Students are required to return from the field in October with a rough draft of their Dissertation, which has to be in final form by the beginning of the Spring term. Two typed and bound copies have to be submitted, one of which is returned to the student after graduation. It was because of the practical problems associated with the system that we first turned to the use of computers for text processing. For with the escalating cost of professional typing this was becoming an increasingly unreasonable demand to put on students whose grants did not make any allowance for costs which were often in excess of £100. At the same time the problems of managing the relationship with the typist and the organisation of corrections and the myriad other little problems all too familiar to those who have produced theses threatened to take valuable time from other aspects of their studies.
Micro users are familiar with the kind of word processing software which will give an indication of the final appearance of the printed text, with the current fashion being for the principle of ³what you see is what you get². This is achieved by software which is both an editor for the insertion, manipulation and deletion of text, and a formatter which sets out the text on the screen or page. Mainframes, on the other hand, tend to separate these into two discrete operations. This may seem inconvenient, though it does offer the great advantage of handling large documents which often leave micros struggling, as well as being able to re-format output quickly and conveniently.
The question of micros versus mainframes is one which we have examined and we did originally try using Aples micros running the UCSD p-system software which was officially supported. However, we found that the relative convenience of micros was outweighed by the problem of the scale of the Dissertation enterprise. Files had to be kept to a reasonable size for the machines to handle conveniently and the printing of drafts was a major headache with small matrix devices. Also the files had to be moved to the mainframe for final printing, which entailed learning something of the EMAS system, and so we decided that in future we would start off there in the first place. It is a move which we do not regret at all, for although we appear to have abandoned the presumed flexibility of the micro, that is not how it looks to us. In undergraduate teaching micros are still in very short supply and with only one lab of twenty machines to serve the whole university there was great difficulty in students getting access for the long sessions they needed. Public terminals, on the other hand were more readily available and accede more convenient, since, unlike micros, they do not have to be supervised.
The formatter which we originally used is one which had been available for some time called LAYOUT and is another locally produced package. This is very powerful and offers a great deal of control, but suffers from the disadvantage of being a ³procedural² system. This is the other side of the control coin, in that every format other than standard running text has to be specified with detailed instructions. For special effects this is perhaps acceptable, but with mundane things like lists or quotations it becomes tedious. It is partly for this reason that we have moved over to another formatter which works on the ³declarative² principle. This means that the system offers a large range of format options that are called up by simple directives. The trick with such systems is to provide a sufficiently wide range of options to meet most people's needs, as well as facilities to make modifications for special cases. Fortunately it looks as if our chosen system does just that.
We now use a package called SCRIBE which is marketed by Unilogic to run on VAX and DEC10 machines. Since our mainframe is neither of these it has been necessary to mount a partial implementation initially. It is not yet clear how much of the full SCRIBE will be available, but even now we have more than enough to cater for average needs. SCRIBE works with the concept of ³document types² which in turn contain component units; thus a Report will consist of Sections, Subsections, Paragraphs and Appendices, whilst a Manual can in addition have Chapters and an unnumbered Preface. There is the possibility of having a Thesis document type and it is hoped that this will soon be implemented.
One of the benefits of SCRIBE's sectioning commands is that they automatically number each Chapter, Section, Subsection etc as well as putting the headings in the appropriate place for that document type. All have tables of contents generated automatically as part of the basis process and indexes can be produced if items are nominated for this. It is also possible to construct cross references and proper footnotes should be available soon too. However, for the time being even the simple commands can be a great boon, such as having to merely indicate the beginning and end of a quotation for it to be intended and put into single spacing automatically, or to have lists numbered and laid out automatically.
SCRIBE requires a source file which is created using an editor, usually SCREED. This contains the text to be formatted, which does not have to be set out in any special way, except for the normal conventions of having spaces after punctuation and a blank line between paragraphs. Indeed SCRIBE will format the most elementary text without any directives at all. However, to obtain the benefits of SCRIBE's power, directives need to be inserted in the text (to create sections, compile lists, enter quotations etc). These all consist of the delimiter ³@² followed by a directive. One of the great benefits of the ³declarative² approach to formatters is that these directives are relatively infrequent and very simple. Once the source file has been prepared it is processed through SCRIBE, which puts the formatted output in a named file which it will create if necessary. This file can then be listed to a printer, viewed on a terminal or copied into another process etc.
The way in which students writing their Dissertations use the above facilities is as follows. First the source file has to be created, which can mean anything from an afternoon session at a terminal to a variety of sessions spread over several days, depending on how long the file is and the speed at which the student works. One of the beauties of the mainframe network is that a user can log onto their process from almost any terminal on the system and carry on where they left off last time. Approaches to the creation of the source file vary considerably, from those who write everything out in full first and merely type it into the computer, through those who have a rough draft of their file or chapter and modify it as they enter it. There are also those rare birds who feel able to compose directly onto the keyboard, though most find that this is a skill which comes only gradually.
Once the file is ready for processing it can be run through SCRIBE in a couple of minutes. Any errors such as incorrect directives etc are notified at the time, as well as being put in a separate file for future reference. If they are sufficiently serious the student can now go back to the source file and correct them before running it through SCRIBE again. The output file created by SCRIBE can now be listed to a line printer and collected, which can take from a few minutes to an hour or more, depending on the size of the file and how busy the system is.
For drafts the source file is preceded by a header file which creates output with double spacing and line lengths of seventy characters. The pages are numbered, the text is right justified and appropriate underlining etc is implemented. When this is printed on line printer paper it also gives a wide margin on each side, which allows plenty of space for entering corrections and comments. To make alterations the student goes back to the source file and finds the relevant place for the changes, which is where SCREED's ³find² facility is invaluable. Once all the modifications have been made the file is again run through SCRIBE and a new output file is generated usually overwriting the old one, which is no longer needed. The new file is again listed to a line printer for checking and submission to the supervisor. The process can continue till the file is satisfactory, though in the interests of both economy and efficiency it should not be done too many times.
Although for many purposes line printer output is adequate, for final copies something better is needed, with higher quality printing on good quality A4 paper. One of the features of SCRIBE is its ability to recognise different printing devices and produce output to match the capabilities of these machines. For this the file is processed once more through SCRIBE, but this time with a different header file which generates files for printing on a Philips GP300. This is a high quality dot matrix printer which is capable of offering a range of fonts which can be mixed within the text. Thus not only can headings be in bold or larger fonts (automatically selected by SCRIBE), but also book titles and ethnic terms can be in proper italics. The text is also proportionally spaced as well as being right justified. ERCC has recently taken delivery of its first laser printer, a Xerox X2700, which will shortly offer both higher quality printing and greater speed.
The final copies of Dissertations are bound by students using Faculty equipment, with a small charge for the plastic spiral combs and clear acetate sheets used as covers. There is now a charge of 4p per page for high quality printer output, which means that the real money cost of two bound copies of a 100 page Dissertation would be not more than £10.
The University Library is belatedly in the process of being computerised and there will shortly be a connection between the Library computer and EMAS. Apart from the eagerly awaited delights of being able to consult the catalogue and check the availability of books from within the Department, it will also be possible to copy bibliographical information. This will not only make more convenient the compilation of reading lists, but also the creation of bibliographies for theses and Dissertations. Computer searches of the Library will be possible and with the connection of the system to the wider SCOLCAP network, access to the catalogues of other libraries will be possible, including the National Library of Scotland.
We are convinced that the use of computerised text processing can have a great impact on supervision and examination at both the undergraduate and postgraduate level. Supervision has been transformed, with staff now being presented with typed and formatted drafts which are so much easier to read than the pages of handwriting we used to get, often heavily amended, with additions stapled on the back etc. It is now also possible for a supervisor to make comments on a draft and for the student to go away, make the suggested changes and have a completely revised version back on the supervisor's desk the next day. This kind of control is especially useful at the postgraduate level where the structure of chapters and even the whole thesis can be radically changed and alternative options evaluated. Although there is often not enough time for such experimentation with undergraduate Dissertations, it is nevertheless possible to change text with such facility that students not only produce better Dissertations as a result, but also learn more in the process.
However, one potential problem with computer use is the inevitable time it takes to become familiar with the system and under the pressure of a tight timetable this can be a disincentive. We have therefore decided to extend the learning process for our undergraduates by starting instruction in the Junior Honours (third) year, when there is more time. Thus all new entrants to the Honours years are given instruction on the use of text processing on EMAS And are required to submit their four class essays as formatted computer output. Another problem which remains to be solved is that of students' relative unfamiliarity with the typewriter keyboard, which can slow down beginners considerably. We have therefore taken a leaf out of the book of another Department and given our students access to a number of computer games, including Gollum, the famous dungeons and dragons game. Although this has been a great success it is only an interim arrangement and we would like to implement a proper typing tutor whilst we await the arrival of the already ³computerate² students of the future.
If the proposals of the Computer Board Report on Computer Facilities for Teaching in Universities ever come to fruition we could well see our students operating high powered workstations which will only need to call on mainframe services for such things as specialist printing devices.
In the meantime we will continue to rely on the power of EMAS and the ability of the network of terminals to cater to our students' needs.
Although I have stressed here the benefits of mainframe facilities, this does not mean that we are uninterested in the use of micros, far from it. The Department is already a heavy user of these for administration and research work, with our main emphasis on relational databases, but that is another topic.
(As an experiment this paper was written on an IBM PC using WordStar,
transferred to EMAS, sent to Kent via PSS, formatted by TROFF on a
VAX and finally printed for distribution on a Canon Laser
Printer).