diff options
Diffstat (limited to 'Master/texmf-dist/doc/fonts/junicode/Record_Interpreter_notes.tex')
-rw-r--r-- | Master/texmf-dist/doc/fonts/junicode/Record_Interpreter_notes.tex | 293 |
1 files changed, 293 insertions, 0 deletions
diff --git a/Master/texmf-dist/doc/fonts/junicode/Record_Interpreter_notes.tex b/Master/texmf-dist/doc/fonts/junicode/Record_Interpreter_notes.tex new file mode 100644 index 00000000000..9d10990d1a9 --- /dev/null +++ b/Master/texmf-dist/doc/fonts/junicode/Record_Interpreter_notes.tex @@ -0,0 +1,293 @@ +\chapter{Transcribing records} +%\fancyhead[CE]{\scshape\color{myRed} {\addfontfeatures{Numbers=OldStyle}\thepage}\hspace{10pt}transcribing records} + +This chapter provides guidance for persons transcribing records (laws and other +public documents) in the style of Charles Trice Martin’s +\textit{The Record Interpreter}, \textit{Statutes of the Realm}, and similar guides +and editions. +Unlike most editions of early texts, these retain (or recommend retaining) the capitalization, +punctuation, and abbreviations of their manuscript sources. + +\section{A preliminary note on transcription} +Here are a few observations, based on a long career as a scholarly editor of medieval and eighteenth-century texts. + +Before embarking on the task of transcribing an old document, ask yourself what value you want to add to the document as +it already exists, because different kinds of transcription add different kinds of value. The kind of transcription +that adds the least is that which aims at the exact \textit{visual} reproduction of a document. A transcript is not a +facsimile: it needs to do something that a photograph can't do. + +Converting a document from visual image to Unicode-encoded text adds a good bit of value all by itself, but only if done +with due regard for the semantics of Unicode characters. Every Unicode character has a meaning, and that meaning is a +help to readers. Using the wrong character is a hinderance to readers, even it if \textit{looks} right. + +For example, in transcribing a Middle English text, you may decide that the Unicode ezh (\char"0292,~\unic{U+0292}) looks more like +the yogh in your source than the Unicode yogh (\kern+1.5pt\char"021D,~\unic{U+021D}) and therefore decide to use it for yogh. But the ezh is +not a yogh! It is a character in the International Phonetic Alphabet and a letter in the alphabets of several minor +languages---but not a letter in the Middle English language. If you use it where the yogh is called for, it will make +your text less accessible and less searchable. Indexing, concordance and bibliographical programs may be misled by it; +screen readers will misinterpret it. To solve one problem (that of visual representation), you may well have introduced +a host of far more serious problems. + +Fortunately, Junicode offers a solution for this particular problem. The OpenType feature +\textSourceText{cv63}\index{cv63} substitutes for +the yogh a character that \textit{looks} like the ezh but is semantically a yogh and therefore will be +handled correctly by applications. But neither Junicode nor any other font can solve every problem of this kind. +Sometimes you will have to call to mind the important principle stated above: \textit{A transcript is not a +facsimile}. It is much more important that it should have the same \textit{meaning} as the original than +that it should have the same \textit{look}. + +This chapter concerns the transcription of texts in Latin (and to some extent, other archaic languages, e.g. Old and +Middle English, Old French). It is long-standing custom, when transcribing certain kinds of documents, to retain marks +of abbreviation---for example, the \textex{\hlig{\char"0A753p\char"0363}} you may find in a manuscript or printed edition representing +the word \textit{propterea}. This is okay---and Junicode can help with the task. But when dealing with the +abbreviations, punctuation, and diacritics of an old text it is more important than ever that you use semantically +correct characters for your transcription, as this will help readers who already face significant challenges. + +For example, the abbreviation \textex{\hlig{\char"0A753p\char"0363}} as printed here consists of an underlying sequence of Unicode +characters: \textex{\char"0A753} (\unic{U+A753}, the common abbreviation for \textit{pro}) + \textex{p} + \textex{\char"25CC\char"0363} +(\unic{U+0363}, the combining small \textit{a}). The OpenType feature \textSourceText{hlig}\index{hlig} + (Historical Ligatures) has been +applied to this sequence, changing its appearance but not its underlying value. That underlying value is intelligible +to computer applications in the sense that they can recognize each character. + +This doesn't mean, though, that computer programs can correctly interpret \textex{\hlig{\char"0A753p\char"0363}} as +\textit{propterea}. Many (probably \textit{most}) Latin abbreviations are ambiguous: this one, +for example, can mean \textit{propterea} or \textit{propria}. Some abbreviations (most +notoriously \textex{\char"25CC\char"035B} \unic{U+035B}) can mean many things, depending on context. It takes a human being with a +knowledge of Latin to interpret them correctly. + +So another way you can add value in your transcript is by interpreting abbreviations like +\textex{\hlig{\char"0A753p\char"0363}} and +supplying expansions of them. Fortunately, systems for representing texts often offer ways to handle this task +gracefully. For example, in a TEI (Text Encoding Initiative) text, you would use this construction:\\[1ex] + +\noindent\verb!<choice>!\newline +\verb! <abbr rend="hlig">!\textrm{\hlig{\char"0A753p\char"0363}}\verb!</abbr>!\newline +\verb! <expan>propterea</expan>!\newline +\verb!</choice>!\\[1ex] + +\noindent This kind of structure can be approximated in HTML, with supporting CSS and scripting to allow readers to choose between +a ``diplomatic'' version, with unexpanded abbreviations, and a ``reading'' version, with expanded abbreviations and +perhaps other amenities, such as modern punctuation and capitalization. + +There are other ways to add value to a transcript---for example, by correcting errors, annotating the content, or +writing textual notes. Each of these operations takes your transcript farther from the facsimile and closer to the +edition. + +\section[Common combining marks]{Common combining marks} +A \textbf{combining mark} is a character that combines with another character (called the \textbf{base}) to form a +character with accent (e.g. \textex{\'e}) or an abbreviation (e.g. \textex{\cvd[1]{81}{p\char"035B}} for \textit{prae}). +Unicode and the Medieval Unicode Font Initiative (MUFI) offer code points for many precomposed combinations of base + +combining mark, but it is also possible to place any mark over any base character by entering first the base and then +the combining mark. It is also possible to place a combining mark over another combining mark. For example, to produce +\textex{\cvd[32]{84}{q\char"0363\char"0304}}, enter this sequence: q (\unic{U+0071}) + \unic{U+0363} + \unic{U+0304}. + +Junicode 2 contains many variants of combining marks: for example the curly zigzag \cvd[1]{81}{\char"25CC\char"035B} is a variant of +Unicode's angular zigzag {\char"25CC\char"035B} (\unic{U+035B}), produced by applying the OpenType feature +\textSourceText{cv81[2]}\index{cv81} to +\textbf{both the base character and the combining mark}. Sometimes the combination of base + combining mark + OpenType +feature will not produce the desired effect. When this happens, place \unic{U+034F} (a special invisible combining mark, +included in Unicode for exactly this purpose) between the base and the (visible) mark.\\ + +\noindent\secletter{a.}\ \ For a straight stroke over any letter, use the \textsc{combining macron} (\unic{U+0304}): +\begin{quote} +\=onis \textit{omnis}; o\={m}is \textit{omnis}; d\=apna \textit{dampna}; dam\={p}a +\textit{dampna}. +\end{quote} +The combining macron can also be applied above superscripts and combining marks. Apply the OpenType feature +\textSourceText{cv84[33]}\index{cv84} for a narrower macron: +\begin{quote} +\cvd[32]{84}{antiqua\char"034F\char"0304} \textit{antiquam}; \cvd[32]{84}{q\char"0363\char"0304} \textit{quam}. +\end{quote} +For the superscript \textit{a}, use the OpenType feature \textSourceText{sups}\index{sups} (see r. below).\\[1ex] + +\noindent\secletter{b.}\ \ For a straight stroke through a tall letter, use the \textsc{combining short stroke overlay} (\unic{U+0335}): \textex{f\char"0335\ d\char"0335\ l\char"0335}. But Unicode also has precomposed versions of +\textex{d}, \textex{l} and other characters \mbox{with} stroke, e.g. +\textex{{\dj}} (\unic{U+0111}), \textex{\char"019A} (\unic{U+019A}).\\[1ex] + +\noindent\secletter{c.}\ \ For \textex{\~{}} above any character, use the \textsc{combining tilde} (\unic{U+0303}): +\begin{quote} +\~a \textit{ac}, \textit{apud}; \~a \textit{alias}.\newline +d\~ns \textit{dominus}; ca\~{r}ina \textit{carmina}; \cvd[4]{12}{f\~{c}is} \textit{factis}.\newline +p\~oita \textit{posita}. +\end{quote} + +\noindent\secletter{d.}\ \ For \textex{\~{}} through a vertical stroke, use the \textsc{tilde overlay} (\unic{U+0334}): +\textex{l\char"0334\ d\char"0334} (\unic{U+0303} would be positioned above the letter, e.g. +\textex{\~{l}}, \textex{\~{d}}). For the ligatures +\textex{l̴l̴}, \textex{b\char"0334b\char"0334}, and \textex{f\char"0334f\char"0334}, type the +sequence for \textex{l\char"0334}, etc. twice.\\[1ex] + +\noindent\secletter{e.}\ \ For the tilde positioned above two letters, use \textsc{combining double tilde} +(\unic{U+0360}) between the letters. It is +automatically repositioned to clear tall characters: c\char"0360o t\char"0360o d\char"0360o o\char"0360l. The same is true of +\textsc{double breve} (\unic{U+035D}) c\char"035Do d\char"035Do, \textsc{double macron} (\unic{U+035E}) c\char"035Eo d\char"035Eo, +\textsc{double inverted breve} (\unic{U+0361}) c\char"0361o d\char"0361o, and \textsc{double circumflex} (\unic{U+1DCD}) +c\char"1DCDo d\char"1DCDo.\\[1ex] + +\noindent\secletter{f.}\ \ The figure used to represent \textit{er\textup{ }}(and other similar combinations) is a common medieval +abbreviation which takes many forms. The semantically correct Unicode character is the \textsc{combining +zigzag} (\char"25CC\char"035B, \unic{U+035B}), but the best match in Junicode 2 for the figure as it appears in the +\textit{Record Interpreter} and the \textit{Statutes} is a gothic variant of this, which MUFI +encodes as \unic{U+F1C8} (the curly form zigzag). However, because for technical reasons many applications will not position +the MUFI character correctly over the base, that code point should be avoided. The best way to access this variant is +to apply \textSourceText{cv81[2]} to \unic{U+035B}, as here: +\begin{quote} +\cvd[1]{81}{deb\char"035Be \textit{debere}; int\char"035B\ \textit{inter}; f\char"035B\hspace{0.2em}r\=u \textit{ferrum}; gn\char"035Bo +\textit{generatio}; p\char"035B; \textit{prae}; seru\char"035Be \textit{servire}}. +\end{quote} +The curly form of the combining zigzag may be attached to any letter, and it may change shape depending on the letter it +is attached to (including caps, for which use the \textSourceText{case}\index{case} feature, and small caps: +\textex{\cvd[1]{81}{A\char"035B B\char"035B \textsc{c\char"035B +d\char"035B\hspace{0.2em}}}}).\\[1ex] + +\noindent\secletter{g.}\ \ All letters a--z, and several others too, have combining forms. +You may access these via their code points, when they are standard Unicode, via the +\textSourceText{\textsc{cv84}} feature, or via +Junicode's special entity references. For details, see \hyperlink{ss10}{4.10.3, Character Entities +for Combining Marks}. +\begin{quote} +q\char"0366\ \textit{quo}; q\char"0365\ \textit{qui}; quatt\&\_orr; \textit{quattuor}. +\end{quote} +\section{Spacing characters} +\secletter{h.}\ \ The symbol for \textit{is}, \textit{es} and a number of other abbreviations is the +\textsc{is-sign} (\unic{U+A76D}): +\begin{quote} +for\char"0A76D\ \textit{foris}; o\={m}\char"0A76D\ \textit{omnes}; \char"0A76Ft\char"0A76D\ +\textit{competentes}; inf\char"0A76D\ \textit{infortunium}. +\end{quote} +This character will sometimes ligature with the preceding letter. The italic version differs from the roman +stylistically (\ \textex{\textit{for\char"0A76D\ o\={m}\char"0A76D\ \char"0A76Ft\char"0A76D\ inf\char"0A76D}}), but it will be +intelligible to informed readers.\\[1ex] + +\noindent\secletter{i.}\ \ There are two characters for \textit{{}-us} in Unicode: +\textsc{spacing us} \unic{U+A770} (do not +confuse this with \textsc{con} \unic{U+A76F}) and \textsc{combining us} \unic{U+1DD2}. The \textit{Record +Interpreter} and \textit{Statutes} appear to use only the spacing character: +\begin{quote} +\cvd[1]{81}{i\~{p}i\char"0A770\ \textit{ipsius}; u\char"035Bs\char"0A770\ \textit{uersus}; p\char"0A770tea +\textit{postea}; p\char"0A770\ \textit{post}.} +\end{quote} +\noindent\secletter{j.}\ \ The three-like sign is the \textsc{et sign} (\char"25CC\char"0A76B, \unic{U+A76B}, also used for +\textit{us} in the Latin ending \textit{-ibus}). Do not use the numeral three (\textex{3}) +or the Middle English yogh (\textex{\char"021D}, \unic{U+021D}): +\begin{quote} +quib\char"0A76B\ \textit{quibus}; lic\char"0A76B\ \textit{licet}; s\char"0A76B\ \textit{sed}. +\end{quote} +\noindent\secletter{k.}\ \ For \textit{{}-rum} the Unicode \textsc{rum rotunda} (\unic{U+A75D}) is like the one in MUFI/Junicode. +The one in the \textit{Record Interpreter} and \textit{Statutes} is a late stylized version of +this. Use \unic{U+A75D} and apply OpenType feature \textsc{\textSourceText{cv80}}\index{cv80} to obtain the correct shape: +\begin{quote} +\cvd{80}{a\~{i}a\char"0A75D\ \textit{animarum}; co\char"0A75Dpere \textit{corrumpere}; beato\char"0A75D\ +\textit{beatorum}}. +\end{quote} +\noindent\secletter{l.}\ \ For \textit{cum}, \textit{con}, etc. use \textsc{small letter con} (\unic{U+A76F}): +\begin{quote} +\char"0A76Fputus \textit{computus}; \char"0A76Fa \textit{contra}; \char"0A76Fnouit \textit{cognouit}. +\end{quote} +\noindent\secletter{m.}\ \ For \textit{per} (or sometimes \textit{par} and other similar sequences), use +\textsc{p with stroke} \unic{U+A751}: +\begin{quote} +\char"0A751s\={o}a \textit{persona}; \char"0A76F\char"0A751et \textit{comparet}. +\end{quote} +\noindent\secletter{n.}\ \ For \textit{pro,} use \textsc{p with flourish} \unic{U+A753}: +\begin{quote} +\char"0A753ceres \textit{proceres}. +\end{quote} +\noindent\secletter{o.}\ \ For \textit{prae}, \textit{pr{\ae}}, \textit{pre}, there is no separate character; +use a variant of the \textsc{zigzag} (f. above) with \textex{p}: +\begin{quote} +\cvd[1]{81}{p\char"035Bs\~{e}s \textit{praesens}}. +\end{quote} +\noindent\secletter{p.}\ \ For \textex{q} with stroke through the descender, there are two Unicode points: \unic{U+A757} for a +straight stroke, and \unic{U+A759} for a diagonal stroke (the \textit{Record Interpreter} appears to use only the +former, and neither is listed among the \textit{Statutes} abbreviations): +\begin{quote} +\char"0A757\ \textit{quod}; \char"0A757d \textit{quid}; \char"0A757b\char"0A76B\ \textit{quibus}. +\end{quote} +\noindent\secletter{q.}\ \ For \textit{quae}, \textit{que}, use \textex{q} +followed by \textsc{et} (\unic{U+A76B}) with or without \textSourceText{hlig}\index{hlig}: \textex{q\char"0A76B} +\textex{\hlig{q\char"0A76B}}. For the semicolon-like \textsc{et} sign (\textex{\cvd{83}{q\char"0A76B}}), use +\textSourceText{cv83[1]}\index{cv83}; for the subscripted version (which can also form a ligature via \textSourceText{hlig}), +use \textSourceText{cv83[2]}: +\textex{\cvd[1]{83}{q\char"0A76B\ \hlig{q\char"0A76B}}}.\\[1ex] + +\noindent\secletter{r.}\ \ All of the letters a-z are available in superscript form. Access with the \textex{sups} +OpenType feature: +\begin{quote} +q\sups{o}s \textit{quos}; c\sups{i}lo \textit{circulo}; cap\sups{i} \textit{capituli}. +\end{quote} +The basic Latin letters a--z have anchors that allow you to position combining marks over them (see a. above)\\[1ex] + +\noindent\secletter{s.}\ \ Tironian \textsc{et} sign \char"204A\ \unic{U+204A}, cap \char"2E52\ \unic{U+2E52}. +With \textSourceText{cv69[1]} \cvd{69}{\char"204A\char"2E52}; with +\textSourceText{cv69[2]}\index{cv69} \cvd[1]{69}{\char"204A\char"2E52}.\\[1ex] + +\noindent\secletter{t.}\ \ For \textit{est}, use \textex{\char"223B} \unic{U+223B} \textsc{homothetic}. Use of a mathematical sign for this +purpose is not ideal, but Unicode offers no better solution.\\[1ex] + +\noindent\secletter{u.}\ \ For \textit{tz} (Old French), use \textex{\char"01B6} \unic{U+01B6} \textsc{z with stroke}.\\[1ex] + +\noindent\secletter{v.}\ \ For an abbreviation for \textit{Rex}, +use \textex{\textrecipe} \unic{U+211E} or \textex{\char"211F} \unic{U+211F}.\\[1ex] + +\noindent\secletter{w.}\ \ At least one edition uses a spacing version of the \textsc{combining zigzag} +(\textbf{f.} above). +Neither Unicode nor MUFI has a matching character: with Junicode, apply \textSourceText{cv67}\index{cv67} to the spacing +\textsc{macron} (\unic{U+00AF}): \textex{\cvd{67}{\char"25CC\char"00AF}}. + +\section[Other formatting]{Other formatting} +\secletter{x.}\ \ For underdotted text, use Stylistic Set 7, Underdotted. For letters that lack an underdotted form, use \unic{U+0323} \textsc{combining dot below}. + +\section[On the web]{On the web} +Because Junicode is a very large font, web pages should use a subsetted version to speed loading +(see \hyperlink{OnTheWeb}{Chapter 9, Junicode on the Web}, for instructions). +The variable version of the font is better for web use than the +static fonts, since one variable font file can do the work of many static font files. + +All major web browsers (Firefox, Chrome, Safari, Edge) are capable of accessing all of Junicode's characters via +OpenType features, use of which promotes accessibility and searchability. When building a web page, study which +features will be needed and write them into the appropriate element or class definition of the page's CSS style sheet. +For example, if you use the curly form of the zigzag (\unic{U+035B}) anywhere, you are likely to want it everywhere, and so it +should be included in the CSS styling for the <body> element: +\begin{verbatim} +body { + font-family: Junicode; + font-feature-settings: "cv81" 2; +} +\end{verbatim} + +\noindent But the \textSourceText{hlig}\index{hlig} feature, if applied to the whole text, +will produce many unwanted effects, so it should be +included in a class definition to be used in a <span> applied just to the target sequence: +\begin{verbatim} +.que { + font-feature-settings: "hlig" on; +} +filio<span class="que">qꝫ</span> +\end{verbatim} + +\noindent The illustrations here use the low-level CSS font-feature-settings property. +There are higher-level properties for some +OpenType features, but as these are not (yet) universally supported by browsers, and some implementations are buggy, it +is best to stick with font-feature-settings for now. + +For the purposes addressed in this document, the font-feature-settings for the <body> element +should probably be as follows: +\begin{verbatim} +font-feature-settings: 'cv69' 2, 'cv80' 1, 'cv81' 2; +\end{verbatim} +\noindent And the following classes should be defined: +\begin{verbatim} +.super { + font-feature-settings: 'sups' on, 'cv84' 39; +} + +.que { + font-feature-settings: 'hlig' on; +} + +.deleted { + font-feature-settings: 'ss07' on; +} +\end{verbatim} |