Date: Tue, 31 Jan 1995 21:15:16 +0100 Sender: NTS-L Distribution list From: "J%org Knappen, Mainz" Subject: Frequently asked questions of NTS-L, 5th edition X-To: nts-l To: Multiple recipients of list NTS-L Frequently Asked Questions of NTS-L Fifth edition Date: 31-JAN-1995 Archive name: nts-l-faq Currently maintained by: knappen@vkpmzd.kph.uni-mainz.de (J%org Knappen) Remark about the format: This faq is divided into several sections and subsections. Each section contains a subsection general with some ideas which have not yet been discussed. I added a date to some subsections to allow you to retrieve fuller discussions from the archives. The transactions of this group are archived on ftp.th-darmstadt.de [130.83.55.75] *) directory pub/tex/documentation/nts-l Each file in this directory is named yymm, where (guess :-) yy is the year and mm is the month when the mail arrived. (I.e., all postings of one month are bundled in one file.) *) Avoid using the number above ... it is subject to changes. Related stuff which is very worth reading you can find on the CTAN archives under tex/info/tex-implementors. In addition, I recommend reading the list of TeX bugs (tex-archive/systems/knuth/errata/tex82.bug). -1. Contents 0. About NTS 1. Proposed features of a New Typesetting system 1.1. Improvement of Quality 1.2. Internationality 1.3. New Look and Feel 1.4. Changing the ligaturing algorithm 1.5. Writing more than one dvi-file 1.6. Anchors 1.7. dump/restore mechanism 2. Proposed additions to TeX (concrete new primitives) 2.1. \lastmark etc. 2.2. \system 2.3. \skylineskiplimit, \skylinehorizontallimit 2.4. \directioncode 2.5. \textcode 2.6. \afterfi 2.7. \interactionmode 2.8. \mathspacing 2.9. \inputescapechar 2.10. \middle 2.11. \outputunit 2.12. \protected 2.13. \scantokens 2.14. \deferred 2.15. \everyeof 2.16. \readline 2.17. \evaluate 2.18. \ifevaluate 2.19. \everyeof 2.20. \fancyprompt 3. Metaremarks 3.1. TeX is not perfect 3.2. In which language shall NTS be written 3.3. The TeX language 3.4. The TeX engine 3.5. Extensions and Enhancements 4. Deviations 4.1. Automated Kerning 4.2. About Lout 5. Proposed changes to METAFONT 5.1. Writing to auxiliary files 6. Proposed changes to the tfm file format 6.1. Vertical kerning 6.2. Cross-font ligatures 7. Proposed changes to the dvi file format 7.1. Horizontal and Vertical Mirroring 7.2. Rotation of a glphy by 180 degrees 8. Additions to the TeX kit 8.1. dvicopy 8.2. dv2dp and dp2dv 8.3. gftotxt 8.4. qdtexvpl 8.5. fontinst 8.6. afmtotfm 8.7. dvipaste 8.8. METAPOST 8.9. Tie 8.10. dvi2dvi 9. Existing extensions to TeX 9.1. TeX-XeT 9.2. TeX--XeT 9.3. MLTeX 9.4. SiSiSi 9.5. Japanese TeX 0. About NTS (Mar 93, see also Jul 92) At DANTE '93, held at the Technical University Chemnitz last week, Joachim Lammarsch, President of Dante, announced that the NTS project which had been started under the aegis of DANTE, was to be re-formed under a new co-ordinator, Philip Taylor. The old core group, announced at the previous annual DANTE meeting, was to be dissolved, and a new core group established. Membership of the new core group will not be restricted to DANTE members, but will instead be offered to various well-known names (and some lesser-known!) in the international TeX community. see also: F. Mittelbach: E-TeX Guidelines for future TeX, TUGboat v11n3 (1990) P. Taylor: The future of TeX, EuroTeX'92 Prag (Proceedings); reprinted in TUGboat v13n4, 433 (1992) As of September 1994, there are two activities going on, the ETeX group is working on a extension of TeX and is going to produce a first ETeX soon. The NTS group takes the path to reimplement TeX in common LISP to get a deeper understanding of how the modules of TeX are cooperating which eachother. Independently, the Omega project was launched to produce a typesetting system which uses unicode as internal code and can handle arbitrarily encoded input via filters. The beforementioned projects were presented at the 1994 TUG meeting at Santa Barbara and at the EuroTeX'94 conference at Gdansk. For more information, consult the respective proceedings. For the Omega project exists a mailing list, omega@ens.fr. To subscribe, send an e-mail to LISTSERV@ens.fr. Omega is now announced to be presented at March 16, 1995 at CERN (Geneva). 1. Proposed features of a New Typesetting system 1.1. Improvement of Quality 1.1.0 General: Optimised page breaking, avoiding ``rivers'', letterspacing (see also 4.1), Hyphenation (Haupt- und Nebentrennstellen), grid typesetting 1.1.1 Skyline approach to line breaking (Mar 93) You can break paragraphs as usual with the current model, where all lines are simple rectangular boxes. If there's no necessity to insert \lineskip, then you don't have to look at the skyline. Only if two lines are too near (e.g. distance<\lineskiplimit), you have to look into the two rectangular boxes and to check if the boxes inside overlap at one or more places. For the worst case (i.e., you have to look at the skyline for all pairs of lines) processing the skyline model consumes a lot of process time, but this shouldn't hinder us to test this idea and look at the results. Btw, the skyline model seems to be easy to implement in the current TeX, because we need only some changes when the finally broken lines of the paragraph are put on the vertical list. There are more changes needed in the code, if the line break should be changed for the cases where it is possible to avoid an overlap with other break points, but IMHO it's nonetheless a relatively small change. Additionally you have to introduce some new parameters. I think of something like: \skylineskiplimit (b) minimum vertical distance between two boxes \skylinehorizontallimit (a) minimum horizontal distance line 1: ------------ % % % % ---------- ------- <== (a) ==> % % ~ % % (b) % ------- v % ---------------------- line 2: and other parameters, but the necessary parameter set, realization, etc. for "skylines" are subject of discussion. 1.2. Internationality 1.2.0 General: Typesetting in arbitrary directions, unicode support (16bits), ISO10646 support (32bits), ligatures pointing to other fonts, vertical kerning, better accent handling (\topaccent and \botaccent) Remark: UTF-8 (UNICODE) can be handled with `TeX classic', LaTeX2e support is provided by the CJK package written by Werner Lemberg. 1.2.1 Supporting TeX--XeT primitives for right to left typesetting TeX--XeT is an existing extension to TeX supporting right-to-left typesetting an producing a usual dvi-file. TeX--XeT is written by P. Breitenlohner and freely available. It is different from TeX-XeT (one hyphen only). Allthough TeX will be frozen at version $\pi$, this is not true for TeX--XeT. (see also 9.2) 1.3. New Look and Feel 1.3.0 General Windows support, wysiwyg-like features 1.3.1 Interaction with the operating system and other programmes see 2.2. \system 1.4. Changing the ligaturing algorithm Replace the (IMHO over-)optimized ligature builder/kerning routines in TeX by routines which separate between the building of ligatures and the insertion of kerns between the characters/ligatures. This can be realized in a simple way by using two passes: in the first pass the ligatures are built and in the second pass the resulting ligatures and remaining single characters are used to determine the necessary kerns. To ensure compatibility between e-TeX and TeX, it will be possible to switch between the new and the current behaviour. Additionally a flag in the TFM file of each font can be used to specify which behaviour is to be used for the font. This ensures that "old" fonts with some tricky ligature/kerning programs depending on the old behaviour can still be used with e-TeX. (I don't know if this font dependent switching is really necessary. Comments, please!!) Example: A font contains the following ligatures and kerns: o " => ligature (o") (= \"{o}) V o => kern(-smallkern) V lig(o") => nothing Input: V o " Output of current TeX: V kern(-smallkern) ligature(o") Output with change: V ligature(o") Status: Bernd Raichle has written a simple, but running reimplementation of TeX's ligature/kerning routine (in CommonLisp), which still waits to be rewritten as a TeX.web change file. The ligature builder/kerning routine is realized in one pass; kerns are introduced in a delayed manner, i.e., after we are sure that there's no possibility for a ligature. Additionally there's a switch between the current TeX and the new behaviour. (The TRIP test fails with the new behaviour.) PS: IMHO the ligature/kerning routines should be further changed to remove the `shelf{}ful' anomaly (see TeXbook, exercise 5.1), i.e., reinserting ligatures when words are hyphenated. The change should allow ligatures for inputs like `f{}f' or `f\relax f', which will simplify the macros in `german.sty', Babel and changed macros for \", \', ... which are used to select characters from DC fonts or other fonts with national characters. 1.5. Writing more than one dvi-file (Nov 94) I want to have TeX ready in the memory and avoid loading it anew and anew for each run again which costs lots of startup time, if you have to load every file via NFSS. The idea I am toying with is to have something like \level which clones the current state of TeX and allows to start a new dvi-file. Then you do anything you want (start even another \level) and at last you say \exit which performes like \end, except that it restores exactly the state before the last cloning. To communicate some information to the waiting lion, there may be one exeption. You say \exit{some_integer_number} and this number can be read by using \exitstatus. Applications are many. If you have a format which usually requires several runs, you can do all the runs within one job, and automatically decide, whether a further run is necessary. Of course, you should device your format to have a counter maxruns to avoid endless looping... 1.6. Anchors (Nov 94) Proposed by Jiri, an ``anchor point'' would be in some senses be analogous to a mark, but rather than recording textual information it would instead record the co-ordinates of itself, relative to the reference point of the smallest surrounding box. Additional new primitives would be required to return the co-ordinates of a specified anchor point. 1.7. dump/restore mechanism I'd suggest something much simpler, but also more general: Include \dump in all versions of TeX. The initex/virtex distinction is simply not worth the bother of retaining any more - the memory differences are trivial in modern terms. \dump should ideally also be extended to allow dumping from inside of a group. Once you do this, it becomes possible to write programs to analyze format files. Since the analysis programs would not be part of TeX proper, they can be as elaborate as anyone might want to make them. Sorting would be no problem, for example. None of this would cost anything in TeX itself, and except for the ability to \dump inside a group, if added, would not even have an effect on whether the resulting E-TeX is properly "TeX", since it would be invisible to any possible trip test. 2. Proposed additions to TeX (concrete new primitives) 2.0. General (Jun 92, Jul 92, Aug 93, Nov 94) A rather long list of proposed primitives (more or less worked out) was posted by Karl Berry on 10-Jun-1992. It contains suggestions like: \elseif (selfexplaining), \format{foo} (allow the author to select a format), \host{name} \host{os} \host{type} \getenv to extract host information \TeXversion, \usertime, \everyeof, and others. It is currently not possible to get some information about the current mode of TeX and make conditionals dependent on it and/or restore it after some action (see 2.7. \interactionmode). More of this kind are a conditional or primitive to signal, when TeX is in "expand only" mode (\edef, \mark, \write, ...), when TeX is scanning numbers (here I'm thinking---and hating---german.sty's active doublequote, which can also be used as a marker for hexadecimal numbers), when TeX is peeking for some special tokens (first column in an \halign), etc... A list of new primitives, enhancements and extensions has been prepared on the NTS meeting at Lindau in October 1994. It was posted to this list in November 1994 and provoked plenty of discussions and additional suggestions. Under the discussed primitves were \bind, \contents, \futuredef, \ifdefined, \ifcsname ... \endcsname, \unless, \OSname, \defaultextension. 2.1. \lastmark etc. (Jun 92, Jul 92) Currently you cannot remove a \write or \mark or \insert or rule from any list at all. If we allow them to removed, how will the commands appear to the user? If we have \lastmark like \lastbox, then perhaps we need a mark data type so that we can say something like \setmark0=\lastmark. It will probably be difficult in the case of \insert's to think of a good command syntax. Perhaps \lastpenalty, \lastkern, \lastskip should remove the penalty, kern, skip, ... so that they are consistent with \lastbox. Then \unpenalty, \unkern, and \unskip would be unnecessary. (Of course most macro packages would probably want to reimplement them, as macros: \def\unpenalty{\count@\lastpenalty}, \def\unkern{\dimen@\lastkern}, \def\unskip{\skip@\lastskip}.) 2.2. \system (Mar 93) 2.2.0 General Oops, this got rather longish, but this topic has caused plenty of traffic. I decided to quote directly the positions of both sides. The subpoints are 1. Pro 2. Contra 3. Syntax 2.2.1 Pro First comes the proposal as formulated by Phil Taylor: There has been much discussion on how a \system primitive might interact with different operating systems, each with different functionality and a different syntax. My idea was to extend the concept of a `TeX implementation', which at the moment implies the creation and application of a change-file to the master TeX source, to include an implementation- specific macro library. Thus each implementor, as well as creating and applying a change file, would also be responsible for mapping a well-defined set of macros, through the \system primitive, to the syntax and functionality of the operating system for which he or she has assumed responsibility. To cite a specific example: Assume that in e-Lib (a hypothetical macro library to accompany e-TeX), a macro \sys$delete_file {} is partially defined; then each implementor would be responsible for mapping \sys$delete_file { to his or her own implementation of \system. e-Lib would define the effect and the result(s), \system would provide the interface, and the implementor would be responsible for providing the mapping. The question has been asked: ``Why via \system and macros? Why not via explicit primitives to carry out the various functions that are envisaged?'' To which I would suggest that the answer is ``Because `the various functions which are envisaged' is both enormous (requiring many new primitives), and yet not large enough (because no matter what functionality we posit, someone will come up with an idea that has not been considered).'' By implementing just one \system primitive, and an extensible e-Lib macro library, one can create a robust and well-tested e-TeX whilst allowing new system interactions to be added at the simplest points: through the implementation-independent and implementation-specific components of e-Lib. 2.2.2 Contra And here's from the ``Minority Report'' (Tim Murphy and J"org Knappen) May I recall the immortal words of Ken Thompson, "A program should do one thing, and do it well." (TM) I don't like the hackers to decide, making eTeX yet another programme from which I can send e-mail and read news :-) Maybe people will tell me eTeX is a fine operating system, but TeX version $\pi$ is the better typesetter :-) But there is another side of \system, I want to call it the monstrosity side. Many people are thinking now, that TeX is a monster and difficult to tame. \system will add to this monstrosity. It will create a new paradise for hackers creating system hacks. And it will make people turn away from eTeX and use other products, even if they are far less secure. (JK) 2.2.3 Syntax If a \system command is required, should it not have a similar syntax and semantics to the a similar TeX command. I can't think of anything else in TeX (prepares to be shown wrong) that expands in the mouth and has side-effects. Should it not be like \read, \write etc. that is it generates a whatsit that is obeyed at shipout, unless preceded by an \immediate, in which case it is done immediately by the stomach. There seem to be two obvious syntaxes, one like \write: \system{foo} or \immediate\system{foo} and one like \read: \system{foo} to \baz or \immediate\system{foo} to \baz The latter one would produce the exit code into \baz. Should this be done with catcode 12 characters, or should it be done like \read, with the current catcodes? 2.3. \skylineskiplimit, \skylinehorizontallimit see section 1.1.1 2.4. \directioncode (May 1993) A \directioncode (with syntax analogous to \uccode, \lccode, sfcode) to be assigned to each input character. The basic ones are 0 -- transparent (space, full stop...) 1 -- left-to-right (latin letters, digits...) 2 -- right-to-left (hebrew letters, arab letters...), a truly international NTS will also have codes for vertical typesetting and some special cases. The question is how to use this idea consistently. One could extend the notion of TeX's modes. Horizontal mode is in fact left-to-right mode, a right-to-left mode is missing. To be complete, this mode will be equipped with boxen and all the stuff TeX's left-to-right (aka horizontal) mode has. At the beginning of a paragraph NTS decides which mode to choose by the \directioncode of the first input character. Sometimes the first character will have the wrong code, in this case the insertion of an explicit control sequence (like \lrbox{}) is necessary. If a character with another directioncode occurs, NTS starts a \rlbox and finishes it as soon as a character with the original \directioncode appears or at the end of the paragraph. For the building of right-to-left tables a \rlalign is needed. 2.5. \textcode (Sep 93, Nov 93) Some of the character coding discussions in the Technical Working Group on Multiple Language Coordination and some experiences I've made with `german.sty' (specially the problems with an active doublequote and hex integer constants!) lead to this _incomplete_ proposal/idea for the following addition: Introduce something like \textcode (and \textchar & \textchardef) which are the text (hmode) equivalent of TeX's \mathcode (and \mathchar/\mathchardef) primitives. With an equivalent and appropriate implemented \textcode primitive (with the choice to define a character as "pseudo-active"), it would be possible to * relate characters to different fonts (using a generalized `fam' of \mathcode) * suppress expansion of active characters (it will only be expanded, if it is read to form the hlist) (using an equivalent \mathcode="8000 value) [This point allows the use of e.g. a pseudo-active " which expands to non-expandable tokens and it removes the special construct \relax\ifmmode... for active characters, too.] 2.6. \afterfi (August 1993) In the answer to an exercise of the ``Around the Bend'' series, Michael Downes realised the non-existence of an \afterfi primitive (Note: He did not demand it nor really miss it). Perhaps an \afterfi can simplify some obscure mouth-only macros with nested conditionals??? \afterfi should be expandable, because \if...\fi is expandable. 2.7. \interactionmode (Nov 93, was: \currentinteractionmode) Add a new count register \interactionmode, which reflects the user interaction level with the following values: <= 0 batch_mode = 1 nonstop_mode = 2 scroll_mode >= 3 error_stop_mode The commands \batchmode...\errorstopmode, the "options" 'Q', 'R', 'S' in the interactive error routine and all other TeX internal interaction level changes (e.g. after an interruption) access this new register. The level changes in the interactive error routine and the old commands should always work, even if the symbol \interactionmode is redefined (this means that the user can redefine \interactionmode, but the commands \batchmode...\errorstopmode still work). Examples: \ifnum\interactionmode<1 \AskUser \else \UseDefault \fi {\interactionmode=0 % switch to \batchmode \global\font\test=xyz10 % try to load font }% % restore former interaction level % ... now test if font has been loaded % without error (i.e. != nullfont) 2.8. \mathspacing (Nov 93) Add a new count register array \mathspacing with 64 entries (we really need only 64-8=56 entries, because some of them are never used, but to simplify things 64 are used) with the following syntax: \mathspacing = % 0 <= <= 63 The spacing specified in number is inserted between the two math atom types specified in number . The two numbers are coded as = * 8 + = ( ( * 256 + ) * 256 + ) * 256 + This means that is easily expressed in octal and in hexadecimal notation. <..._atom_type> is one of the following seven types: 0 ordinary 1 large operator 2 binary operation 3 relation 4 opening 5 closing 6 punctuation 7 delimited subformula <..._spacing> can be specified separately for each of the four math styles (display, text, script and scriptscript) with the following values: 0 no space 1 thin space (specified by \thinmuskip) 2 medium space ( -- " -- \medmuskip) 3 thick space ( -- " -- \thickmuskip) 4-255 reserved for other things (e.g. other spacings and/or additional penalties, like \relpenalty, \binoppenalty, ...) For more information see TeXbook, pp. 170f & Appendix G and TeX--The Program, \S 764ff. Examples: (using TeX's standard spacing) Between an `ordinary' (= 0) and a `relation' (= 3) atom a thick space (= 3) is inserted, but not in script or scriptscript style. \mathspacing '03 = "0033 Between a large operator (= 1) and an ordinary (= 0) atom a thin space (= 1) is inserted: \mathspacing '10 = "1111 Necessary changes for the sketched proposal are very simple to implement. The syntax of \mathspacing is awful, but this is true for \mathcode, \delimitercode, etc.etc., too. 2.9. \inputescapechar (Nov 93, alternative names were proposed during discussion) Use two new internal count registers \inputescapechar and \outputescapechar to specify the "escape" character to be used for unprintable characters. If \inputescapechar is not in [0..255], TeX's behaviour is used, i.e., two equal characters with category 7 are used as a prefix for a ~~x notated character, otherwise two characters with code \inputescapechar are used for this prefix. If \outputescapechar is not in [0..255], the character `~' is used when an unprintable character has to be written. The default values of these two registers are \inputescapechar = -1 \outputescapechar = `~ to be compatible with TeX's standard behaviour. 2.10. \middle (Jan 94) Generalize the syntax of \left \right to allow one or more optional \middle delimiters, \left \middle \right. The middle delimiter grows to the same height as the left and right ones. Constructions of this kind are used throughout quantum mechanics (Dirac notation), e.g. $$ \left\langle A \middle% \hat{O} \middle% B \right\rangle $$ By default, the \middle element is \mathrel, you can make it \mathord by embracing it {\middle %}. 2.11. \outputunit (Dec 93) \outputunit unit to be used when e-TeX shows some dimen or glue value Usage: \outputunit= (s. TeXbook p. 270) Rationale: Currently TeX forces the user to think in points when it outputs any dimen or glue value (e.g. when issueing an overfull hbox warning). But a program should adapt to the conventions of the user instead the other way round. The addition of \outputunit whould make TeX much more user-friendly, since only a few people are thinking in points. 2.12. \protected (new prefix for macro definitions) (Nov 94) Analogous to \long, \outer, etc., causes the associated macro to be non-expanding in contexts where such behaviour is likely to be undesirable (in particular in \writes and \edefs); an explicit \expandafter \empty may be used to force expansion in these circumstances. 2.13. \scantokens (Nov 94) Allows an existing token-list to be re-input under a different catcode regime from that under which it was created; as it uses all of TeX's present \input mechanism, even %%ff notation will be interpreted as if \input. Causes an `empty filename' to be input, resulting in `( )' appearing in the log file if \tracingscantokens (q.v.) is strictly greater than zero. If the token list represents more than one line of input, and if an error occurs, then \inputlinenumber will reflect the logical input line from the token list rather than the current input line number from the current file. 2.14. \deferred (Nov 94) At the moment, only \writes are deferred; there are cases when it would be desirable for other things, too, to be expanded only during \shipout, and \specials are one of these. (See also 2.2.3) 2.15. \everyeof {} Provides a hook whereby the contents of a token list register may be inserted into TeX's reading orifice when end-of-file is encountered during file reading. Would not be invoked of the file indicated logical e-o-f through the medium of \endinput. Proposed by Phil to allow clean processing of file-handling code which requires a (sequence of characters yielding) \else or fi to be found in a file, where no such sequence can be guaranteed. 2.16. \readline to (Nov 94) Allows a single line to be read from an input file as if each character therein had catcode 12. Intended to be used for verbatim copying operations, in conjunction with \scantokens, or to allow error-free parsing of `foreign' (non-TeX) files. 2.17. \evaluate {} (Nov 94) Intended for use on the r-h-s of \count, \dimen and \skip assignments, it would allow the use of infix arithmetic operators such as +, -, * and /; the type of the result would, in general, be the type of the simplest operand forming a part of the expression, and the normal semantics of TeX would allow this to be further coerced where necessary. Parenthesised sub-expressions would be allowed. Additional operations were proposed: trigonometric functions, square root, pythagorean addition (METAFONT's `++'), angle{}. 2.18. \ifevaluate{}....\else....\fi (Nov 94) The boolean expression should allow for the usual operators like .not., .and., .or., .xor., =, >, <, <=, etc., if feasable it should also allow the evaluation of arithmetic expressions. Some functions are needed to get information old TeX \if's know now, I propose the name \query to these functions. At the moment I see need for the following \query functions: \queryeof returns `1' if eof, `0' else \querymode returns `0' in the infamous `no mode' (do we need to differentiate finer here? `1' in vertical mode `2' in restricted vertical mode `3' in horizontal mode `4' in restricted horizontal mode `5' in math mode `6' in restricted math mode With \querymode \ifmmode, \ifhmode, \ifvmode, and \ifinner can be emulated. \querydefined{\cs} `1' if defined, `0' else \querycsname...\endcsname (ditto) 2.19. \everyeof {} Provides a hook whereby the contents of a token list register may be inserted into TeX's reading orifice when end-of-file is encountered during file reading. Would not be invoked of the file indicated logical e-o-f through the medium of \endinput. Proposed by Phil to allow clean processing of file-handling code which requires a (sequence of characters yielding) \else or fi to be found in a file, where no such sequence can be guaranteed. 2.20. \fancyprompt allow configuration of the TeX prompt (default `*'), e.g. *\fancyprompt{eTeX>} eTeX> 3. Metaremarks 3.0. General Remarks about group efforts vs. one person creating software (Mar 93), ALGOL 68 as a warning example 3.1. TeX is not perfect (Jun 92, Jul 92) The discussion has taken place in June and July 1992. Several details were worked out, where TeX could be improved. Another point of criticism was the programming language of TeX in general, several participants prefer a procedural language over a macro language. 3.2. In which language shall NTS be written (Mar 93) In 1992, there was much discussion, in which language an NTS should be implemented (candidates were LISP, C, and WEB). This has settled in March 1993 (to PASCAL-WEB), because of the acceptance of the idea that rather than wait for an ``all-singing, all dancing'' NTS, the group should develop, in a stepwise manner, small but significant enhancements to TeX. This implies that the enhancements are implemented as change files in WEB. 3.3. The TeX language (Oct/Nov 93) The TeX language is a simple language in the sense, that it contains only few concepts. It belongs to the Lisp family. In particular, it's a list-based macro language with late binding. Actually, that's all one needs to say to characterize it. Its data constructs are simpler than in Common Lisp: `token list' is the only first order type. Glue, boxes, numbers, etc., are engine concepts; instances of them are described by token lists. Its lexical analzsis is simpler than CL: One cannot program it. One can only configure it. Its control constructs are simpler than in CL: Only macros, no functions. And the macros are only simple ones, one can't compute in them. 3.4. The TeX engine (Oct/Nov 93) The TeX engine lies below the TeX language. Glue, boxes, numbers, etc. belong to the TeX engine. The registers of the engine can be changed by TeX's primitives. However, the latter seem to be quite unregular and baroque. 3.5. Extensions and Enhancements (Nov 94) `Extensions' are basically new primitives which have no effect on the semantics of existing TeX documents, except insofaras any document which tests whether such a primitive is, in fact, undefined, will clearly obtain opposite results under TeX and e-TeX; `enhancements' are more fundamental changes to the TeX kernel which may affect the semantics of existing TeX documents even if no new primitive is used or even tested. Such changes may be, for example, differences in the construction of TeX's internal lists, or perhaps different hyphenation or ligaturing behaviour. 4. Deviations 4.0. General (empty) 4.1. Automated Kerning (Oct 92) Kindersley's "optical kerning": for the purposes of kerning, each character is replaced by a circle centred on the centre of gravity of that character; the radius of the circle is determined by the fourth moment of the character (that is, the fourth root of the sum over all black pixels of the fourth power of their distance from the centre). On the UKTUG trip to Kindersley's studio, I tried to extract the reason why the fourth, as opposed to third or fifth or whatever, moment is used; the reason is apparently that it "looks right". We can construct elaborate schemes for kerning (Kindersley's fourth moments, FontStudio's (convex?) envelopes, Calamus' eight widths, etc), but the proof of the typographical pudding is in the eating of the resulting words, so to speak. 4.2 About Lout (June 1993) In June 1993, the new system Basser Lout caused several questions and suggestions on this list. The following is taken from a short review of Lout by Bernd Raichle: `Lout' is a (yet another) document formatting system, released under the terms of the GNU General Public License and available on some ftp servers. IMHO it's more like a `troff' (with a better input language and some newer concepts) than a `TeX'. A few citations from the documentation of lout: Lout is a high-level language for document formatting, designed and implemented by the author. The implementation, known as Basser Lout, is a fully operational production version written in C for the Unix operating system, which translates Lout source code into PostScript, a device-independent graphics rendering language accepted by many high-resolution output devices, including most laser printers. [...] When expert users can implement such applications quickly, non-experts benefit. Although Lout itself provides only a small kernel of carefully chosen primitives, Lout has 23 primitive operators... missing, for example, the simplest arithmetical operators (there is only the operator "@Next" which increases a number by one). packages written in Lout and distributed with Basser Lout provide an unprecedented array of advanced features in a form accessible to non-expert users. The features include rotation and scaling, fonts, These features are mostly based on the output language... Postscript (if you are looking inside a Lout package, you find large portions of embedded Postscript code). paragraph and page breaking, TeX does a better job for these two items, because Lout is missing most of TeX's paragraph/page breaking parameters. (Note: Lout uses TeX's hyphenation algorithm and the hyphenation patterns.) displays and lists, floating figures and tables, footnotes, chapters and sections (automatically numbered), running page headers and footers, odd-even page layouts, automatically generated tables of contents, sorted indexes and reference lists, bibliographic and other databases (including databases of formats for printing references), equations, tables, diagrams, formatting of Pascal programs, and automatically maintained cross references. TeX's math setting abilities are better. Lout uses a package named `eq' derived from the `eqn' preprocessor used with `troff'. And there are other packages named `tab' (for tabulars) and `fig' (drawing figures). [...] Lout is organized around four key concepts -- objects, definitions, galleys, and cross references -- [...] The concept of `galleys' and the "expansion" of recursive defintions are IMHO the only new concept in Lout: `galleys' are a way to describe a page, dividing it in certain regions which can be filled from different sources (e.g. a footnote galley is filled with footnote text, etc.). Recursive definitions are very simple, e.g. def @Leaders { .. @Leaders } defines the command (Lout calls it `object') to "expand" to a `..' and if there is place for another "expansion" it is called again. For example \hbox to 4in{Chapter 7 \dotfill 53} is in Lout 4i @Wide { Chapter 7 @Leaders 53 } With this recursive definitions, a whole document is defined as a @PageList consisting of a @Page and a @PageList with an incremented @PageNum. A @Page is defined as a set of `galleys' (header, text body, footnotes, footer), which are also defined as a list of text/footnotes/... and so on. Perhaps others can add more impressions, mine are based on the documentation coming with the Lout package and some tests done in 1-2 hours. 5. Proposed changes to METAFONT 5.O. General Most of the General points in the discussion of TeX also apply to METAFONT. 5.1. Writing to auxiliary files METAFONT should be able to write to auxiliary files. Desparately needed for packages which allow to draw figures in METAFONT and label them in TeX (BoF session at EuroTeX '92, Prague). 6. Proposed changes to the tfm file format 6.1. Vertical kerning (Jun 92) This may sound exotic to you, but the AFM format can do it. And I desperately need it for non-Latin alphabets (everytime I need a character in a ligature raised or lowered, I have to reserve a position in the font for it). 6.2. Cross-font ligatures (Jun 92) Ligatures pointing to other fonts!! Yes, imagine for example that your 256 chars are full, you could still access characters by ligatures...but they have to be in the same font. 7. Proposed changes to the dvi file format 7.0. General A longer discussion about colour inclusion into the dvi-file occured in Oct/Nov 93. The outcome was, that high-quality colour handling is a difficult and device dependent task, better left to the printer driver. Colour can be handled by the \special command which is sent directly to the driver. On comp.text.tex I saw a remark, that specials may occur on places in the dvi-file, where TeX can't put them. Maybe eTeX should take care of this point. 7.1. Horizontal and Vertical Mirroring (Nov 94) Allow a glyph to be mirrored in a vertical axis, placed centrally; mirrored in a horizontal axis, placed at half the x(?)-height. It was noted, that the specification of the axes might cause problems. It was also noted, that several drivers understand specials for the requested action. 7.2. Rotation of a glphy by 180 degrees (Nov 94) This would greatly increase the power of vpl files (e.g. define schwa by rotating e), and I hope that it would be possible for drivers to implement. It was noted, that this is fine, but rotation by 90 degrees cannot be device independent, because there are printers with rectangular (non-square) pixels. 8. Additions to the TeX kit 8.0. General The TeX kit, meaning the bundle of programmes building up a complete TeX installation, contains besides the major programmes TeX and METAFONT additional programmes, traditionally divided into several groups: a) TeXware: dvitype, tftopl, pltotf, pooltype, pktype, patgen, vftovp, vptovf, (outdated: pxtopk, pktopx) b) MFware: mft, gftodvi, gftopk, gftype, (outdated: gftopxl) c) webware: weave, tangle And, of course, drivers for the screens and printers. Other utility programmes more loosely associated with TeX are indexers (makeindex, idxtex, glotex, ...) and bibliography tools (bibtex, bibclean, ...). Makeindex and Bibtex are standard tools, because they are described in the LaTeX manuals. Another group includes drawing programmes (TeXCAD, xfig, gnuplot, ...) to create graphics to be imported to TeX documents and graphics converters (ghostscript, xv, bmt2font, ...). 8.1. dvicopy Written in WEB by Peter Breitenlohner Available from CTAN Resolves all references to virtual fonts in a dvi file. Can be used to create a dvi file without virtual fonts (e.g. for drivers which cannot handle vf's). 8.2. dv2dp and dp2dv Written in C by Geoffrey Tobin Available on request from the author dv2dp makes a human readable file from a dvi file. The companion programme dp2dv makes a dvi file frome a dvi property list file. It allows you to change dvi files with a text editor. 8.3. gftotxt Written in C by Yannis Haralambous Available ? Extracts a certain kind of specials from the gf file into a text file. Workaround for the famous non-ability of METAFONT to write auxiliary files. 8.4. qdtexvpl Written by Eberhard Mattes Available from the CTAN archives Generates a virtual font from TeX macros. qd = quick and dirty. 8.5. fontinst Written in TeX by Alan Jeffrey Available form CTAN Makes virtual fonts from afm files (PostScript fonts). Used by the LaTeX team to create virtual fonts and fd files for LaTeX2e. 8.6. afmtotfm Written in C by Thomas Rokicki Available from CTAN (part of dvips distribution) Makes tfm files from afm files. 8.7. dvipaste Part of LAMSTeX (Michael Spivak) Allows to paste a dvi-file into another one. 8.8. METAPOST Written in Web by John Hobby Now freely available from the CTAN A tool to build graphics to be included in TeX documents, uses a METAFONT style language. 8.9. Tie Written by Klaus Gunthermann Available from CTAN Allows to process multiple change files to a given Web source. A tool of this kind is needed to build any eTeX distribution, since one has tex.web, etex.ch, tex.ch-system, and etex.ch-system files to deal with. 8.10. dvi2dvi Available from CTAN Postprocessor for dvi files, allowing e.g. 2-up or 4-up layout of full pages. 9. Existing extensions to TeX 9.0 General Here I include short description of existing extensions to TeX. This section does not include commercial TeX implementations because of lack of solid enough information on them. It also does not yet include ETeX, NTS, nor Omega. 9.1. TeX-XeT (Don Knuth and Pierre MacKay) CTAN: tex-archive/systems/knuth Adds new primitives \beginR, \endR, \beginL, and \endL for right to left typesetting. Despite their names, the new primitives are insensitive to TeX's grouping mechanism. Outputs a special file in dvi-ivd format, which need be resolved by another programme or requires special drivers. 9.2. TeX--XeT (Peter Breitenlohner) Available from CTAN tex-archive/systems/knuth/tex--xet Implements the same primitives as TeX-XeT, but outputs a normal dvi file. Passes the trip test (as far as I know). 9.3. MLTeX (Michael Ferguson) Available from aldebaran.ee.mcgill.ca in pub/mltex Adds a primitive \charsubdef, which allows the following: input of 8-bit characters, participation of accented characters not present in the output encoding in hyphenation, ligatures and kerning with accented letters. Especially popular in francophone countries. emTeX supports MLTeX via the /ml option. 9.4. SiSiSi (Wilhelm Barth, Heinrich Nirschl, Heini Hofstaedter und Harald Mueller) Available from CTAN: tex-archive/systems/vms (sic!) Implements another hyphenation algorithm (the name means SIchere SInnentsprechende SIlbentrennung, which is translated `secure sensitive hyphenation') specific to the german language (with Haupt- und Nebentrennstellen). Fails the trip test, therefore not a legitimate TeX. In use at TU Wien (Austria). 9.5. Japanese TeX [in preparation] The End.