% notes on 1st meeting of UK TeX Users' Group % held on 15 March 1989 at London School of Economics % % hastily put togther under conditions of pressure and duress by % David Osborne, Cripps Computing Centre, University of Nottingham % email: d.osborne@uk.ac.nott.clan % % ==use plain format== % \def\Item{\item{$\bullet$}} \def\Itemitem{\itemitem{\ --\ }} \def\\{$\backslash$\relax} \def\specials{{\tt \b special}s} % define the LaTeX logo in two parts (we want the La separately) \def\La{{\rm L\kern-.36em\raise.3ex\hbox{\sc a}}} % now some other useful odds and ends \article{\bf Notes on 1st meeting, UK \TeX\ Users Group} The London School of Economics was the venue on 15th March for the first meeting of a UK \TeX\ Users' group (numbering November's inaugural Nottingham meeting as the `zeroth'. Unlike the earlier meeting, the LSE meeting took a theme for all the talks: `Graphics and \TeX'. Because more time had been available to plan and publicise the event, it was good to see a broader representation of \TeX\ users instead of the bias towards those from academe and a rough head-count gave between 40 and 50 attendees. \section{`Business meeting' ({\it Malcolm Clark\/})} As the gathering was intended to be the first official meeting of a national UK user group, Malcolm Clark began with what he called a `business meeting' to discuss matters relating to the organisation of a user group, such as a constitution and the formation of an organising committee. Malcolm had prepared a draft constitution for the group, copies of which he distributed. This covered not just administrative matters, such as arrangements for annual meetings and the number of members at these to form a quorum, but also raised some questions which he asked those present to consider and decide at the next meeting. These covered important issues such as the relationship of a UK group to other European \TeX\ user groups and to TUG itself, together with the financial (and political) implications of membership fees. Clearly, these will require a good deal of discussion to sort out but Malcolm argued for co-operation with other groups, where possible. Communication of information between users in the group and between this and other groups is obviously an important point and some suggestions were made regarding a newsletter. Malcolm offered to make the \TeXline\ mailing list available, though \TeXline\ will continue as an independent publication. An {\it ad hoc} election of a committee to organise the group's future activities was then conducted. Finally, to lead in to the theme of the day's meeting, \TeX ed copies of a recent article by David Rogers were distributed; this had first appeared in the \TeXhax\ digest, 89(8), later printed in \TUGboat\ as `Computer Graphics and \TeX\ -- A Challenge'. \section{Picture Languages ({\it Sebastian Rahtz\/})} Sebastian had, at last, a chance to give this talk, which had to be postponed from the Nottingham meeting for lack of time. It was worth waiting for and he described the many approaches which can be taken to include graphics in \LaTeX\ documents. The hardware he used at Southampton for this was a Sun 3/50 workstation with some unspecified (large) number of megabytes of memory. He began by surveying the field to assess the range of pictures/graphics which someone might want to include in their (\La)\TeX\ documents. \Item colour photographs -- not yet possible; \Item halftones; \Item line (`art') graphics -- hand drawn; \Item computer generated, bitmaps -- `black boxes'; \Item structured computer graphics -- object-oriented; \Item algorithmic graphics -- generated from data; \Item bitmaps -- e.g., digitally transformed images; \Item character graphics \Itemitem simple {\sc ascii} characters \Itemitem Costabel's sprite macros (as described in \TeXhax 88 \#106) which take a bitmap diagram and generate a character; \Item dots -- \PiCTeX's way of generating graphics: hungry of time and memory; \Item Renner's category structure diagrams -- these use \LaTeX's picture commands and fragments of lines and arcs; \Item {\tt epic} and {\tt eepic} -- as above, but supplemented using \specials; {\tt eepic} is the extended version \Itemitem these require a \dvi\ driver which understands {\tt tpic} \specials \Itemitem {\tt tpic} is the front-end software \Item \MF -- single \MF\ characters representing diagrams -- again, a black box \Item \TeX\ halftones -- e.g., Knuth's \TUGboat\ (8,2) paper \Item graphics using \specials\ -- no standardisations yet; this introduces the problem of \PS\ versus non-\PS\ printers and what is acceptable as a lowest common denominator \Item graphics produced by post-processing the \dvi\ file -- \specials\ in the \dvi\ are treated by a \dvi\ transformer which writes a new \dvi\ file; e.g. John Renner's \TeX tyl, but this introduces problems in reliability and restrictions on text inclusion. \smallskip\noindent He suggested that it would be useful to agree that standard graphical objects, (e.g.~a circle), are available -- not everyone wants to use a picture language to obtain these. He then discussed some examples of `interactive graphics' ({\it MacDraw\/}; \PS\ `in the raw'; and more sophisticated tools such as {\it Adobe Illustrator\/}) before considering Fig, a {\it MacDraw\/}-like package on Sun workstations (also available as {\tt xfig} to users of the X-Window system). Fig generates a description of objects in the drawing in its own graphics language. This is translatable into \PS\ (richer than Fig, so presenting no translation problems); {\tt tpic} code (with some loss of Fig facilities); and into \LaTeX\ picture primitives. Sebastian showed some examples of output from each of these methods on a sample graphic and demonstrated that some produced good approximations to the desired results, while others diverged a long way from them. He drew the following conclusions: \Item graphics primitive objects (lines, rectangles, circles) can be filtered through drawing programs and included in (\La)\TeX\ documents with varying success; \Item the best results are naturally obtained when there are no constraints (from hardware, for example; graphics standardisation is doomed when we have to conform to the lowest common denominators; \Item \TeX's graphic primitives require picture drawing in ways that people don't find convenient. \section{\PS\ ({\it David Marx\/})} Dave Marx of Imperial College described the basic features and capabilities of the \PS\ page description language. He noted that for typesetting applications, it is not intended to be used directly and is consequently not as powerful in this area nor expressed at as high a level as \TeX. His talk was a useful reminder of \PS's capabilities, since most of the other speakers during the day introduced the topic of \TeX's relationship with \PS, or methods for introducing graphics described in the \PS\ language into their documents. \section{\PS\ in \TeX\ ({\it David Brightly})}% Following the \PS\ theme, David discussed \psfig\ as a means of incorporating \PS\ graphics into a (\La)\TeX\ document (see David's article on this subject in \TeXline\ 7, Feb 1988. \psfig\ was written by Trevor Darrell and described by him in `Incorporating \PS\ and Macintosh figures in \TeX', part of the \psfig\ documentation and available from good archives everywhere. David began by stating what \psfig\ is not: it is not a drawing program like {\it fig} or {\it MacDraw\/}; not a picture language like {\it pic\/}; not a \TeX\ macro package like \PiC\ or the \LaTeX\ picture environment; nor is it a \dvi\ post-processor like \TeX tyl. He described it as `electronic cut and paste for \PS\ graphics'. \psfig\ is available as a \LaTeX\ style option or a \TeX\ macro package and allows the user to scale, translate and clip an image described in \PS. The \psfig\ code consists of about 300 lines of \TeX\ and requires a \dvi\ driver which supports two simple, though \psfig-specific, \specials. These allow `inline' \PS\ and the insertion of a file of \PS\ code and can be easily modified in the \psfig\ code to work with specific \dvi\ drivers which support \specials\ of this form. David said he had successfully used \psfig\ with {\tt dvi2ps} (which one?) and with Arbortext's {\sc dvilaser/ps}. The \psfig\ macros propagate information from pre-existing \PS\ files into the \dvi\ file by means of \specials, though the \PS\ files are required to include the |%%BoundingBox| comment. This gives the co-ordinates of the bottom left and top right corners of a rectangle enclosing the graphic described by the \PS\ code and is used by \psfig\ to scale the graphic to fit the space chosen by the document author. This space need not correspond to the aspect ratio of the graphic, since \psfig\ is capable of unequal scaling in $x$ and $y$. In fact, by making the `box' into which the graphic will be included of zero size, it is possible to overlay the \PS\ with normally \TeX-typeset material. David outlined the advantages of the \psfig\ approach: it is not limited to specific types of graphic, e.g., line graphs; graphics from any package which can generate \PS\ can be included; and it is simple to use, so there is not much to go wrong. There are disadvantages: output can only be to \PS\ devices -- in the light of the \TeX-and-\PS\ coverage during the meeting, plus some comments that no-one venturing into `serious' typesetting would consider a non-\PS\ printer these days, this doesn't seem much of a restriction; and there would be typeface mismatch between annotations in the graphic and the body text of the document (assuming use of CM fonts in the body text). \section{Chemical diagrams ({\it Tony Norris\/})} The typesetting of chemical disgrams integrated with scientific text was described by Tony Norris, of the Polytechnic of the South Bank, London. The advantages of being able to accomplish this task were obvious, with `traditional' methods costing in the region of \quid40 per page and being very time-consuming. However, the nature of the diagrams introduced particular problems: they have a precise, predefined geometry; must have consistent ring-sizes between diagrams and it must be possible to tesselate different rings; correct relationships of atoms and bonds must be preserved; and the annotation of the diagrams, which makes heavy use of subscripting, should ideally be unified with the body text. One approach which had been tried and rejected was to store each ring or bond as a \MF\ character; however, this was not feasible as the diagrams might require a bond at any angle and do not allow `quantised' angles. Tony described the DrawChem program, an interactive graphical editor which had been designed for this application. It allows structured editing of a diagram, permitting duplication and modification of structures (i.e., rings), and results can be exported to PageMaker or Ventura (maybe \TeX, too). \section{Typesetting pictures in \TeX\ ({\it Adrian Clark\/})} Adrian discussed the techniques for using \TeX\ to typeset pictures which could be included in a \TeX\ document. The ability to do this brings advantages of economising on cost, improved quality of results, shortened turn-around time compared with traditional methods, and user convenience. He pointed out that traditional methods of including computer graphics as illustrations in documents introduce distortions from photographing display screens and that there is the added cost and time due to film processing. Adrian went on to describe commercial uses of halftoning. A magazine illustration, for example, would use a resolution of approximately 150\,dpi. Resolution of book illustrations is usually expressed in terms of the screen frequency used; monochrome illustrations using a screen of 600 lines/inch, and colour using 1800 lines/inch. Methods for halftoning in \TeX\ were then covered. These include \Item \specials\ and bitmaps -- these are device-specific, though efficient for processing with \TeX; however, the size of the picture must be known and specified in advance; \Item using \MF\ -- the picture can be set as a single, large glyph; this can cause problems on some printers and you might run into problems with your \dvi\ driver; \Item halftone fonts -- as described by Don Knuth in his TUGboat article; care is needed with calibration of the typesetter to obtain the best results. Adrian showed some very successful examples of images he had typeset using this last method. \smallskip\noindent Summing up, Adrian looked to future developments in this area. He confirmed that, as might be expected, typesetter output gives publication-quality results. For non-typesetter output, higher resolution devices (e.g., 1200\,dpi laser printers) will make a big difference to the results which can be achieved, although current laser output is probably adequate for internal reports. Following a question and answer session, the meeting closed with thanks being expressed to Carol Hewlett of LSE who, with Malcolm Clark, had arranged the meeting. The date of the next meeting wasn't announced, but has subsequently appeared in the bytes of {\sc uk\TeX}. It will take place on Wednesday, October 11th at Aston University, Birmingham, the theme being `Fonts -- Design and Use'. \author{David Osborne}