that customers have to input to a general pool of funds by joining DISC and then support participation in national and international meetings and work. The UK academic community have supported standards development and use for many years both in graphics standards and elsewhere (languages, networking, documents). We are a major UK customer. The Information Systems Committee of the UFC have recognised this and are members of DISC. Their nominees on the graphics standards committee (IST/31) are Ken Brodlie of the University of Leeds and Anne Mumford from Loughborough.
Chris Cartledge from Sheffield is the Chairman and many other community figures have nominations from elsewhere.
Funding for participation is limited and the ISC have made some limited travel funds available for participation in the graphics standards work. It is important that those people representing the community reflect the needs of the community. IUSC are going to discuss priorities at its next meeting. Anyone with any views on this subject should let me know. The choices are very real and involve decisions as to which standardisation efforts we should fund. Areas of current activity include application programmer interfaces (PHIGS PLUS, GKS review), file exchange standards (CGM Profiles, 3D metafile), the new imaging work, graphics reference model. There are also issues about which language binding work we should focus on, should we monitor some areas but not participate actively and so on. Views would be input into the priorities discussions.
Some new technical reports are being worked on at the moment. One of these reports will be on Visualization and will include the Introductory Guide to Scientific Visualization by Rae Earnshaw and Norman Wiseman which was described in the October issue of the Graphics Newsletter. This will be packaged with the report of the Visualization Workshop, an update and the report of the Software Evaluation. This will be available as Technical Report No 9. Other technical reports about to be published include: CGM Software Evaluation by Alan Francis (No 6); How to Put 3D Graphics into X from Liant Software (No 8); File Formats for Computer Graphics by Lesley Carpenter and Anne Mumford (No 10) and the PHIGS Software Evaluation (No 11). Not all of these will be available by February but if you would like to obtain a copy as soon as they are please send details of the technical reports you require giving the number of the report(s) and a postal address to: j.t.barradell@uk.ac.lut and they will be sent to you when they are available.
Technical Reports are free of charge for the academic community. I am working on producing an up-to-date sales list for AGOCG which will be widely circulated.
The colour printer evaluation is under way and will comprise a study of central service and devices suitable for lower volume. The report should be available after Easter.
We are hoping that there will be funding for a visualization support person and details will be circulated as soon as the situation becomes clear.
Following Norman Wiseman's article in the October newsletter, we will be holding a workshop on supporting computer graphics. It is hoped that the output of this workshop will form the basis of the AGOCG work programme for 1992/93.
I am surveying the video board market and will be producing a technical report on this when it is complete.
The Cricket Software has been available for some months, and now is an appropriate time to start reviewing the software, and the deal in general, in the light of experience. As part of the CHEST deal, Computer Associates, who supply the software, have agreed to meet with representatives in the community to discuss technical problems and give us an opportunity to suggest enhancements to the products. The IGWP hopes to organise input from the community in the first part of 1992.
An operational requirement for this software has now been completed and it is hoped to send out a request for information to suppliers soon. Fred Hopper would be pleased to hear from anyone with the appropriate expertise who would be willing to help with the evaluation of different software packages when they arrive.
The questionnaire, mentioned in the last Newsletter, received 49 replies, 33 from Universities, 14 from Polytechnics and one from Research Councils. (This is probably more a reflection on the e-mail lists used than the relative interest of the different sectors :-) ).
The results confirmed the team's assessment that it was sensible to restrict the short-list of packages for evaluation to no more than three well-known products.
Visualization has been identified as a key activity for the SERC Engineering Applications Support Environment (EASE) programme of work over the next few years. Within EASE, the idea of a Community Club has proved a successful mechanism for bringing together researchers with common interests - and so a Visualization Community Club has been formed.
Its first meeting was held at RAL on 11 October 1991, and attracted around 50 people (about half the initial membership of the Club). The meeting heard talks on a variety of topics, including:
There were also talks on current work at RAL, and on the requirements of NERC researchers.
There are a number of technical activities on-going at RAL in support of visualization:
The group at RAL is led by Julian Gallop.
A Steering Group has been formed to guide the Club and plan a programme of events. There is a feeling in the group that for 1992 the Club should look at three areas:
Thus two meetings are planned: one to look at the issue of output, where topics such as video and colour hardcopy will be discussed, the other to look at the different visualization systems and what researchers have achieved using them.
In addition, a workshop will be held on data - looking at data models, compression, etc. Finally it is planned to hold a joint event with the Computational Fluid Dynamics (CFD) Community Club.
If you would like to join the Visualization Club, please contact: Rajka Popovic. Informatics Department, RAL.
Richard Voss (IBM) and Bob Devaney (Boston University) are the two Keynote Speakers at the International Conference on Applications of Fractals and Chaos to be held in the Grosvenor Hotel, Buckingham Palace Road, London, on 19-20 February 1992.
Richard Voss will speak on the use of fractal techniques for studying DNA base sequences. Dr Voss has worked in the area of fractals for number of years with Dr Benoit Mandelbrot, and has published many papers in the area. He has also been a contributor at numerous ACM SIGGRAPH Tutorials on Fractal Techniques and Fractal Modelling. Richard Voss will also speak on Fractal Analysis of Early Chinese Landscape Paintings.
Prof Robert Devaney will give a presentation on Complex Dynamical Systems. He is a leading authority on chaotic aspects of systems and has published and lectured widely. He was an Invited Lecturer at a recent Workshop in Singapore on Chaos and its Applications.
Other invited presentations from European experts in the field include 2D and 3D modelling of marine organisms, predicting ship capsize, digital signal processing, image compression using fractals, using fractals in geological exploration, sociological modelling, and fractal structures in the brain. Further topics include chaos in road traffic flow, chaos in the behaviour of the stock market, and the rotation of bearing systems. Chaotic dynamics will be presented in a number of application areas.
Professor John Lansdown will give an Invited Lecture on the subject of Chaos: Reality or Mirage? to present some of the challenging aspects of modern chaos theory.
This conference therefore brings together a number of state-of-the-art contributions in the applications of fractals and chaos and the interrelationship between the two domains. The wide variety of areas covered indicates the extent to which fractal and chaotic methods are being used as real tools to advance understanding of natural and scientific phenomena.
This conference will be useful for all scientists and mathematicians who are interested to explore the potential of these new tools for their own work. A full written proceedings will be supplied to all delegates who register for the Conference.
This IEEE Conference took place in San Diego in October 1991. It is the second of what is now intended to be an annual conference and was a week long. Of the 400 people who attended a small number were from Britain and elsewhere in Europe.
It began with two days of workshops, for which participants were admitted by position paper only. Overlapping with these were tutorials for a day; these were followed by three days of conference papers, panels and case studies. In parallel with the conference itself was a small exhibition of only one and a half days. The workshop is reported on in a separate article in this issue.
Rather than give a blow-by-blow account, a number of themes are worth extracting.
As visualization is intended to assist humans to perceive data (and, by that means, lead to an understanding of the underlying phenomena), it was good to see human perception given a place in the conference, including the opening panel which was in memory of Edwin Land who died earlier this year.
Although the simulation of lighting effects is useful to help perceive the shape of objects (we perceive shape from shading), more than one non-ambient light can be a distraction and people tend to assume that light comes from above.
Orientation is important. Recognising symmetries is difficult when the axis of symmetry is not horizontal or vertical. Recognising familiar shapes is difficult when rotated to a non-orthogonal angle. Some of these factors are the result of our built-in receptors which can help our recognition process apparently without conscious thought.
Another panel, Color vs Black and White in Visualization, was unanimous in saying that colour is over-rated in this field. One problem is that it can present false boundaries and deceive the eye into seeing specious patterns. Often colour is best used when distinguishing a small number of distinct properties. This was rated the best Panel session in the conference.
There is a need to encourage visualization techniques which help the brain's receptors, not hinder them!
It is plain that new ideas for visualization techniques are not drying up. These include
There will need to be a way of allowing these and future methods to be available in some way from a common visualization software framework.
With the increased trend towards the use of distributed heterogeneous computing environments, the need to address topics such as the interoperability of software and portability increase in importance. At the root of such considerations is the data model upon which software is built. A data model for scientific visualization based on the abstract mathematical concept of fibre bundles (fields) was proposed in Haber et al. The model is an extension of previously published work by Butler et al and purports to be the basis of the underlying data model of IBM Data Explorer. It is applicable to both the representation of continuum fields and discontinuous geometries such as randomised data. The question of interoperability is also addressed by Nadeau et al in their work on image format conversion tools.
Some, indeed many, visualization methods are demanding of time and space. Work on massively parallel systems is being reported, with the Connection Machine CM2 (or CM2(0) being common. An example is rotating a volumetric dataset on a distributed memory parallel system (P.Schroeder et al). However there was a lack of reported work on distributed memory parallel systems that could be afforded by a department (for instance, at the 10 to 30 processor level).
It is clear that parallel systems shared memory and distributed memory - have their part to play in improving the performance of visualization software. Current work seems to be on improving specific algorithms on distributed memory systems; integrating such work into general purpose visualization systems would be one step further towards making a useful and powerful system.
Virtual reality did not make a big impact at this conference although the one talk on this topic was rated the best paper. This was from NASA Ames (Bryson et al) on applying a wide field of view stereoscopic system to (apparently) immerse the user in a simulated flow field. Unlike the head mounted displays that have been much talked about, this was a boom mounted unit employing what appeared to be an angle-poise method to hold the weight of the two displays, claiming to give the user a preferable working environment. The necessity for smooth changes placed enormous demands on a top of the range Silicon Graphics 4 D system and the degree of detail (number of particles modelled) was judged to be not yet of sufficient quality. This work was built on known wind-tunnel techniques, hence the use of the term Virtual Wind tunnel in the title.
There were papers and talks on remote visualization, using long distance networks to access visualization facilities.
It is clear that, in support of massive data sets and problems requiring numerically intensive computing, remote visualization will have its part to play. In the USA, serious money (not to say awesome money - some $IB per year) will go into the High Performance Computing and Communications Initiative starting from the 45MB its/second long distance lines available in the USA now and still faster lines in five year's time. The conference keynote speaker Larry Smarr talked on this topic. One of his examples was steering a remote radio telescope through a visualization system. The user would alter the parameters, responding to the data that he was viewing online. It would be an interesting task to specify the quality of service needed!
Of the general systems, apE was demonstrated by Tara Visual, but the Ohio State University development team was dispersed shortly before the conference. Ironically the team was awarded the prize for the best IEEE Visualization paper of 1990. AVS was demonstrated on several stands. Iris Explorer was demonstrated on Silicon Graphics only. Data Explorer (IBM only) was demonstrated on the IBM Power Visualization System, a powerful combination of Intel i860s, which had its effect on the hotel's air-conditioning - Jackets were being worn despite the equable temperatures outside. Of the special purpose systems, Visual 3 from MIT and Sterling Systems SSV software based on NASA Ames' FAST were shown. Also on view was Fieldview from Intelligent Light. These are suitable for some problems in CFD.
The aim also was to show work from University research groups, lowering the cost of the exhibition to such groups. Some work from Lowell was shown, but the space was not widely taken up by University groups. We understand that the same scheme will be used next year.
Both of us felt that it was a useful meeting, not only for the diversity of techniques and applications but also for the contacts.
Immediately prior to the main Visualization '91 Conference a number of workshops, each concentrating on a specific topic took place.
Attendance at one of these organised by Dave Butler (Sandia) and Chuck Hansen (LANL) to discuss the rather broad subject area of Scientific Visualization Environments was by invitation based on submitted position papers; we formed the UK contingent.
The Workshop commenced with a brief introduction, by Dave Butler, who outlined his view of what he hoped might be achieved in the 2 days available to us. This was followed by each participant (twenty of us in all) giving a brief overview of their position statement. The remainder of the workshop was devoted to in-depth discussions on three topics, each taking half a day of the programme schedule; Data Types and Interfaces, Control Structures and Architecture and Visualization Reference Model.
The session on data types and interfaces commenced with Julian Gallop presenting the work of the AGOCG workshop subsection on techniques (led by Ken Brodlie) which received favourable comments although was felt to be inadequate for defining certain data types (eg unstructured grids and masks (areas of undefined data)). Using the Brodlie model as a basis from which to work a model was developed which used the concept of data being represented as a function of several variables which could be sub-divided into a number of subcategories of a hierarchy. It was felt that where possible visualization systems should internally try to use a standardized set of generic data-types, but that from a user perspective these should not be apparent, ie they should be able to use their own favourite data formats.
Turning to control structures and architectures, again consideration for the user and their requirements figured highly. The impact of the availability of Visual programming was discussed and the possibility of this type of interface actually constraining the user to use only those tools made available to them. The need for an explicitly defined and well-documented programming interface was recognised, existing type of interfaces being considered deficient.
It was agreed that data flow consideration must be divorced from control and that the present data-flow paradigm of many existing visualization systems was too constricting, what was probably required was a more object-oriented approach. Visualization should be seen as a tool available to the application developer, we must necessarily restrain from forcing applications into a visualization environment. The problem of enabling computational steering was seen as an area which needed to be addressed.
Finally we discussed the need for a reference model for Visualization. It was agreed that it is too early to consider trying to produce a standard what was needed was a working model which could evolve. The model should allow traversal back up the pipeline (to use Haber and McNabb terminology). A proposal for the basis of a model was put forward by Polly Baker (NCSA).
The Workshop concluded with a discussion about the need for benchmarking and validation techniques. Dave Butler summarised the main points covered during the two days and, in conjunction with Chuck Hansen, has undertaken to write-up the Workshop with the aim of publishing the main conclusions/pointers for future work. Although the workshop was instructive, we were a little disappointed in the lack of progress made at furthering mutual ideas and identifying key research areas - perhaps after the AGOCG Workshop last February our expectations were too high!
The second conference devoted to exploring how visualization is being used to extract knowledge from data was held on 21-25 October 1991 in San Diego. Visualization '91 was sponsored by the IEEE Computer Society Technical Committee on Computer Graphics in cooperation with ACM/SIGGRAPH. The conference consisted of tutorials, case studies, panels, demonstrations and refereed papers.
After a day of tutorials, the conference began with a keynote address by Dr Larry Smarr, Director of the National Center for Supercomputer Applications. The title of his talk was, "Scientific Visualization from Inside the Metacomputer." He discussed the problem of how scientists are going to be able to analyse all that data which is currently being generated by the advances in scientific computing and sensing devices. He predicts a future where scientists will be able to do advanced real time visualization analysis at remote locations with the simulation and other scientific calculations being performed at supercomputer centres. This will require very sophisticated and fast networks connecting global computing resources.
The keynote panel for this year was devoted to the memory of Edwin H Land and was entitled, From Perception to Visualization. It included the panelists: Richard Mark Friedhoff, Nahum D Gershon, Vilayanur S Ramachandran and Robert L Savoy. They discussed how some of Land's early work is tied to current research in visualization and made the argument that knowledge in visual perception can lead to more effective methods for scientific visualization. In addition to this keynote panel, several other panels were scheduled during the remainder of the conference. Their titles give some indication of the diversity of the topics covered:
Rather than have a continuing vendor display area, a single block of time was set aside for demonstrations. Research organizations and commercial companies presented state-of-the-art software and hardware for visualization.
A case study is intended to deal with the interdisciplinary issues of visualization and how to progress from the research to the application. Nine sessions were scheduled. This unique part of the conference covered a variety of application areas which are indicated by the titles of these sessions:
This year a capstone speaker was included. Michael F Marmor of Stanford University School of Medicine gave a talk entitled Vision and Art. He used many and varied artistic works to convey some rather interesting concepts and messages in the perception of images. It was a very enjoyable end to the conference.
The proceedings of the conference includes the 41 refereed papers presented at the conference. These papers represent the state-of-the-art in visualization research and applications. The authors are from leading research centres, laboratories and universities. These proceedings (ISBN 0-8186-2245-8) can be obtained from: IEEE Computer Society Press.
Visualization '92 will be held on 19-23 October 1992 in Boston. The deadline for papers and proposals for participation is 31 March 1992. See you there!
Prof John Vince chaired this 25th Anniversary Conference of the BCS Computer Graphics and Displays Group. There were three speakers from the USA and seven from the UK.
Prof Bob Hopgood began the day with a presentation on Pioneering Images, a look at the early history of computer graphics. The BCS Displays Group was formed in 1967 to address the area of graphical VDUs, and was originally a sub-group of the MultiAccess Group! The first ACM SICGRAPH Newsletter appeared in 1966, the first SICGRAPH session at an ACM Conference in 1970, and the first SIGGRAPH Conference in 1974. Thus it appeared at first sight that computer graphics was only around 25 years old! However, a closer look at the history revealed that most of the key tools and techniques were in use much earlier. The following examples were outlined:
Some film clips from 1963-72 were shown to illustrate the above. Many of these topics seem to have been reinvented at subsequent dates!
Prof Jack Bresenham followed with a presentation on, From Algorithms to Applications and analysed consistency considerations that had to be taken into account when implementing basic algorithms for lines, areas, line attributes, curves, and clipping. An explicit basic reference model was required.
Prof Mike Pitteway drew attention to some of the mathematical disadvantages of using splines based on Bezier curves to define font outlines, compared to general conic arcs. The three point Bezier curve is the parabola, which cannot be used to achieve curvature continuity at the knot points where consecutive parabolas are joined together; more arcs are therefore required for an aesthetically satisfying font.
Derrick Grover (British Technology Group) gave a review of the evolution of display technology. The current problem of producing flat panel displays of good quality without pixel drop-outs is restricting the resolution and size of the display. At higher resolution the low manufacturing yield causes the displays to become prohibitively expensive. The trend to develop transistor arrays for dynamically controlling each pixel offers the option of a degree of fail-safe operation by duplicating the drives per pixel.
Prof John Woodwark (Information Geometers Ltd) examined the relationship between computer graphics and shape modelling. Early examples of interactive computer graphics in the early 1970s concentrated on the graphics rather than the underlying model, whereas engineering design was concerned with the production of shapes via NC machine tools (ie emodels but no graphics!). With the development of solid models, there was increasing interest in realistic pictures of engineering design, brought about by shading algorithms and rendering techniques. A current trend is towards physically based models for statically deformed materials (eg draped fabrics) and dynamic phenomena such as collisions and the flow of liquids. A further area is the visualization of data with more than three dimensions.
Dr Rae Earnshaw gave a review of tools and techniques for transforming numeric data into visual information. Large volumes of data are now being generated by satellites, instrument arrays, and scanners. What is needed are effective ways of analysing this data to reveal the information it contains. Supercomputers can provide the raw computation power required; indeed the 1987 Report produced on Visualization in Scientific Computing in ACM Computer Graphics was a strategy for seeking to capitalise on this infrastructure. A number of visualization software products were now available which had high functionality and data conversion utilities. These systems ranged from graphics systems to turnkey visualization systems where the user specified their data and selected the functions they required (eg UNIRAS agX, Wavefront Data Visualizer, PV Wave, Sun Vision, VoxelView and NCSA tools). At the higher level were systems where the user built their own application out of modules provided by the system (eg A VS, apE, Khoros, Explorer). In these systems the user selected functional modules on the workstation screen and connected them together by means of pipes and networks. Thus a complete customised application is built up without any need for programming. It can be modified and adapted as required.
Dr Peter Dzwig, Director, London Parallel Applications Centre, reviewed the current developments in parallel processing and the implications for display subsystems. Possibilities for new application areas were reviewed. These included combining images and databases, analysing high resolution imaging from satellites, and virtual reality. NASA have already built a Virtual Reality system which allows the user to roam around on the surface of Mars. The surface has been generated from satellite data.
Dr Adrian Thomas (University of Sussex) surveyed real-time display hardware for virtual reality and animated symbolism. Current work was reviewed in the area of display devices with image refresh fast enough for the eye to fuse the separate images and perceive a stable unflickering image. The real-time task was to generate the images fast enough to match the critical refresh rate for image fusion.
Lincoln Hu (Intelligent Light and Magic) reviewed the methods used for designing the computer generated sequences for the films, The Abyss (1989) and Terminator 2: Judgement Day. The sequences tackled new problems in human modelling and animation, film digitalization, digital compositing, and the integration of live action. Film clips illustrated the animation sequences which had been produced. At $4m for 7 minutes, they must be some of the most expensive sequences to date!
Matt Elson, formerly of Symbolics, reviewed computer animation and virtual reality in the film, Virtually Yours and illustrated the work with clips from the film.
It was a full day with much to see and think about. Graphics, interaction, and visualization are indeed the challenges for the 1990s! A good number of Graphics Pioneers were present and all are still very active. But hopefully not reinventing the wheel!
This month we return again to the subject of X Windows. There are two announcements about the availability of X Windows training material. There is also an article on the use of eXceed, a PC Server, and a synopsis of the last European X User Group meeting.
The purpose of this article is to make readers aware of what material is being prepared to help with X Windows training, so that they may plan accordingly.
UK Universities, and other centres of higher education, have been acquiring software through a central agency for a number of years. This agency is called the Combined Higher Education Software Team (CHEST). The deals struck by CHEST generally provided site licences, at very reasonable rates, together with manuals. The first, and one of the most popular, was for the UNIRAS software. This comprehensive set of software came with over 20 manuals, and computer centre staff were faced with the problem of how to train their users.
Several supporters of the software came together to build a training pack. This activity was brought together by Anne Mumford, the coordinator of the Advisory Group on Computer Graphics (AGOCG). This activity has been highly successful with some 140 sets sold to UK universities and polytechnics, and a few commercial sales.
The success of this venture was an influence on the Universities Funding Council's Information Systems Committee. This set up a 3 year venture, called the Information Technology Training Initiative (IITI), to enhance this area within higher education.
One of the projects, at the Computing Services of the University of Edinburgh (EUCS), is in the field of X Window Training. This is aimed at producing a training pack, available at nominal cost to all centres of higher education and research, which will be used for self-teaching, assisted learning or traditional lectures. The project duration is for 18 months.
Three classes of student have been identified as having distinct viewpoints on the X Window System, and there will be three parts to the training pack. There are first the users. These are the people who sit down at a display and use X to do their work. Second, there are the installers who maintain the software and the systems. Third, there are the managers who are the senior people responsible for running computer services.
The real work on this project started in October, when the project author, Jean Ritchie, took up her post. In the first month, there was much activity on defining content and wrestling with the timetable. The activity now has settled down into a regime of furious writing, evaluation and giving trial training courses.
The ITTI could not support a programming component from the limited funds available. An approach from IXI Limited, of Cambridge, has enabled EUCS to tackle that area, and the Motif project author, John Wexler, started at the beginning of November.
The above is some background to the project, but what probably is of more interest to the reader is the content of the material and when it will become available. First, from the X Windows part, it is expected that a complete training course for users will be available in April of next year. Second, from the Motif part, a complete 4 day training course will be available at the same time.
Details of precise content, cost and how to order will be made available in March and will be sent direct to computer centres.
If you are unlikely to receive the information by this route, then send your name and address to me.
Consideration is being given to running one-day workshops covering the use of the training materials. Also, full Motif and agX training is being considered. Details of this will be in the March mailing.
Recently CHEST reached agreement with Control Data on site licences for the Vista-eXceed X-server products. Vista-eXceed is a suite of software which enables an IBM-PC or compatible to run as an X-server. This effectively turns your PC into a (colour) X-terminal with access to all the facilities of X Windows systems.
The product can be run using either DOS or MS- Windows. The DOS servers will run in protected mode allowing up to 16Mb of RAM accessible to the client application. However it is also possible to run in real mode on machines with only 1Mb of memory. The Windows server operates in standard or 386 enhanced Mode.
The basic hardware requirements are IBM PC/AT or PS/2 compatible, with 286/36/486 processor. A number of TCP/IP transports are supported including PC-NFS which is available under a CHEST agreement.
I have been using both the DOS Plus and Windows version of eXceed for some time now on a 20 Mhz 386DX machine with SVGA and DGIS graphics accelerator card on a large display screen. There were some irritating problems with the original evaluation copy using PC-NFS 3.0.x, such as having your session involuntarily terminated every hour or so. The official release used in conjunction with PC-NFS 3.5 seems to have generally cleared these problems.
Most users would be well advised to decide on either DOS or Windows. I only continue to use both because I have to fulfil a support role on the software.
If your main concern is using an existing PC as a general purpose Xterminal, the DOS version is the better choice. It needs a less powerful machine and less memory, a 20Mhz 386SX machine with 2Mb memory gives a very adequate performance. One annoying aspect of the DOS versions available under the CHEST agreement is that their is no support for the more advanced graphics cards, (eg DGIS and TIGA). Most users may find the 800×600 SVGA servers quite suitable, but the accelerator cards give a superior performance, particularly if you produce a lot of graphics or use a postscript pre-viewer. They are also becoming much more reasonably priced.
Although you can make do with less, users of the Windows version would be well advised to have at least a 4Mb machine available, otherwise unless you have a very good reason for using Windows stay with the DOS version.
The Windows server operates through the standard MS-Windows drivers, this has the advantage of making a much larger selection of graphics cards available, but does very slightly slow the performance. You can use either the standard X fonts or alias one of the Windows system fonts. The system fonts give much faster scrolling, but there is a limited choice.
The server can operate in single or multiple window mode. Basically the difference is that in single window mode you are presented with the root window and, say, an xterm client. The window manager operates from the host rather like the DOS version. In multiple window mode the window manager is MS-Windows and each new client is part of the MS- Windows system, just like any other Windows application. You can carry on all the normal Windows functions such as cutting and pasting etc between X and Windows, and switching to other applications. The only constraint is that you cannot paste graphics to X. As more and more PC software is migrating to using Windows, the addition of X availability gives a very useful integrated working environment.
Here in Edinburgh, X is available on both our Unix and VMS central services, and about 25% of the access to these machines in through X. The availability of an X-server for PC's is seen as a very useful addition, and take-up of the product appears to be proceeding quite rapidly. Users intending to use eXceed who do not at present subscribe to chest-x windows are advised to do so. There has already been a considerable amount of useful information published there, especially on the associated set-up procedures when using MS- Windows. I have not mentioned network versions of the product as I have not at present had direct experience of these, however I know that DOS versions have been successfully networked.
The eXceed products do not support file transfer using tftp. However a number of utilities, referred to as Vista extend, are supplied which allow access to the basic DOS functions from the host machine. This allows you to do all sorts of DOS things such as creating new directories, renaming and deleting files etc without leaving your X window. It also allow the transference of files between DOS and the host machine without leaving the server.
The X-server uses a password mechanism to protect against unauthorised access to the PC. The password may be dispensed with but this leaves open the possibility of a third party deleting all your files when the server is up and running!
Unfortunately it doesn't stop you doing it yourself.
Availability of these utilities will depend on local System Managers. NB. Watch out for Cardinals.h on your system, it may not be available.
A Training package for using the software, based on the material produced for the X-server Workshop recently run at Edinburgh University, will be placed in the public domain. Details of this will appear on chestxwindows when it is available. The format of the material will be plain text and postscript files.
The package will be suitable for either training sessions or for individual use and will consist of some introductory OHP's (with notes) plus two workbooks (DOS and MS Windows). The workbooks will take users through the basic configuration procedure followed by a simple X session running some Clients. It will include the concepts of using the RGB and Font databases, restricting access to the server and using the extend utilities. The Windows session will also cover cutting & pasting between X and MS-Windows.
The original problems using PC-NFS have mostly been resolved with the release of version 3.5. SUN have now released their toolkit version TK2.0 to Application Builders and it is rumoured that considerable improvements can be expected in future releases of eXceed. An XGA server version should be available shortly.
As some other CHEST software, there is a discussion list on the Newcastle Mailbase system. Send mail to <chest-xwindowS@uk.ac.newcastle>
This year the second annual conference of the European X User Group (EXUG) was held at Queen's College, Cambridge. The conference followed the successful format of last year, with a set of tutorials preceding a full lecture programme, and an exhibition of X related hardware and software. The speakers consisted of some of the leading experts from the USA and throughout Europe. Response to the conference was very good with in excess of one hundred enthusiastic delegates.
The opening paper What's New: X today and tomorrow was presented by Keith Packard of the X Consortium at MIT in the USA. Keith outlined the current status of the X consortium, emphasising the international membership of the group and the commitment to open standards. He then went on to present an overview of the content of XI IR5, the latest release of the X window system. He rounded his talk off by giving an indication of the future direction of X. XIIR5 contains many improvements and new features which stimulated some interesting questions and discussion from the audience. Of particular interest to the graphics community is the inclusion of a sample implementation of PHIGS extensions to X, PEX-SI, which provides standard 3-D functionality, albeit implemented in software initially. Mike Herbert of Tektronix explained some of the Technical issues with PEX terminals in his paper while Evans and Sutherland Computer Ltd demonstrated stereoscopic viewing implemented in PEX on an ESV workstation at the exhibition.
There were many other excellent papers covering a broad spectrum of topics ranging from server internals, through characterising network performance, programming graphical user interfaces and user interface management systems, to the discussion of X training in the UK University sector and interesting X applications. It is hard to pick out the best paper from such a good conference but for me personally, A scheme mechanism for managing X-resources, by Doug Young of Silicon Graphics USA stood out both for the entertaining presentation and the content. Like most good ideas when presented it seemed so simple and obvious that you wondered why the approach had not been adopted before.
In conclusion, I thoroughly enjoyed the conference and I will look forward to next September when the third EXUG conference will be held in Heidelburg.
If you were unfortunate enough to miss the EXUG conference this year you can get a copy of the proceedings from the EXUG office.
