\centerline{\bf The uphill struggle of \TeX} \medskip\noindent The following article is a translation of a piece by Laurent Siebenmann which originated as a paper delivered to a gathering of the French \TeX\ user group GUTenberg (Groupe des Utilisateurs de \TeX) and which is to appear in the Gazette of the French Mathematical Society. I am proud to say that the author himself concurs with my translation. In fact, he could very well have written it himself in English. Even so, both of us had difficulty with a few translations. For example, what should the English title of this piece be? In French it was ``Evolution en amont de \TeX''. Should this be rendered as ``The uphill struggle of \TeX'' or more felicitously as ``The progressive evolution of \TeX''. There are also some wonderful French phrases which acquire colour by virtue of linguistic confusions to which the French themselves are probably immune. Take for example ``dactylographe chevronn\'e''. For me, this conjures up the image of a she-goat. However, nothing so ungallant is intended; and I think that the translation can only be ``seasoned typist''. The piece serves to show just how vigorously \TeX\ is progressing ``on the continent''. \smallskip \rightline{\sl Stephen Pollock} \bigskip \rm\def\tem#1{\par\noindent#1\enspace\ignorespaces} \def\bull{\par\noindent$\bullet$\ignorespaces} \centerline{\bf The evolution of the \TeX\ user interface} \medskip \leftline{\sl Preprocessors for Mathematics Texts} \noindent The task of transforming a manuscript into a printed document of high quality is particularly difficult whenever mathematics are involved. Since Donald Knuth conceived \TeX\ partly for the purpose of resolving this notorious problem --- which was directly affecting his own books --- it will probably be of interest to the participants of the Gutenberg Seminar (including the non-mathematicians) to hear that a variety of solutions are now on offer. In fact, the ideas which are being pursued have a wide range of application. However, the specific preprocessors which I am going to itemise here, and, in particular, my own Sweet-teX, do not, at present, have the applicability outside the realm of mathematics that \TeX\ has. These preprocessors are specialised tools whose uses are largely confined to the tasks which they perform to the greatest advantage. At present there are two well-established classes of preprocessors for mathematical manuscripts: \smallskip\noindent \frenchspacing {\sl Category I:} Text-Orientated Preprocessors \bull STRATEC: a preprocessor for the Victor(``Sirius'') by D. Foata, J.J. Pansoit and Y. Roy, Laboratoire de Typographie Informatis\'ee, 7, rue R. Descartes, Univ. L. Pasteur, 67084-Strasbourg, France; Tel (16)-88-41-63-45. A long-awaited version for the IBM PC has recently appeared as well as a version (unexpected) for the Macintosh. See [TSD1] \bull Easy\TeX\ for the IBM PC, by a team from Milan: E. Crisanti, Alberto Formigoni and P. La Bruna; Te.CoGraf. snc, via Pacini 11, I-20131 Milan, Italy. This preprocessor has a twin in Category II. \bull Sweet-teX is a preprocessor for the Macintosh by Laurent Siebenmann, Math\'ematique, Bat. 425, Univ. de Paris-Sud, 91405-Orsay, France; Tel (1)-69-41-79-49. Available for the classic Mac and the Mac Plus; versions planned for the Mac II and the Mac SE. French version and English version. See [Sieb] \bull Mathor, for the IBM PC, Novedit, avenue Hoggar, ZI Courtaboeuf, 91 Les Ulis, France. Tel (1)-69-07-36-88. \bull Easy\TeX\ (bis!), for the Macintosh, written by O. Pironneau and distributed by Numerica, 23 Boulevard de Brandebourg, BP 215, 94203 Ivry-sur-Seine, France. A desk accessory which works with the MacWrite word processor. \bull Grife, for the IBM AT and the Sun, by Vincent Quint and Ir\`ene Vatton, INRIA et Laboratoire de G\'enie Informatique, BP 53X, 3801-Grenoble : Tel (16)-76-51-48-59. See [TSD], [FQA]. Grife is not yet widely available; but it is the only processor which aims to be a fully-featured ``dream system''. \bull ChiWriter, for the IBM PC, by the American Cay Horstmann; available in France from Churing, 20 rue Condorcet, Paris 9. Tel (33-1)-40-15-07-47. ChiWriter is a mathematical word processor which has recently been equipped with a \TeX\ code generator. \bull Publishing companion 1.03 for WordPerfect (and others?) for the IBM PC, K-talk communications, 3920 Olentangy River Rd., Columbus, Ohio 43214, USA, tel (614)-459-9711. \bull \TeX mate for the IBM PC, by A. Lannes, Tr\'ema PAO, 18 Place Margnac, 31520 Toulouse. \smallskip \leftline{\sl Category II: Equation Editors} \noindent An equation editor offers an interactive environment for constructing mathematical formulae in isolation. The formul\ae\ created in this way are inserted subsequently into the mathematical article. (See also [Pal] and the advertisements in Notices of the AMS, Jan. 1988.) \bull Easy\TeX, for the IBM PC. See Category I. \bull MathWriter, Cooke Publications, PO Box 448, Ithaca, NY14852, USA. Tel (607)-272-2708. \bull MathType, Design Science, 6475-B E. Pacific Highway, Suite 392, Long Beach, CA9080. Tel (213)-433-0605. \bull Formulator, by Bob Pollard, Icon Technology Ltd., 9, Jarrom Street, Leicester LE2 7DH, England. \nonfrenchspacing \medskip I have listed only those preprocessors which are (or are claimed to be) in service and which, during the composition of a document, display on the screen something resembling the final printed text. The list is already considerable and, doubtless, it is incomplete. The preprocessors of Category I are not homogeneous. One can begin by distinguishing those which are word processors in their own right. This is the case for Mathor and Grife. By contrast, STRATEC Sweet-teX and easy\TeX\ (Mac) are grafted onto preexisting word processors; which serves to make them inexpensive and adaptable. (Sweet-teX works with any word processor for the Mac using standard menu resources.) Next one can distinguish the sub-category of preprocessors which exploit the concept of ``high-level formatting''. This is true of Sweet-teX and Grife. In contrast, this concept is largely neglected by Easy\TeX\ (Mac) and Mathor. In Category I, one sees a Latin dominance, which has a large French contingent, whilst in Category II one sees an Anglo Saxon dominance. This should improve international trade in both directions; for the two types are complementary and, moreover, they can collaborate easily in the same manuscript without any special programming to integrate them. Suppose that, in using the Type I preprocessor, one encounters an excessively complicated formula. One might format it with the Type II formula processor (which is often a desk accessory on the Macintosh) and proceed to transfer it into the Type I manuscript using the clipboard to give visual continuity to the document. To guarantee the correct insertion of the formula in the version of the document formatted by \TeX, one could, for example, type, at the appropriate insertion point, the command |\input Formula.tex| and then give the name |Formula.tex| to the \TeX\ code generated by the Type II preprocessor. This presupposes that the Type I preprocessor accepts \TeX\ code whenever one wishes, and that it will ignore the corresponding graphical object. I must admit that I have not tried all of these preprocessors; and therefore my intention is to make people aware of their potentialities rather than to conduct a detailed assessment of them. \medskip \leftline{\sl What one may expect of a \TeX\ Preprocessor?} \bull A device to help in learning \TeX{}. It is no longer necessary to memorise accurately a long list of \TeX\ commands. STRATEC and Easy\TeX\ (Mac) are both good tools for learning. (My own Sweet-teX is inferior in this respect because its high level formatting, which is put into effect by a system of signposts or Icons, itself demands some practice. Moreover, to enable it to collaborate with several different word processors, Sweet-\TeX\ has hitherto employed a system of invisible characters, which have proved hazardous for beginners. To overcome this problem, I have recently constructed an enhanced version which exploits the internal file structure of MacWrite). \bull A way of preparing \TeX\ files which is significantly faster. The speed advantage of preprocessors is not manifested for all typists nor for all articles; but it does seem to me that it is overwhelmingly confirmed when complicated mathematical articles are being processed by professional typists. \bull A working environment for the mathematician which, under favourable conditions, can also become an environment for drafting and even for research. This is likely to work best when the screen image is convincing and the effort involved is modest. \bull A better environment for correcting and polishing \TeX\ files. Even typists who learn to write the \TeX\ code of complicated articles with fluency tend to tear their hair when it comes to corrections; because the effort of decoding (ie. of reading) \TeX\ is an exacting task. By contrast, with a preprocessor, one can often, in making corrections, rely upon a display of the mathematical formulae which is infinitely more readable. \bull Versatility in formatting. One can hope to be able freely to change the version of \TeX\ (Plain, \AmSTeX\ , \LaTeX\dots) and even to be able to quit \TeX\ altogether to the extent that the article can be typed entirely in the language of the preprocessor. This will certainly require, for each version of \TeX\ , a specialised configuration of the transcoding mechanism. (At present, Sweet-teX is adapted to plain \TeX\ and to \AmSTeX.) \bull A more effective use of the (rare and expensive) talents of the \TeX perts. A \TeX pert printer can supervise the work of a large number of (less expensive) typists trained only in the use of the preprocessor, and he can busy himself in proofreading their work in order to make corrections and to enhance it with snippets of pure \TeX\ . The readability of the display generated by the high-level formatter is the crucial factor here. For the benefit of professional typesetters, I should emphasise that such a system allows a master printer effectively to supervise the work of his apprentices and to guarantee its quality. \bull A compatibility between \TeX\ and the conventional way of typing papers destined for journals printed in the traditional manner. Few typists can be expected to work simultaneously in the more than one style of technical publishing. It is too much to ask of them. Sweet-teX, Mathor and Grife attempt to achieve this compatibility by offering a document preparation style which is essentially the same for both \TeX\ and traditional printing. To summarise, there exist several new user interfaces for \TeX\ which, for the first time, are compatible with the traditional methods of technical typing. I predict that, in consequence, the seasoned professional typists within our scientific institutions will progressively take over an large proportion of the \TeX\ installations. As for commercial printers, it is to be expected that they will replace their current keyboarding systems by up-market preprocessors such as Grife. \medskip \noindent{\sl What the Preprocessors Can Expect from the Mathematics Community?} \bull A greater social awareness on the part of mathematicians who are experts in \TeX! Following the lead of those rare colleagues who have authored a collection of macros for \TeX\ , it behoves them to question whether, instead of coding their own manuscripts in \TeX\ , their time would be better spent in guiding typists in the use of a \TeX\ system allied to a preprocessor. My colleagues are often quite proficient as \TeX\ typists; but, in view of their superior salaries, I often ask myself whether their undeniable talents for abstraction would not be better employed in another task which is a bit more subtle and a lot more pressing: namely to acquire a knowledge of how to maintain such a system. Fortunately there exists, to support such activity, an excellent little manual --- in French --- which is about to be published [S\'er]. Also, fraternal assistance is now available via GUTenberg [GUT] which is the French \TeX\ user group. \bull A greater rigour on the part of mathematicians in the way they present their articles. Only by grasping a complete and correct grammatical structure can a typist settle questions of formatting. The advent of \TeX\ has abridged many traditional steps on the way to printed mathematics; the remaining ones require and deserve additional care and attention. \bull Training for typists involved with \TeX. Occasionally a typist who starts with a preprocessor will become, by stages, a seasoned \TeX pert typesetter. \bull A greater appreciation of good typography on the part of scientific administrators, allied with a concern for quality/price ratios. One wonders what the typography of the Bourbaki seminar volumes etc. will be like a few years hence. These are publications which are often the highest mathematical quality, but which are still, in the main, produced with typewriters. \medskip {\leftline{\sl References}\frenchspacing\parindent0pt\parskip3pt [TSD1] Proceedings of the First European Conference on \TeX\ for Scientific Documentation, Como, Italy, Addison-Wesley, 1985. [TSD2] \TeX\ for Scientific Documentation, Springer Lecture Notes in Computer Sci. 236, Ed. J. D\'esarmenien. [FQA] R. Furuta, Vincent Quint, Jacques Andr\'e, Interactively editing structured documents, Electronic Publishing 1, (April 1988), 19-44, Pergamon Press. [Pal] R. Palais, Mathematical Text Processing (last article), Notices Amer. Math. Soc. (Mar., 1988), pages 391-396. [Sieb] L. Siebenman, Towards Wider Use of \TeX\ Typesetting: A pre-\TeX\ Manifesto, Notices Amer. Math. Soc. 33 (Aug., 1986), pages 597-607. [S\'er] R. S\'eroule, Le Petit Livre de \TeX, (in preparation), Laboratoire de Typographie Informatis\'ee, 7 rue R. Descartes, Univ. L. Pasteur, 67084-Strasbourg, France. [GUT] GUTenberg: Groupe francophone des Utilisateurs de \TeX{}. The group's electronic message board has recently been inaugurated at the E.N.S., 45 rue d'Ulm, Paris V.} \smallskip \rightline{\sl Laurent Siebenmann}