Jump Over Left Menu
Engineering Computing Newsletter: Issue 49
- IT Awareness - Important to many
- The Software Experience Column
- ITTI Product Reports
- Graduate Schools 1994
- Fortran - The Once and Future Language
- Community Clubs
- World Transputer Congress
- Conference and Meeting Notices
IT Awareness - Important to many
My leader article in the last ECN on IT in Engineering Awareness clearly struck a chord with many. As a result I have been inundated with views, offers of help and advice. I was very pleased with such a response and I will reply to all personally, in due course, but please bear with me as it will take some time because of the number of communications. Please rest assured all your views will be taken into account in refocusing the Initiative to make it more appropriate to engineering researchers' needs.
On the IT Awareness seminars front, you will be pleased to know that the first event is now fully planned and details are included as an insert to this newsletter. You will also see from the programme that Academician Professor Y Tzypkin from the Moscow Institute of Control will be with us to give a Russian perspective on neural computing. If you feel you can contribute to the discussion on neural networks or neuro-fuzzy learning systems, please get in touch with me, ideally sending an A4 summary of your views and suggestions for future research sub-themes.
I have just started planning the second event which will concern engineering decision support. Once again I am seeking help from any author/researcher who is or is becoming involved in any informing technology developments that really do help engineering decision making.
Such decision support tools could be for use at the design stage or help in the management of the decision making processes (simultaneous engineering or CSCW); it could relate to product development, testing or introduction; it could also concern any IT tool development which would aid engineering design creativity, manufacture or marketing. If you have any ideas, whatever the area of engineering decision support, do contact me to see if these ideas can be worked into the June meeting (20-21 June). This will be at the ICE, our venue for all London meetings. The SERC (Engineering and Physical Sciences Research Council (EPSRC) from 1 April) has listened to the summary of views I expressed on your behalf concerning the IT topics for future events and feel that the next two seminars, to be run later in the year, should be on rapid prototyping and virtual engineering and object technology. In line with its latest strategy they have also asked the coordinator to focus these future seminars towards the major themes of the innovative manufacturing initiative (more of this in the next ECN). Again, please let me know your views on the appropriateness of these topics or if you would like to be involved in these seminars, especially if you would like to run one of them with our help.
The SERC has appointed an IT Awareness/EASE Advisory Panel to ensure its IT into Engineering concerns are dealt with fairly. Members include: Professor John Perkins (representing Process Engineering); Professor Agnes Kaposi (Electro Mechanical Engineering); Professor Peter Brandon (Construction); Howard Kirby (Environmental Civil Engineering); Hilary Kahn (Engineering Design); Dr Brian Hoyle (Integrated Control Systems Engineering) and Dr Brian Murray (Marine Technology). Do feel free to contact them instead of me if you feel they would better represent your views to SERC. Finally, in recognition of the need seen by all to develop good quality software, I have been asked by SERC to chair the Advisory Panel's discussion concerning a proposal to set up a Software Quality Assurance (QA) Facility.
The need for such a Facility was originally suggested by the CFD community, but SERC believe the need may be more general. They believe such a QA Facility: could give advice on the development of high quality software according to an appropriate set of (QA) standards; should meet the general needs of the engineering research community; and therefore should monitor/assure the quality (against those agreed standards) of engineering software.
The Panel have been asked to develop a specification for such a Facility that would be put out to open tender. SERC would welcome your views about:
- the scope of such quality assurance in terms of languages supported (Fortran, C, C++ etc) and the areas of engineering to be considered (CFD alone or other areas, please specify)
- the relative merits of a central versus university based Facility.
We would also like to hear from any team who feel they might have something to offer to such Quality Assurance.
More details on this or any other matter can be obtained from the coordinator. All communications and contacts via the ITE Awareness Initiative Office.
James Powell, Brunel University (Salford University after 1 April 1994)
The Software Experience Column
Surfing the Internet
A quiet revolution has been taking place on the global Internet over the past two years and without the right sort of tools, you would have been unaware of the changes. The problem sparking the revolution has been the exponential growth both of computer hosts attached to the Internet (1.75M in August 1993) and the volume of stored information, measured in terabytes, that is freely accessible. Several new systems have emerged to index, structure and retrieve the vast quantities of information distributed on computer networks. Together, these systems support the activity of "net surfing" or, more technically, Internet Resource Discovery - which describes the process of searching through a large information space often with a poorly defined set of search criteria and goals.
The three principal systems are Gopher, WAIS (Wide-Area Information Server) and WWW (World Wide Web). They will be described here in overview along with their relationship to engineering computing. Each system is supported on Unix, IBM-PC and Macintosh platforms by public domain software which can be obtained via an Archive server. The pre-requisite for their use is that the platform has a connection to the Internet.
The service began as a campus-wide information system at the University of Minnesota in April 1991. Its model is a hierarchical menu system in which the menu items can be directories or documents. The documents can be located anywhere on the Internet and are retrieved in a transparent way. A document is taken in its widest sense because it may be not only a text or graphics file but also a query form to carry out indexed search by means of WAIS (see below) or on-line access by Telnet or a link to a Gopher server at another site. Users of Gopher access the server by means of an interface program (the Gopher client) which runs on their workstation; for example <xgopher> for X-Windows. Gopher service can be sampled from one of the public login Gophers before you install the client program. Use Telnet either to <info.sunet.se> or <consultant.micro.umn.edu> - login as gopher in either case.gopherspace
However, Gopher seems most useful for browsing at one target location. There are many hundred Gopher sites world-wide with about 80 sites in the UK (December 1993). The directory of UK gophers is held at the <ukoln.bath.ac.uk> gopher. What kind of documents will you find at a Gopher site? It will almost certainly have general, local information with a mixture of technical reports, project descriptions, searchable databases and access to other network services.
Useful Gophers for engineering applications include:
- US NIST (ISO-10303 STEP and for CAPP): <gopher-server.nist.gov>
- IEEE publications and activities: <info.ieee.org>
- Numerical Algorithms Group Bulletin Board: <nags2.nag.co.uk>
- ACM's Special Interest Group on Design Automation: <kona.ee.pitt.edu>
- University of Waterloo, Computer Aided Process Engineering: <cape. uwaterloo.ca>
Wide Area Information Server (WAIS) is a full-text information retrieval system which searches databases distributed over the Internet. Each database (source) is a collection mainly of text files accompanied by an inverted index of file contents. The database topics have very wide coverage - providing, for example, a thesaurus, bibliographies, technical reports, news archives, weather reports and so on. The search strategy has several novel aspects.
A user defines a question using natural language and specifies a set of one or more sources to be searched with the locally run WAIS client software. WAIS performs the search, hiding the network location of each source from the user. It responds with a list of names of document files each of which is ranked according to frequency of occurrence of the query terms in its contents. So, document files from different sources will be interleaved in the returned list. Any of these files may then be retrieved from their source site and viewed. The query language is deliberately simple because WAIS promotes the method of relevance feedback in order to search for similar documents. Users tend to have mixed feelings about this mechanism.
How can you discover what sources are available on the Internet? One source - directory-of-servers.src - is maintained on the machine <quake.think.com> as a master list of individual sources. Therefore a query can be formed in fairly broad terms to search this master source - perhaps using the broad term robotics returning a list of relevant sources which can then be given as the set of sources for a refined, specific query perhaps the narrow term SCARA. The latter query returned 4 sources and 15 documents when tested recently.
WAIS provides excellent support for information discovery as a one-stop search but is not suited to browsing documents at a single source. Sources relevant to engineering include comp.robotics.src, SIGHyper.src, bibszenon-inria-fr.src, comp.doc.techreports.src, comp.software-eng.src, patentsampler.src.
The rate of addition of information to the WAIS system appears to have levelled off with just over 500 sources in January 1994. WAIS can be sampled by Telnet access to <quake.think.com>, login wais.
WWW and the MOSAIC PROJECT
World Wide Web (WWW) is a client server system which presents a hypertext interface to information distributed over the Internet. It is currently receiving tremendous attention because it incorporates access to WAIS, Gopher, News and FTP services alongside its own native HTTP protocol for information transfer. The RARE report proposed that, for network access to multi-media, the European academic community should concentrate its effort on WWW.
Users browse a Web document by following links from a top-level home page to related nodes in the hypertext. The hypertext model hides detailed information behind buttons - highlighted text or icons - which, when selected, retrieve further documents. These nodes support a rich set of media including text, images, graphics, audio and video. Access is also provided to remote services by Telnet and many Web servers support indexed search of their contents. A Web document consists of literal text that is formatted with HTML - Hypertext Markup Language - a subset of SGML. So, users commonly build their own personal Web file which has pointers to relevant local and networked information ref1ecting their own perspective. There are Web client programs (hypertext readers) for a range of workstations but the best GUI interface is provided by the Mosaic reader developed by NCSA at the University of Illinois.
WWW lends itself both to information discovery and browsing. The Mosaic client supports it with excellent indexes - look out, particularly, for the Global Network Navigator. Examples of information resources with relevance to engineering research include:
- US Natnl. Institute for Standards and Technology: <http://www.nist.gov.edu>
- Cambridge University Engineering Department: <http://www.eng.cam.ac.uk>
- Electronic Engineering Group at Delft: <http://olt.et.tudelft.nl/index.html>
Growth in the number of WWW sites is rapid as the advantages of network-based information exchange are identified.
Further details of these systems and a Guide to Internet Resources for Engineering Research are on the CAD Centre Web server <A HREF="http://www.cad.strath.ac.uk">CAD Centre</A> along with a hypertext version of this article.
Iain Donaldson, University of Strathclyde
ITTI Product Reports
In 1991 the ISC (Information Systems Committee) of the UFC (Universities Funding Council) set up a project called the Information Technology Transfer Initiative (ITTI). This funded universities and colleges all over the UK to develop training materials for Information Technology. the materials were to be made available to all Higher Education Institutions in the UK for a nominal price.
EUCS (Edinburgh University Computing Service) set up two IITTI projects; one to prepare training material for the X window system, and one to develop help materials for new users of UNIX.
X window system
The X window system training packs
There are three X window system training packs, one each for users, systems administrators and managers. The material is modular, allowing courses to be tailored for different audiences. The packs consist of workbooks (which can be used as self-study material as well as for taught courses), overhead slides, and configuration files.
This is presented as three lecture workbooks, which aim to give managers an understanding of what X is for, when it would be used, and staff implications - setup and learning time, user training etc. Technical information is included but can be left out for non-technical managers.
System Administrators Course
This is an intensive and highly technical introduction to preparing equipment for use with the X window system, configuring xdm (the X display management system), font files, and security using the Magic Cookie system.
The workbooks for this course contain mainly practical exercises with comprehensive explanatory text. They cover an introduction to X, use of common MIT clients, how to use the mouse, open, iconise and close windows, the client/server model, importance of the window manager, how to configure the window manager, session scripts, resource files, remote clients and magic cookies. This material is usually presented as two half-day courses.
The courses are available to academic users from Jean Burgan or by anonymous ftp from xtraining.ed.ac.uk in the directory xtraining/user. Commercial users should contact Cliff Booth at Unipalm Ltd.
UNIXhelp for Users
Three products are being developed through this project:
- UNIX help for Users, an on-line help system for first time users of UNIX
- UNIXhelp for Users: a reference and guide, paper documentation for a similar audience
- UNIXTutor a CST course for people wanting to learn UNIX.
The on-line help is available in two formats: for use with Deskhelp for those sites which have taken up the CHEST license for Xdesktop from IXI Ltd and in HTML format for installation on any World Wide Web (WWW) server. EUCS is presently running a W3 server featuring UNIXhelp, which is available over JANET and Internet at URL: http://www.ucs.ed.ac.uk/.
Source for both versions is available by anonymous ftp from the host unixhelp.ed in the directory unixhelp. Source for the WWW version can also be retrieved using a suitable WWW client browser from URL: http://www.ucs.ed.ac.uk/Unixhelp/get_source.html.
For details of how to get the user documentation, contact Jean Burgan. If required, the FrameMaker source files can be made available. These have been built into a book with a hypertext table of contents and index, and will be of interest to sites with a license for FrameViewer. The CBT product is currently being developed using IconAuthor, and will be delivered at end second quarter this year. This product will run on PCs under Windows.
John Smith, Edinburgh University
GIS for Engineers - A need for Action?
The broad range of mapping, information handling and modelling technologies known as Geographic Information Systems (GIS) have attracted increasing attention across the engineering professions. Some applications have a CAD background, some are based within asset management, yet others playa flexible role in project planning, impact assessment and environmental modelling. From urban street furniture to evaluating the economic implications of a dam construction project, engineers are turning to GIS for increased efficiency and that all important competitive edge.
Such a bullish introduction would be misleading if it implied that GIS was now a standard component of every engineer's desktop or laptop. There is a significant gap between the technology's potential in the eyes of its advocates, and the realisation of this potential at a pragmatic level. Increased GIS awareness as a precursor to serious study of the topic is widely seen as an important and long overdue step. Recognising that the awareness deficiency was a hurdle both in the academic sector and in the professions, the Universities Funding Council sponsored a three-year programme at Southampton University's GeoData Institute to develop GIS awareness support materials. Now entering its third year, this Information Technology Training Initiative (ITTI) project has spawned a series of approaches that could serve the engineering profession effectively. The first in a series of awareness products comprises three PC based non-interactive demonstrations (running on Storyboard) accompanied by a 15-page introductory booklet. The first presentation titled "GIS Awareness" carries the text with full colour illustrations and images. The topics are covered in sufficient detail to provide an understanding of the unique qualities, issues and operations relating to GIS.
Other demonstrations cover two powerful groups of GIS functions, Modelling and Overlay Analysis and Digital Terrain Modelling. The demonstrations require no presentation software other than that provided on the discs, and they will run on any IBM compatible PC with VGA display or better, and MS DOS version 3.30 or later. The GIS Survival Guide is a 44-page booklet designed to support the next stage of GIS Awareness, or to act as a brief introduction in its own right. It has a wide audience, with particular relevance to managers involved in system selection or the GIS business case, and is arranged in nine main sections ranging from GIS - The Basics through GIS Technical Specifications and Data to Management Implications of GIS and GIS System Selection.
Julian Ball, The GeoData Institute, University of Southampton
Graduate Schools 1994
SERC runs a number of Graduate Schools each year to provide skills development and career workshops for PhD students. In addition the Schools also have places for Executives (graduates in employment, nominated by their organisation), who give the students the benefit of their additional knowledge as well as improving their own personal skills such as team building, communication and business awareness.
If you are a graduate, under 35, have been in employment for at least 2 years and are interested in attending a School please contact me for further information and an application form.
Les Sims, SERC Swindon Office
Fortran - The Once and Future Language
Fortran was the first of the true general purpose high-level. languages. It was invented by a team in IBM in the late 1950s and became so popular on IBM machines that other vendors produced variants to run their competitive products. Early in the 1960s the spread of incompatible dialects lead to the establishment of the first attempt at the standardisation of a programming language. The result was the ANSI definition of Fortran commonly known as Fortran 66.
The existence of this Fortran 66 standard led to the widespread implementation of conforming Fortran compilers and the widespread adoption of Fortran as the dominant programming language for scientific and engineering applications. In spite of the existence-of the standard, dialect differences began to develop, so ANSI decided to revise the definition of the language. This resulted in the version of the language known as Fortran 77. This definition of the language was subsequently (1980) also adopted as a formal international standard by ISO. By the early 1980s Fortran 77 was the dominant version of the language in use world-wide for technical computing.
It was obvious by this stage that major programming languages were necessarily evolving; as the technology of computers developed and the applications changed and expanded, the forms of expression used to program them had to adapt to accommodate the change. The relevant ANSI committee, X3J3, set off to produce another, the third, revision of the Fortran standard. This time working both to produce an American standard and simultaneously an International standard. Although the technical work was still being done by an ANSI committee according to the formal A NSI rules, the final arbiter of the content of the standard was the international community working formally through the international committee known by the memorable designation of ISO JTC1/SC22/WG5. The standard produced by this process was finally published by ISO in mid 1991 and was subsequently adopted by ANSI and the European equivalent, CEN, in late 1992. This latest version of Fortran, is destined to be the dominant technical programming language into the next millennium. It is informally but officially known as Fortran 90.
The making of Fortran 90
The 1966 standard was largely the codification of the limitations in compiler technology in the early 1960s. To a large extent it was the intersection of the dialect variations of the then existing Fortran implementations. The Fortran 77 committee remained very conservative. However, the language they produced was more like the union rather than the intersection of the- mid-1970s extended Fortran implementations. There were, however, a few moves in the direction of modernisation, such as block-IF and file manipulation I/O. It was clear, from many of the public comments that had been received by both the ANSI and ISO committees following the publication of Fortran 77, that there was a significant requirement for a more radical revision of the language for its 1980s version. The new, 1980s, X3J3 committee and the WG5 committee involved a much greater diversity of interests that hitherto. It is probably fair to say that from the forty or more people that were regularly involved in the process, it would have been possible to find fifty or more different philosophical views of exactly how and in what ways Fortran needed to be extended. However, it is also fair to say that there was always a substantial majority who believed that extensive modernisation of the language was essential. The predominant view was that it was neither acceptable nor adequate for the design of what became known as Fortran 8x to be based on linguistic models drawn exclusively from existing vendor extensions to Fortran 77. Clearly existing Fortran 77 extensions would be one source of ideas, but the committee actively looked at many other languages for concepts and functionalities that could usefully be adapted to the Fortran framework.
In spite of many years of aggressive antagonism from computer scientists and repeated attempts to replace Fortran with some new theoretically superior language, Algol 60, PL/l, Pascal, Algol 68, Simula, Ada, etc., Fortran had remained totally dominant in the technical programming field. However, most of these more "desirable" languages had demonstrated some useful extra functionalities and they had been a product of much increased understanding of the language designers art. For some members of the committees at least, the motivation behind much of what they were doing to modernise Fortran was to enable the general technical programmer and user to benefit from this improved language technology by designing into Fortran the most important of these functionalities. The tricky job was to do this in a way that preserved upwards compatibility with Fortran 77. This considered it essential to preserve the enormous investment in Fortran programs that already existed. Although, much less consciously a motivating factor in the early years of the process, designing a language that would be implementable efficiently on a variety of new machine architectures also became a factor.
As might be imagined, such a radical agenda resulted in technical and political differences within the committees and sometimes between the committees. Major modernisation of such a venerable and economically important facet of the computing scene as Fortran, was not and never could have been an easy process. Fortran 8x took so long to produce, involving so many painful compromises, that it finally appeared as Fortran 90, having taken 12 years.
What is Fortran 90?
It is impossible to give even an outline of Fortran 90 in an article the size of this. All that can be done is to indicate the directions in the major changes have taken the language and to suggest where more detailed information can be found. This latter information is included in a few, not entirely unbiased recommendations at the end of the article.
A major feature of the Fortran 90 standard is its establishment of a procedure whereby the language may evolve both by the addition of features and by the orderly removal of no longer needed features. To this end the Fortran 90 standard defines three categories of feature:
- Deleted Features
- These are those features of the previous language definition removed from this version.
- Obsolescent features
- These are features that were redundant in the previous language and considered to be undesirable. By identifying these features at one revision notice is given that should they become largely unused by the next revision they could be deleted.
- Extended features
- These are the features added to the language at this revision.
As far as Fortran 90 is concerned the set of deleted features is empty, all of Fortran 77 is included. However, the Fortran 90 standard does identify a small set of obsolescent features:
- Arithmetic IF
- DO loops with non-integral index
- nested DO loops sharing a single terminator
- branching to an ENDIF from outside the block IF
- PAUSE statement
- alternate return
- ASSIGN statement, Formats and GOTO
- nH edit descripters.
It is probable that some at least of these archaic features of early Fortran will be deleted from the standard definition of the language at the next revision, possibly in 1996.
The list of extended features in Fortran 90 is long. As was stressed above, Fortran 90 was a major attempt at modernising Fortran. However, the areas of major extension can be summarised as:
- Free form source code
- This is the only area of conflict with Fortran 77 so as a result the old card oriented fixed form for source is retained but the two source forms may not be mixed within a program unit.
- Whole array processing
- In this area in particular Fortran now incorporates array handling facilities that rival, or even surpass those of specialist array languages like APL. It is also in this area that Fortran's optimisability on parallel processor machines shows itself most clearly. Fortran 90 allows arrays to be manipulated as whole objects, e.g. if variables A,B,C are all arrays of the same shape (dimensionality and number of elements along each dimension) then A=A+B*C would be a valid statement and it would mean the corresponding elements of B and C being multiplied and added to the corresponding elements of A, followed by the result being assigned to the corresponding element of A. No order of calculation is implied for the different elemental operations so a processor would be at liberty to perform these in parallel if it could. Functions can deliver whole array objects as their result. Arrays can also be of dynamically determined sizes; both stack and heap storage management is covered.
- User defined or derived types (structures)
- The user is able to define new types of data as structured aggregates made up of components which are objects of intrinsic or previously defined derived types. Fortran 90 procedures can have arguments that are of derived type and can deliver results of such types.
- Global storage and program packaging
- This involves a new type of program unit, the MODULE, in which data and procedures can be defined. These data objects and definitions can then be accessed in other program units where ever they are needed. This provides global storage without the storage sequencing and separate declaration requirements of COMMON. Because derived types, data objects and procedures can all be defined together in a module extensive packages of reusable program facilities can be produced. The module is such a powerful mechanism for structuring programs it is likely that almost all Fortran 90 programs of any size will involve their use.
- Pointers and dynamic storage management
- A full pointer facility is provided which allows dynamic data structures such as lists and trees, etc., to be handled easily in Fortran. Constraints are included in the language definition so as to preserve as much of the optimisability of Fortran as possible even when pointers are used.
- User defined generic procedures
- The users are provided with the capability for extending to their procedures the generic property that was a valuable feature of the intrinsic functions of Fortran 77. Whole sets of procedures can all have the same generic name provided they can unambiguously be distinguished by the types and dimensionality of their arguments. This functionality allows the user to define new meanings for the operators, for example, it is possible for the user to define a new type, say called VECTOR, and an extended meaning can be defined for + between vectors. Extended meanings can also be defined for assignment by a similar linguistic device.
- Parameter selectable representations of intrinsic types
- The popular extension REAL*4, REAL*8 as a mechanism for selecting the machine representation for real data is extended and generalised to allow all intrinsic types to select the kind of representation by parameter. The major feature of this facility is that the analogue of the 4 or 8 may be a symbolic constant and hence may be set in one place but used all over the program. This makes changes in such things a real precision or integer range very much easier to program.
- Block control constructs
- A full set of fully blocked control constructs are added to complement the IF-THEN-ELSE-ENDIF of Fortran 77 are added. In particular, DO-ENDDO blocked loops and SELECT CASE multiway branching blocks are fully supported.
- Input/output extensions
- In many ways less has been done in the I/O area than elsewhere. A number of small tidying up additions have been made to the basic Fortran 77 I/O structure. The major extension, however, must be the ability for the first time in Fortran to process partial records, thus allowing the handling of variable length records in a more straight forward way.
There is much more in Fortran 90 than could be described here. The language is now essentially similar in overall functionality to Ada or C++. Of course there are functionalities that either of these languages handle more easily, but there are areas in which Fortran 90 is markedly more advanced than either. Fortran 90 is, however, something that none of the other languages can be. Fortran 90 is a superset of Fortran 77. This fact alone probably justifies the implication of the subtitle of this article.
Programming Language Fortran, ISO/IEC 1539: 1991
Metcalf and Reid, Fortran 90 Explained Oxford University Press (1992)
Morgan and Schonfelder, Programming in Fortran 90
Dr J L Schonfelder, Director, Computing Services Dept, University of Liverpool
Parallel Processing Evaluation (PPE) Community Club
A multi-processor PC has been added to the complement of parallel architectures and systems supported by the Parallel Evaluation Centre (PEC) at RAL. The system uses Windows NT to provide its parallel processing power.
Windows NT is Microsoft's new portable operating system. We became interested in it because of its support for multi-processor systems, which seemed to offer a route to affordable parallel processing, and because of its portability to a variety of platforms (including Intel, MIPS and Alpha). We joined the beta-test programme early on and have since worked with all versions through to an Advanced Server domain controller. NT was formally released in July 1993, and our experience to date is that it is a very stable and useful environment.
The user of an NT system need not be aware that they are using a multi-processor, as all scheduling and load balancing takes place automatically. Applications can however be written to take advantage of multiple processors if they are available.
The Win32 programming interface (an upgraded, 32-bit version of the standard Windows API) provides system calls for the creation and control of multiple threads of control within a process. The kernel provides synchronisation objects such as events, semaphores and critical regions, as well as communication objects such as pipes. This allows multi-threaded applications to be written using either shared-memory or message-passing abstractions.
Transparent processor management is a two-edged sword, of course. The programmer cannot explicitly place processes onto processors, and this can result in performance degradation. We found that a four processor machine gave speed-ups of around 2.5 to 3 on a variety of applications. Multi-threaded applications are fully integrated into the NT environment, and may partake of all system services and interaction media. Applications may use multiple threads to improve their interface's responsiveness, to perform calculations in parallel, to service multiple clients simultaneously, or a combination of all three.
NT integrates its network services into the main body of the operating system. The usual file services are available, with security being maintained across the network. A domain of NT and DOS machines may be administered centrally. Communications objects may be created and acquired by name from other computers on the network. A library provides a full set of Unix-style socket calls, which allows an NT application to interwork with other, non-NT, applications. (We have used this feature to interface an NT-hosted front end to a PVM daemon running on a SparcStation.) Interworking is further aided by support for DCE-compliant Remote Procedure Calls (RPC).
Access to this and other parallel systems within the PEC are available to all interested researchers. Our current range includes a wide range of small-scale parallel systems based on i860, C40, T800 and 486 processors.
Brian Henderson, RAL
Computational Fluid Dynamics Community Club (CFDCC)
CFDCC Winter School
In the second week of January, close on the heels of Christmas, the Third SERC School on Computational Fluid Dynamics (CFD) was held at The Cosener's House in Abingdon. The week consisted of a mixture of lectures and practical sessions, with a swollen Thames lapping at the foot of the garden and threatening to provide an all too real example of turbulent fluid dynamics in action.
Twenty four students were involved, with backgrounds ranging from studying blood flow from the heart, to designing wind turbines, to modelling ocean gyres; demonstrating the breadth of subjects to which CFD modelling is applicable. Lectures were given by Mr Steve Fiddes (Bristol), Dr Bassam Younis (City), Prof. Derek Causon (Manchester Metropolitan) and Dr Jim McGuirk (Loughborough) all of whom were happy to be cornered at coffee time (or in the pub!) with specific queries.
The week began with an introduction to the Navier-Stokes equations, an outline of their derivation and the need for simplifications if solutions to are to be found for realistic flows. This was followed by a consideration of different flow types (e.g. steady, inviscid or incompressible) and the simplifications which may be made for the different flows. The mathematical techniques for solving these equations were then introduced. This began with the application of discretisation schemes and time marching methods, initially for explicit methods but then extended to implicit methods. Finite volume methods were also considered. A discussion of the disadvantages of classical methods followed leading on to the introduction of shock capturing schemes. Pressure correction methods and the practical simulation of turbulent flows were also introduced.
The remainder of the lectures dealt with practical aspects of fluid modelling, such as grid generation and the impact of recent developments in computer hardware. Each of the lecturers also presented examples of practical applications of fluid dynamics, in situations ranging from Formula 1 racing cars to explosions on oil platforms. For someone like myself with little knowledge of CFD the course provided an ideal blend of theory and practical applications.
As the eyes of the students began to glaze over towards the middle of the week, the emphasis of the course shifted from lectures to practical sessions. This provided a chance to put the theory into practice and to obtain hands-on experience of the problems of designing and implementing computer codes for fluid dynamical modelling. (This also gave those of us with no CFD experience a chance to get our revenge on those with no previous experience of UNIX!).
The dedicated participants often continued their studies late into the evening, risking the Thames to reach Abingdon's threatened fluid research centres to conduct experiments in the different flow properties of Flowers and Speckled Hen. By Friday, these experiments had started to take their toll, (along with the endless excellent meals!) and most people where popping the peppermints and overdosing on coffee. Despite this, the majority staggered through to make their presentations on the week's practicals and to put the lecturers in the hot seat and quiz them about the course. Then, a final coffee and, laden with notes, we crept off to recover.
Dr Tricia Matthews, NUTIS, University of Reading
CFDCC Welcomes Mr G Thompson as the new Chairman
Mr Graham Thompson will be taking over the role of Chairman of the CFDCC at the next meeting. He is a partner of Binnie & Partners, which is a group of private consulting companies specialising in the water and environmental sectors. The Binnie Group has approximately 500 employees in the UK and 1500 world-wide. Graham manages the specialist support given to the worldwide group in the areas of hydraulics, hydrology, maritime and geotechnics together with related environmental areas. This includes providing extensive computing support in the application of simulations, GIS and databases.
Graham graduated from Queens University Belfast in 1973 and obtained a postgraduate degree from Manchester University in Maritime Engineering in 1975. He has been a member of various SERC committees since 1987, primarily in the Built Environment area. In 1989 he was a member of the 5-man Littlejohn Committee which reviewed the SERC Civil Engineering Research Programme and recommended a restructuring of the committees to give the researchers freer access to the decision makers. From 1991-1993 he was Chairman of the Water, Environment and Coastal Engineering Committee, which promoted and managed several large central research facilities including the Anaerobic Sludge Digestion Plant, the Flood Channel Facility and the Coastal Facility.
Graham has been a member of CFDCC since 1992. He says that during his session as Chairman he intends to further increase the responsiveness of the Club to the users and increase the influence of the Club, through increased support, from the subject committees.
Thanks to Prof G P Hammond
The Computational Fluid Dynamics community owe a considerable debt of thanks to Prof G P Hammond for his chairmanship of the CFD Community Club Steering Group. It has been through his careful leadership that the Club has successfully grown to its present membership and the range of activities it organises.
It is now four years since the first Steering Group Meeting (21 June 1990) and Prof Hammond is retiring from the chair as his role as Head of Department and other commitments make increasing demands on his time. Prof Hammond is also of the mind that change is good and that after four years perhaps the Steering Group would benefit from a change.
At the Group's recent meeting the members expressed their thanks to Prof Hammond for his hard work and he wished the Club continuing success.
New Opportunities and Directors in Aeronautical CFD
CFD has proven itself as a vital industrial tool of strategic and economic importance. This has led to the future direction of CFD research being actively discussed by both academics and industrialists. This debate is particularly important in view of the proposal for more co-funded research in the new Research Councils. As part of this debate, this one day meeting will address some of the frontiers of CFD research, with emphasis on research with eventual aerospace applications.
Invited academic speakers will discuss recent results and debate future directions for two key areas of CFD research:
- advanced turbulence and transition modelling
- unsteady compressible flows
There will also be brief presentations on the role of Engineering and Physical Sciences Research Council (EPSRC) in funding future CFD research and the recently announced national massively parallel processing system. The chairman for the day will be Steve Fiddes, University of Bristol.
The meeting will be of interest to active workers and students in CFD development and to industrial representatives, who will be shown a glimpse of where the engineering methods of the future may come from.
The seminar will be held on Wednesday 20 April 1994 at RAL. A Registration Form and Provisional Programme is included with this newsletter and more details are available from me.
Mrs Debbie Thomas, Secretary CFDCC
Visualization Community Club (VCC)
The Visualization Group at RAL runs visualization case studies. A case study is based on the work of an engineering research group with a need for visualization beyond what they already use and is in cooperation with that group. The purpose is not only to be a benefit to them, but to make available to Community Club members the case study report, any visualization software generated, and the data on which the visualization is based. A case study with the Engineering Department at the University of Oxford has been completed and a report is close to completion. This is based on data from wind tunnel testing of a 45 degree semi-angle pyramid probe. We expect a normal case study to be about I to 3 staff months of our time, if our existing funding is used.
If you are interested, then do contact me via the Community Club contact point.
The Visualization Community Club plans a series of topics for events in 1994 on the topics of:
- visualization in building research
- multimedia in engineering
- flow visualization (a course)
- visualization of experimental data
- visualization of multidimensional data
Exact titles, detailed programmes and dates will be publicised shortly. In planning these topics, we have been considerably helped by the replies from the survey of the engineering research community which we carried out last year.
New Community Club Secretary
The Secretary of the visualization Community Club, since its first meeting two and a half years ago, has been Rajka Popovic. She now has new responsibilities - still in Informatics Department at RAL - on Esprit funded work. Her replacement is Janet Haswell who is already experienced in scientific and engineering visualization at RAL. We wish both Rajka and Janet well in their new work.
Julian Gallop, RAL
World Transputer Congress
WTC '94, Villa Erba, Cernobbio, Lake Como, Italy 5 - 7 September, 1994
The WORLD TRANSPUTER CONGRESS 1994 (WTC'94) is the leading international transputer conference and exhibition and is the second in a series sponsored by and run under the overall management of TTC. SGS-Thomson, the Commission of the European Union and WoTUG are also sponsoring WTC '94.
WTC '94 will be held in conjunction with the newly formed Italian Transputer User Group (ItTUG). The Conference will incorporate the Inaugural Meeting of ItTUG. It will be the first major conference where significant applications of the new T9000 transputer and its associated technologies (e.g. packet routers) will be extensively reported.
The location is the magnificent Villa Erba Conference and Exhibition Centre, Cernobbio, Lake Como, Italy. Cernobbio is 4Km from Como.
It is set in the beautiful landscaped grounds of the Villa Erba on the shores of the lake. The Mannerist style Villa, with its steps down to the lake, was built in 1892 and is of both historic and artistic importance. Como has excellent air, rail and road access, being within easy reach of two international airports, the main motorways and the trans-European rail networks.
An associated exhibition attracting the world's leading suppliers of transputer-based and other relevant hardware, software and application products will be held at the Villa Erba Exhibition Centre. Opportunities also exist for posters and demonstrations of academic achievements.
A number of tutorials will be held on 3 and 4 September 1994 in the Villa Erba itself. These will cover the fundamental principles underlying transputer technologies, the design paradigms for exploiting them, and workshops that will focus directly on a range of specialist themes (e.g. realtime issues, formal methods, AI and image processing).
If you are interested in attending WTC '94 as a delegate or exhibitor (commercial or academic) please contact me.
Susan Hilton, RAL
CEU Support for Young Researchers to Attend WTC'94
At the very successful WTC '93 in Aachen, Germany last year, 56 young researchers from all over the European Union (EU) were funded under the Commission's Human Capital and Mobility Programme, to attend the Conference, Exhibition and two day series of Tutorials. Financial help in the form of delegate fees, travel and accommodation costs was given.
Those young researchers funded to attend WTC '93 were delighted that they had been given the opportunity to experience a major international conference, which otherwise they would not have been able to do. The general consensus was that WTC '93 was a very exciting event which all had benefited from attending.
It is expected that funds will be available again this year to enable young researchers to attend WTC'94 at the Villa Erba, Cernobbio, Lake Como, Italy.
Those eligible to apply for an award must be either i) citizens of an EU Member State or ii) persons resident, and working in research for at least one year, in a European Union Member State.
Generally, applicants should be under 35 years of age for men and 40 years of age for women. But what will be more important will be that applicants are post-graduates with their careers in the early stages of development.
For information on the Conference and Tutorial Programmes together with a CEU Award Applications Form please contact me.
The deadline for return of Application Forms is 18 April 1994. Successful applicants will be notified by 16 May and they will be required to complete and return the appropriate registration form by 13 July.
Susan Hilton, RAL
Provisional Call for Papers and Tutorial Proposals
Harrogate International Centre, Harrogate, West Yorkshire, UK 4-6 September, 1995
The Transputer Consortium (TTC) is pleased to announce that the WORLD TRANSPUTER CONGRESS 1995 (WTC '95) will be held on 4-6 September 1995 at the Harrogate International Centre, Harrogate, West Yorkshire, England. WTC'95 is the leading international transputer conference and exhibition. It is the third in a series sponsored by and run under the overall management of TTC. It is also sponsored by the Commission of the European Union.
An associated exhibition, attracting the world's leading suppliers of transputer-based and other relevant # hardware, software and application products will be held in the Exhibition Hall.
Susan Hilton, The Transputer Consortium, Rutherford Appleton Laboratory
Conference and Meeting Notices
- Neural Networks, Neuro- fuzzy and other Learning Systems for Engineering Applications and Engineering Research, 18-19 April 1994, Institute of Civil Engineers and Royal Horseguards Hotel, Whitehall Court
- 1994 UK IT Forum, Wealth Creation from Information Technology 22-23 March 1994, Heriot-Watt University, Edinburgh
- Eurographics UK'94, 22-24 March 1994 St Hugh's College, Oxford
- Object Technology, 94 28-30 March 1994, Christ Church, Oxford
- Neural Networking for Engineering Research, 18-19 April 1994, ICE, London
- Massively Parallel Computing Systems (MPCS), 2-6 May, 1994, Ischia, Italy
- Applications of Virtual Reality, Early June 1994,
- Multimedia/Hypermedia in Open Distributed Environments, 6-9 June 1994, Graz, Austria
- The 6th Transputer occam International Conference, 16-17 June 1994, Tokyo, Japan,
- Engineering decision support, 20-21 June 1994, ICE, London
- International Conference on Massively Parallel Processing, 21-23 June 1994, Delft, The Netherlands
- WTC '94, 5-7 September 1994, Lake Como, Italy
- Eurographics 1994, 12-16 September 1994, Norway
- SERC Seminar Numerical Modelling in Geotechnical Engineering, 14 April 1994, James Watt Building, Glasgow University
- Recent Advances in Image Processing, 6-10 June 1994, University of Strathclyde Glasgow