diff options
Diffstat (limited to 'usergrps/uktug/baskervi/6_3/ht.tex')
-rw-r--r-- | usergrps/uktug/baskervi/6_3/ht.tex | 693 |
1 files changed, 693 insertions, 0 deletions
diff --git a/usergrps/uktug/baskervi/6_3/ht.tex b/usergrps/uktug/baskervi/6_3/ht.tex new file mode 100644 index 0000000000..5da3bb488c --- /dev/null +++ b/usergrps/uktug/baskervi/6_3/ht.tex @@ -0,0 +1,693 @@ + +\MakeShortVerb{\|} +\def\cmd#1{\texttt{\char'134#1}} +\title{How does \acro{HTML} handle mathematics?} +\author{Malcolm Clark} +\begin{Article} +\section{Introduction} +There is a very short answer to the question posed in +the title: not at all. However, as an instantiation of +SGML (see, for example, \cite{EvH}, \cite{Goossens} and \cite{Bryan}), +we can look first at how mathematics is handled in other +\acro{SGML} \acro{DTD}s; and then examine how the expired \acro{HTML}3 draft +proposed to include mathematics. + +The \acro{DTD}s already available which are designed to +handle mathematics include \acro{ISO}\,9573 (\cite{ISO9573}) +(confusingly, also known as \acro{ISO}\,12083), +which is part of \acro{CALS}, \acro{AAP} (Association of +American Publishers) (\cite{AAP}), and the \acro{HTML}3 draft (\cite{math}). +The Euromath \acro{DTD} might also be relevant (\cite{Euromath}), +but since its status is closer to that +of proprietary it is too awkward to consider here. +Van Herwijnen (\cite{EvH}) comments on the first two, +comparing them to \TeX\ and |eqn|. +A longer and more detailed examination of the \acro{AAP}, Euromath and +\acro{ISO}\,12083\slash \acro{ISO}\,9573 is given in Poppelier, van +Herwijnen and Rowley (\cite{PvHR}). + +Van Herwijnen provides an example from physics for the decay of a particle + together with representations in +\TeX, |eqn|, \acro{ISO}\,9573 and \acro{AAP}. +The equation is: +\begin{displaymath} +\Gamma(J/\psi\rightarrow\eta_c\gamma) += \frac{\alpha Q_c^2}{24} +\left\vert A(J/\psi\rightarrow\eta_c\gamma)\right\vert^2 +\frac{m_\psi^2}{m_{\eta_c}^2} +\left(1-\frac{m_{\eta_c}^2}{m_\psi^2}\right)^3 +\end{displaymath} +The entire expression is too extensive to compare here, but +the left hand side of the equation in \acro{ISO}\,9573 may be +given as +\begin{Verbatim} +<mfn> <fname> Γ <of>J/ψ → η +<sub> c</sub> γ +\end{Verbatim} +while using the AAP dtd, it could be +\begin{Verbatim} +<g>G</g>(<fr sol>J</><g>y</g> &ar; <g>h</g> +<inf>c</inf><g>g</g>) +\end{Verbatim} +For the sake of completeness, the |eqn| alternative is +\begin{Verbatim} +Gamma(J/psi rarrow eta sub c gamma) +\end{Verbatim} +Eric's comments are interesting. +He comments that the two \acro{SGML} representations are cumbersome +and difficult to read, especially when contrasted to \TeX\ and |eqn|. +He also rails against the obsession with representation. +For example $\Gamma$ `means' decay width, but as far as the \acro{DTD}s are +concerned we have \texttt{\Γ} and \texttt{<g>G</g>}. +To be fair, \TeX\ and |eqn| hardly +fare better, but at least we do know that we could have +provided a more meaningful command. The second representation (\acro{AAP}) is +particularly unfortunate, since instead of treating the symbol +as a symbol, it treats it as a Greek letter. Of course, Eric is deeply +imbued with the basic notions of \acro{SGML}, and would be very +sensitive to this. + +His contention is that someone who already knew \TeX\ or |eqn| would +have no motivation for learning or using these rather baroque alternatives. + +There is a point to be made about the rather cumbersome nature of +the \acro{SGML}. Writing it by hand will be cumbersome, but surely no-one +ever wants to write in this way. Structure editors are available. + In the \TeX\ world, Scientific Word gives a structure editor +for \LaTeX. This can be done since it is possible +to hold an equation as elements in a tree structure, so that modification +or correction to an element can be managed quite simply, and changes can +propagate down the tree. The same sort of thing exists within the \acro{SGML} +world. Euromath uses the Grif (\cite{Grif} and \cite{GrifSA}) + editor for just this, and it would +be easy to see other similar editors, like Chamberlin's +Quill (\cite{Chamberlin}) +maintaining the information. There is a question lurking whether +mathematicians would actually like to input in this way. +Just as experienced keyboarders find +\acro{GUI}s very difficult and slow to use, perhaps the same sort of +resistance would +be found. However, the real point is that humans should not be expected +to write \acro{SGML}. If they really must write \LaTeX, then an approach +like Scientific Word, which could be coerced into generating a tree structure +which could be mapped onto a \acro{DTD}, is potentially more valuable. + +A general issue, which Eric raises implicitly, is that none of the \acro{DTD}s +offer a way of encoding meaning in a flexible way. Either the +element is present already, or it is not. There appears +to be no straightforward way of extending the range of elements. In the world +of high energy physics and mathematics this must be something +of a straightjacket. On the other hand, the bane of many editors' +lives is the ease with which individual authors +can `extend' \TeX\ or \LaTeX\ by adding a few new definitions. + +\subsection{SGML Notations} +If we really did have existing equations, then one way to +handle them within \acro{SGML} is through a Notation (see also \cite{Bryan}). +A Notation permits +a document to include data which is not to be parsed. It is +therefore possible to include \TeX\ or \LaTeX\ and assume that +at that point a convenient processor will be magicked to deal with +it. He gives the example of the definition in a \acro{DTD}: +\begin{Verbatim} +<!NOTATION TeX SYSTEM ""> +<!NOTATION LaTeX SYSTEM ""> +<!ELEMENT Formula - 0 CDATA> +<!ATTLIST Formula #NOTATION (TeX|LaTeX|eqn) + #CURRENT> +\end{Verbatim} +which may then be used later with the |Formula| element as +\begin{Verbatim} +<FORMULA NOTATION=TeX> +... +</FORMULA> +\end{Verbatim} + +A scheme which already maps \acro{SGML} to \LaTeX\ (e.g.~\cite{Flynn} or + \cite{Goossens2}) would +find this a very easy way to absorb maths, provided all the +equations were in the same notation. The prospect of a +\cmd{newcommand} or \cmd{def} within the Notation could be +worrying. + +Although this sort of expedient is plausible, it is not entirely +successful. One of the arguments behind the use of \acro{SGML} is that it codes +structure or meaning, rather than appearance (to echo one of Eric's points). +\TeX\ and \LaTeX\ sometimes code meaning, +but not in a consistent and reliable way. +How do we extract information? If we have \acro{SGML}, it is relatively +easy to find corresponding structural elements, which may then be extracted. +Once we start including a Notation, this chance is all but gone; +and if we include alternative Notations (say \TeX, \LaTeX\ \emph{and} |eqn|) +it becomes even more problematic. + +It is not clear to me how these Notations, or even the \acro{DTD}s +differentiate between in-line and displayed equations. I assume +that an attribute could be included which specified the style. +On the other hand \acro{ISO}\,12083 distinguishes between +in line, displayed and `display formula groups' styles, through the +use of different elements. + +There is perhaps a deeper question here, +which is this, should it be at all relevant? +Should the author be able to specify that +some equations are in-line and others are to be displayed? +It should make no difference at all to the content, although +it would make great changes to the appearance. But to deal with maths +is to deal with appearance, to a large extent. The display seems +to be a key issue. Since many equations are strongly two dimensional +(as opposed to the one dimensional nature of most text), it is a key question +whether it is reasonable to expect this aspect to be reflected in +any linearisation. + +To give \acro{ISO}\,12083 its due, it says ``Since there is no consensus +on how to describe the semantics of formulas, it only describes +the presentational or visual structure.'' + +\section{\acro{HTML}3 (expired draft)} +\acro{HTML}3 supports a |<math>| element which provides some +capability for the inclusion +of maths expressions. The draft (\cite{math}) + does indicate that this capability is limited. +It describes the functionality as similar to that found in ``common word +processing packages''. In itself this may be seen as a hint of one of +the driving +forces within \acro{HTML}3 -- a desire to emulate word processor +functionality. For a language which derives initially from a high +energy physics community, this is a modest aim. The other design aim +of |<math>| in \acro{HTML}3 is to be ``concise and comparatively easy +to read''. As a rider, it is suggested that this will make formulae +longer than \LaTeX, but shorter than Euromath or \acro{ISO}\,12083. + +It is stated that the maths owes ``a lot to \LaTeX's math mode''. In some cases +it uses names for elements which are derived from \TeX/\LaTeX. Immediately a +misconception springs up. The second paragraph of the section on maths gives +examples of \LaTeX\ commands: namely \cmd{atop}, \cmd{choose} and \cmd{sqrt}. +Only \cmd{sqrt} is a \LaTeX\ command. The \cmd{atop} and \cmd{choose} + are unrepentant \TeX. Ignoring +for a moment the fact that there is a clear confusion between \LaTeX\ and \TeX\ +in the mind of the author of the draft (Dave Raggett),\footnote{I suspect +this is quite a widespread misunderstanding.} there is actually a deeper problem. +The use of an operator like \cmd{over} requires much more work on the part of a +processor, since it is often not until the whole expression has been parsed that +sense can be made of what is actually \cmd{over} what (see \cite{seb}). +From years of teaching +\TeX, I can confirm that the \cmd{over} command (and its buddies, \cmd{above}, +\cmd{atop}, \cmd{choose}, \cmd{brack} and \cmd{brace}) can lead to immense +frustration and confusion. + +\acro{HTML}3 also has an |<above>| token, but this is quite different from +\TeX's \cmd{above}. Since +the \TeX\ command is quite typographic, specifying the width of the line +separating numerator and denominator, this is probably not too ambiguous. The +function of the \acro{HTML}3 tag is to allow something to be drawn above an +expression. There is a similar |<below>| tag. +|<above>| is a sort of numerator operator: +\begin{Verbatim} +<math> +<above>x+y</above> +</math> +\end{Verbatim} +gives something like +\begin{displaymath} +\overline{x+y} +\end{displaymath} +with the nuance that the element takes an attribute |sym| which can specify +other symbols: |line| (the default), |cub|, |larr|, |rarr|, |hat| and |tilde| (in +a sense |hat| and |tilde| correspond to \cmd{widehat} and \cmd{widetilde}). +For example, +\begin{displaymath} +\overbrace{a+b+c} +\end{displaymath} +would be +\begin{Verbatim} +<math> +<above sym=cub>a+b+c</above> +</math> +\end{Verbatim} +the corresponding |<below>| has the same list of possible |sym| attributes, +though I don't immediately understand what |hat| and |tilde| would do. + +There are also some terminological surprises. |<math>| almost borrows +the use of underscore and carat/circumflex for sub-scripts and +super-scripts, except that they are referred to as index and exponent. +The |^| and |_| are actually |shortref|s for |<sup>| and +|<sub>|.\footnote{The notion of exponent and index may be welcome, +since we would expect \acro{HTML} to be concerned with the underlying +content. Perhaps \TeX\ does tend to overload the idea, since a superscript may +signify more things than just an exponent. A semantic separation +might be a very good +idea, but this is unlikely to be the intention, since they are +|shortref| characters for |<sub>| and |<sup>|. +Poppelier \textsl{et al.}\ put this very succinctly: +``What is the function of the 2 in SU$_2$, $\log_2x$, $x_2$, $x^2$ $T_2^2$? +In SU$_2$ it is the number of dimensions in the Lie group; in $\log_2x$ it is +the base of the logarithm; if $x$ is a vector, the ${}_2$ in $x_2$ is an index; +the ${}^2$ in $x^2$ could be a power, but if $T$ is a tensor, the ${}^2$ in +$T_2^2$ is a contrainvariant tensor index.'' +} Since \acro{HTML} syntax and \LaTeX\ +syntax are rather different, the ease of \LaTeX's sub- and super-scripting has +to be abandoned. The \acro{HTML} tag has to be terminated. It is +unfortunate that there +was not a way of employing an implied end tag. +For example +\begin{small} +\begin{displaymath} +a^{23}_n +\end{displaymath} +\end{small} +is given from +\begin{Verbatim} +<math> +a^23^_n_ +</math> +\end{Verbatim} + +Should you need to subscript a +subscript, the |shortref| form cannot be used. Although |<math>| does support a +grouping operator, +\begin{displaymath} +a_{b_c} +\end{displaymath} +is obtained from +\begin{Verbatim} +<math> +a<sub>b<sub>c</sub></sub> +</math> +\end{Verbatim} +Perhaps cleverly, superscripting an expression with a binary operator results +in the expression being placed over the operator, like \cmd{stackrel}. +For example +\begin{displaymath} +A \stackrel{\alpha'}{\longrightarrow} B \stackrel{\beta'}{\longleftarrow} C +\end{displaymath} +would be +\begin{Verbatim} +<math> +A →^α′^ + B ←^β′^ C +</math> +\end{Verbatim} +In passing note that the use of the `prime' operator is different from +\TeX\slash \LaTeX\ use, and the use of space. I am not entirely clear about the +use of space in |<math>|. It is certainly not ignored, as in \TeX, and +the draft does comment on the use of different horizontal white space +within the equation. The draft runs +\begin{quote} +Spacing between constants, variables and operators is determined +automatically. Additional spacing can be inserted with entities such as +| |, |&sp;| and |&quadsp;|. White space in the markup +is used only to delimit adjacent variables or constants. You don't need +spaces before or after binary operators or other special symbols as these +are recognised by the \acro{HTML} math tokeniser. White space can be +useful, though, for increased legibility while authoring. +\end{quote} +This does imply a rather different use of space. +This use of space does have the effect that a `prescript' can be made +quite unambiguously simply by ensuring it is preceded by a space. +It would imply +a string |sin| is recognised as `some sort of function'. The string |xyz|, +would also presumably imply `function', while in \LaTeX\ it would imply +the three variables $x$, $y$ and $z$. +It does leave unclear how +\(\sin^2\theta \) and +\begin{displaymath} +\max_{i=1}^n x_i +\end{displaymath} +would be handled with ease. If the $\sin^2\theta$ uses |<sup>| +but the +\begin{displaymath} +\max_{i=1}^n x_i +\end{displaymath} +requires |<above>| and |<below>| then +we have an interesting inconsistency. + +|<math>| does adopt \TeX/\LaTeX's notion of binary operators, +and in general claims +to reflect the assumptions of \TeX/\LaTeX. It does not however provide support +for multi-line equations, stating that `this can be effectively handled by +combining math with the |TABLE| element'. To me this wanders far from the basic +concepts of \acro{SGML}. However, what it appears to mean is that +the |<array>| tag uses the same sort of syntax as |<table>|, not that +an array uses the table tags. + + +From the draft, it is anticipated that chemistry could be set from within +the |<math>| tag. I would view this as a mistake. It may be (almost) defensible +from within \LaTeX\ to use math structures, although the various chemistry +packages at least try to separate the notions. It seems unfortunate that HTML3 +should not attempt something similar.\footnote{And of course this +emphasises the +inadequacy of referring to a subscript as an index and a superscript as an +exponent. The terms are pretty meaningless for chemical notation.} An example +might be +\begin{Verbatim} +<math class=chem> +Fe_2_^2+^+Cr_2_O_4_ +<math> +\end{Verbatim} +for +\begin{displaymath} +\mathrm{Fe}_2^{2+}+\mathrm{Cr}^{}_2\mathrm{O}^{}_4 +\end{displaymath} +where the different notational style of chemistry is tackled, notably +its use of an upright font and consistent baselines for subscripts. + + +Some hints on appearance are provided: it is expected that functions +(operators), numbers and other constants are portrayed in an upright font, and +variables are italic. Unlike \TeX/\LaTeX, limits for integrals and summation +signs are said to be placed directly above or below, or to the immediate right +(depending on the symbol). Unfortunately, the draft does not indicate quite +what this ambiguous term means. I suppose it does not mean `emulate' the +\TeX/\LaTeX\ mode, though that is obviously plausible, and from the point of view +of a browser author could be a reasonable path. + +What does it look like? +\begin{Verbatim} +<math> +∫_a_^b^{f(x)<over>1+x} d x +</math> +\end{Verbatim} +for +\begin{displaymath} +\int_a^b\frac{f(x)}{1+x}dx +\end{displaymath} +Note that the sub- and +super-scripts, like \TeX/\LaTeX\ also denote limits. + + +Some maths accents are available: |<vec>|, |<bar>|, |<dot>|, |<ddot>|, |<hat>| +and |<tilde>|. There are no explicit equivalents for \cmd{check}, \cmd{breve}, +\cmd{acute} and \cmd{grave}, although they could be created with |<above>|. + +Another borrowing from \TeX/\LaTeX\ is the notion of grouping: +\acro{HTML}3 uses a |<box>| element where \TeX/\LaTeX\ would use parentheses. +|<box>| can be replaced by a |shortref| form of |{| +and |}|, which greatly aids brevity and comprehension.\footnote{Should +you need the symbols themselves, they are obtained by +the entities \texttt{\{} and \texttt{\}}.} +Although \TeX nically a braced group is a sort of `box', +perhaps |<group>| might have +been a better, though less concise term, in the context. +It is perhaps an unfortunate choice, since `box' carries overtones for +many \TeX\ +users. Still, it does ensure that all the power of grouping is present (fairly +essential in view of the |<over>| element). To overload slightly, one of the +attributes of the |<math>| element is |box|, which causes a box to be drawn +around the formulae. The |<box>| element is used in a number of ways; it is +used, for example with the |<left>| and |<right>| commands for delimiters which +grow. This leads to a rather strange construction: +\begin{Verbatim} +<math> +f(x)=<box> + (<left> 1+x <over> sin x <right>) + <\box> +</math> +\end{Verbatim} +where |(<left>| gives a left parenthesis of appropriate size and |<right>)| +gives the corresponding right parenthesis. +As with \TeX, it is recognised that sometimes it +may be necessary to have a delimiter larger than `default'. |<box>| therefore +has a |size| attribute to enable this to happen. The permitted values are +|normal|, |medium|, |large| and |huge|. The |shortref| form cannot +take attributes. + +Integrals (and other large operands which are stretchy) also need the use of +|<left>|, without any corresponding |<right>|. For example +\begin{Verbatim} +<math> +<box>∫_-&inf;_^&inf;^ + <left> f(x,y) <over> x^2^+y^2^ + </box> d x d y +</math> +\end{Verbatim} +should give +\begin{displaymath} +\int_{-\infty}^\infty\frac{f(x,y)}{x^2+y^2}dx\,dy +\end{displaymath} + +Although |<math>| is said to recognise functions as maths entities, there is no +list of all the entities given in the draft. However, straightforward elements +like |<sqrt>| and |<root>| exist which work the same as their \TeX\ counterparts. +The \LaTeX\ notation for $\root n \of x$ however is \cmd{sqrt} with an +optional argument: |\sqrt[n]{x}|, not as HTML3: +\begin{Verbatim} +<math> +<root>n<of>x</root> +</math> +\end{Verbatim} +which is rather closer to \TeX. + + What the draft terms `Greek letters' are available in a similar way to +\TeX/\LaTeX\ as entity references (and includes omicron). +Some control over font styles is available through |<b>| and |<t>|. The first +emboldens, while the latter makes upright (Times?). You may even combine +the two as |<bt>|. Note that these changes apply to variables and constants, +and not to numbers, delimiters, operators and other symbols. An interesting +attribute is |class|, so that we could identify a vector as +\begin{Verbatim} +<math> +<b class=vector>a</b>=A′ +</math> +\end{Verbatim} + +Arrays or matrices are quite verbose, but broadly similar to \LaTeX. They are +introduced by |<array>|, while each row starts with |<row>| and each cell with +an |<item>|. This is in line with HTML's |<table>| model. Adapting some aspects +of \LaTeX, column definitions can be added, for example +|coldef="CCCC"|, the default, where columns are centred. The alternatives are +|R| and |L|. This is one of the few instances in HTML where case is vital. If the +attributes are separated by |+|, |-| or |=| this will propagate down the array +as a separator. For example +\begin{Verbatim} +<array coldef="R=C+C+C"> +\end{Verbatim} +Other attributes include |ldelim| and |rdelim| to specify the right and left +delimiters of the array or matrix. Unlike other instances where a name is used, +the symbol itself appears to be used in this context. + + +\subsection{Summary} +I tend to think that this makes it more difficult for someone with a \LaTeX\ +background to interpret |<math>|. When two languages are quite different, there +is rarely confusion in flipping from one to the other, but when they share +many similarities it can be frustratingly simple to converge at all the wrong +times. Note also that |<math>| uses the ISO entity names for +symbols (\cite{SandS}) rather than +the \TeX\ names. In a few cases this sows potential confusion. +The confusion which exists within the draft between \TeX\ and \LaTeX\ +is not of itself a problem, except that people coming to HTML3, being told +it is `like \LaTeX', would find some key differences. + +The main divergences are +\begin{enumerate} +\item the interpretation of space\label{space} +\item need to close most \acro{SGML} elements (e.g.\ |^| and |_|) +\item |<over>| +\item interpretation of functions (a consequence of \ref{space}) +\item entity names similar but not identical to corresponding commands +\item use of |′| +\item missing commands +\item poor support for cross-referencing +\end{enumerate} + + +\section{Future of maths in HTML} +At this point, it may be worth considering the extent to which any maths +expression can divorce the semantic component from the form on the page. Often +the way equations are portrayed can assist in their interpretation. Both tables +and maths seem to be examples where the meaning and the appearance are very +closely intertwined. There are instances where an author changes notation in +order to pursue an argument. One assumes that the essential meaning does not +change between changes in notation, and that perhaps a markup system might not +even note the change, except perhaps as an attribute. + +The draft document which forms the basis of this discussion +expired in September, 1995. Some of the new structures which +it introduced, notably tables, form part of most browsers now, +but mathematics didn't make it (although |<sub>| and |<sup>|, without +|<math>| did) (see \cite{html32}). An email from Dave Raggett +(to David Carlisle), who wrote the draft, notes that \acro{W}3\acro{C} +\begin{quote} +has set up a small working party on math, and we expect +to publish a detailed proposal by early Summer. The March'95 spec +will provide a starting point, but we may end up with something +rather different. +\end{quote} + +\noindent He goes on to state: +\begin{quotation} +\noindent The \acro{W}3\acro{C} math group has the goal to develop an +open specification +for \acro{HTML} math that: +\begin{itemize} +\item[]Is suitable for teaching, and scientific publishing. +\item[]Works with symbolic and numerical math applications +\item[]Supports filters to/from other math formats, e.g.\ \TeX +\item[]Is easy to learn and to edit by hand +\item[]Is well suited to template and other math editing techniques +\item[]Can be rendered to: +\begin{itemize} +\item[]graphical displays +\item[]speech synthesisers +\item[]plain text displays e.g.\ \acro{VT}100 emulators +\item[]print media, including braille +\end{itemize} +\item[]Supports lengthy expressions via fold\slash unfold and +line breaking with author control. +\end{itemize} +This is shaping up as the need for simple macros and declarations +that define terms etc. for use across multiple |<math>| elements, +and parsing of |PCDATA| using ``models'' referenced by |<math>| elements. +These models define how to interpret stuff at a level sufficient to +support symbolic manipulation without having to make all these +distinctions explicit in the markup itself. +\end{quotation} + +This is quite a bold extrapolation from the original specification. +Some of the suggestions seem to me to be incompatible. The inclusion of +symbolic and numerical applications is interesting and goes far +beyond the existing maths \acro{DTD}s. + +The relevant \acro{W}3\acro{C} web page (\cite{W3C}) contains a +reference to \acro{QED}, an ambitious +programme to build a `single, distributed, computerized +repository that rigorously represents all important, established +mathematical knowledge'. If this genuinely represents part of +\acro{HTML}'s solution to maths, we will have some time to wait before +a system is available. + +\section{Conclusion} +It is not clear to me at present whether \acro{HTML} will ever be rich enough +to do the sorts of things which mathematicians and physicists want to do +with maths. Of course, it could be that these are not the market at all. +I have long argued that one of \TeX's (and \LaTeX's) major problems is +that the population of users who benefit by it is small -- very much +a minority. And the expansion of the use of computers has made them +an even smaller minority. At best we are a niche market. Some of the +simpler problems are already tackled quite conveniently by word processors, +further eroding the niche. Why bother with mathematics at all? Is it +really worth the effort, compared with something sexy like |<frame>|s? + +Having said that, it appears that Public Entities in \acro{ISO}\,8879 +(\cite{SandS}) +are sufficient +to encompass most of the symbols I have seen in \LaTeX\ and \AmSTeX. The +potential is there. What makes this especially intriguing is that there +is software around like Panorama from SoftQuad which is designed to +enable any \acro{SGML} document whose \acro{DTD} is known to be +rendered on the screen. +Therefore for truly `heavy' applications, this seems a much better +way to go. In fact, I would see it as an altogether better way to go. +Browsers which could examine the |Doctype|, find it on an appropriate +server and then render it would be much more flexible, and enable us to use +existing \acro{SGML} documents easily on the Web. \acro{HTML} would +simply be a lightweight +\acro{DTD} used because it had a lower overhead. + +On the other hand, browsers like Netscape Navigator +are becoming larger, are starting +to include `plug-ins' and be Acrobat-aware. +If \acro{HTML}3 version 2, as outlined by Raggett does support filters +(hopefully to \LaTeX\ rather than \TeX), as well as symbolic +manipulation (Mathematica, Maple, etc.), it could be a very powerful +tool. +However, the latest information available through the World Wide Web Consortium +does not encourage belief that this is much more than a dream. + +\begin{thebibliography}{} + +\bibitem[ANSI, 1994]{AAP} +\acro{ANSI}/\acro{NISO}, (1994). +\newblock Information and documentation -- Electronic +manuscript preparation and markup. +\newblock Technical Report, ISO 12083:1994 {\acro{ANSI}/\acro{NISO}}. + +\bibitem[Bryan, 1988]{Bryan} +Bryan, M. (1988). +\newblock {\em {SGML} an author's guide}. +\newblock Addison-Wesley. + +\bibitem[Chamberlin, 1988]{Chamberlin} +Chamberlin, D.~D. (1988). +\newblock An adaptation of dataflow methods for \acro{WYSIWYG} +document processing. +\newblock In {\em ACM Conference on Document Processing Systems}. + +\bibitem[Euromath consortium, 1996a]{Euromath} +Euromath consortium (1996a). +\newblock Euromath system. +\newblock \url{www.math.ethz.ch/~shared/emb/GeneralFiles/Euromath.html} + +\bibitem[Flynn, 1995]{Flynn} +Flynn, P. (1995). +\newblock \acro{HTML} and {\TeX}: making them sweat. +\newblock {\em {\BV}}, 5(2):7--11. + +\bibitem[Goossens and Saarela, 1995a]{Goossens2} +Goossens, M. and Saarela, J. (1995a). +\newblock {\TeX}\ to \acro{HTML} and back. +\newblock {\em {\TUB}}, 16(2):174--214. + +\bibitem[Goossens and Saarela, 1995b]{Goossens} +Goossens, M. and Saarela, J. (1995b). +\newblock A practical introduction to \acro{SGML}. +\newblock {\em {\TUB}}, 16(2):103--150. + +\bibitem[Grif, 1996]{GrifSA} +Grif (1996). +\newblock {GRIF} {S}. {A}. -- \acro{SGML} and \acro{HTML} solutions. +\newblock \url{www.grif.fr} + +\bibitem[\acro{ISO}, 1988]{ISO9573} +\acro{ISO} (1988). +\newblock Information processing -- \acro{SGML} support +facilities -- Techniques for using \acro{SGML}. +\newblock Technical Report, \acro{ISO}/\acro{IEC} \acro{TR} 9573:1988, +\acro{ISO}. + +\bibitem[Poppelier et al., 1992]{PvHR} +Poppelier, N.~A.~F.~M., van~Herwijnen, E. and Rowley, C.~A. (1992). +\newblock Standard \acro{DTD}s and scientific publishing. +\newblock {\em {EPSIG} {N}ews}, (3):10--19. + +\bibitem[Quint and Vatton, 1986]{Grif} +Quint, V. and Vatton, I. (1986). +\newblock Grif: an interactive system for structured document manipulation. +\newblock In van Vliet, J.~C., editor, {\em Text processing and Document + Manipulation}, pages 200--213. Cambridge {U}niversity {P}ress. + +\bibitem[Raggett, 1995]{math} +Raggett, D. (1995). +\newblock \url{www.w3.org/pub/WWW/MarkUp/html3/} +\newblock Expired \acro{HTML3} draft. + +\bibitem[Rahtz, 1995]{seb} +Rahtz, S. P.~Q. (1995). +\newblock Another look at {\LaTeX}\ to \acro{SGML} conversion. +\newblock {\em {\TUB}}, 16(3):315--324. + +\bibitem[Smith and Stutely, 1988]{SandS} +Smith, J. and Stutely, R. (1988). +\newblock {\em {SGML} the user's guide to {ISO}\,8879}. +\newblock Ellis Horwood. + +\bibitem[van Herwijnen, 1990]{EvH} +van Herwijnen, E. (1990). +\newblock {\em Practical {SGML}}. +\newblock Kluwer Academic Publishers. + +\bibitem[\acro{W}3\acro{C}, 1996b]{html32} +\acro{W}3\acro{C} (1996b). +\newblock \acro{HTML} 3.2 specification. +\newblock \url{www.w3.org/pub/WWW/MarkUp/Wilbur/} + +\bibitem[\acro{W}3\acro{C}, 1996c]{W3C} +\acro{W}3\acro{C} (1996c). +\newblock Math markup in \acro{HTML}. +\newblock \url{www.w3.org/hypertext/WWW/MarkUp/Math/} + +\end{thebibliography} +\end{Article} + |