Summary. Headings and sub-headings should be added to a transcript, and it should be framed in generous margins. A bulleted list is among the most effective typographic devices to present the core section of a deed. Finer detail that needs highlighting can also be displayed with bullets. Bigger documents should be structured into DIV rather than P elements. Dynamic styling is a way to hide dull text till it is needed. Run-in headings enable a more compact view.
For hundreds of years, lawyers and registrars have been writing deeds and employing only the most primitive layout for them. On the one hand, a great deal of redundant "aforesaids" and formulaic material had to be included in these documents because of the lack of what legal scholars call rechtssicherheit. On the other hand, deeds were written in a cramped, abbreviated style with neither spacing nor headings because paper and the services of clerks were costly.
The reader therefore has a dual problem: learning to skate over the junk text, while divining important meaning that may be concealed behind just a squiggle of the pen.
The introduction of pre-printed forms meant some typographical devices such as italics could be employed to make the old formulas clearer for the 19th-century reader. In the 20th century, the advent of the typewriter brought paragraphs, numbering and all-capitals highlighting to legal deeds. The IBM "golfball" electric typewriter brought bold fonts.
But it is only on the computer screen, where white space, colour and enlarged headings cost nothing more than a little alteration to the styling code, that we can make such documents much more readable.
Provided by the editor, headings will be the key signposts guiding the reader through the document. To take full advantage of styling, all the documents in a collection should be headed up in the same way, and a general "anatomy" should be devised to describe every document so that all its parts have a place within the whole. This is explained more fully below.
As examples of deeds in this guide we will use English wills, but the principles apply equally well to testaments from other countries and indeed to most other types of legal deed.
Even if you are only transcribing a single document, choose headings that would potentially help a user find their way among a large number of documents: your collection may grow later.
For a collection of wills from disparate sources, five elements are likely to be important enough to go into the main heading: the name of the testator, the person's abode, the document type (e.g. will), the validation date and the jurisdiction. These headings, as in the example, should use a modern spelling and dating system so they can be re-purposed for use in a table of contents or searchable index. Personal names should be converted from Latin to vernacular forms (for English names, use Judith Werner's names list, and see the related links there).
It should be plain to readers from the larger font size and colour that any headings provided by the editor in the online version were not present in the original document.
Sub-headings, which may have to be enclosed in square brackets to make clear they are not original material, should pick out whatever structural features recur in different examples of the chosen type of manuscript. Most wills, for example, contain at the very least a signature and are witnessed as demonstrated in an example where a reader who does not know German will be grateful for the guidance the sub-headings provide.
In English-speaking lands, the testator must also appoint some person as the executor and it is usually appropriate to label this section too. Note how three such sub-headings bring a fresh sense of order to an earlier example.
These sub-headings not only help to break up the text, marking it off into sections, but can also provide sorting criteria for small computer programs that operate within the user's browser program (client-side scripts). Documents marked up this way can easily yield up information to indexers as well.
That is why it is useful to always include the date of signature in the Signature heading, even though the date may be stated elsewhere in the document in text form, such as in the preamble. I have used the International Standards Organization (ISO) format for dates (see explanations by Meyer and Kuhn) not only because this sidesteps a source of conflict between American and British readers, but also because the year-first order allows simple left-to-right sorting.
The layout recommended here for wills is a simple one that allows several transcripts to be presented on one web page. A border or background— I chose a grey underlay— helps to mark off each document. This has been demonstrated in the examples so far. All that is needed is to enclose the entire will between <DIV> tags, to then give this DIV a class name, and to finally prescribe it some styles, for example:
.deed {
background-color: #dddddd;
margin: 0px 15% 10px 15%;
padding: 1px 2% 15px 3%;}
Remember that the CSS shorthand for multiple margins and padding usually consists of four items starting from the top, like a clock, and sweeping around, like a clock hand, in the order right, bottom, left. So in the above example, the block has these margins: top 0px, right 15%, bottom 10px, left 15%.
Generous white space to the left and right sharply improves legibility compared to both handwritten and typewritten documents, where line lengths are generally too great for comfortable reading. Percentage units for the left and right white space ensure that the chosen structure stays in proportion horizontally when the computer window screen is resized by a user.
One of the greatest advances in 20th century macro-typography was the spread of the bulleted list. The organizational principle has been familiar to lawyers for hundreds of years, with each recurring statement being prefixed by the Latin term item meaning "moreover". But it is only today that we have a graphic way of displaying itemized lists so that their structure is plain at first glance. This allows us to "retrofit" old documents. Look at an example from the year 1575: a typography based on headings and a bulleted-list greatly improves its legibility.
Note that nothing other than the headings has been added to the original, painstaking transcript in which Peter Alefounder used a variety of HTML entities (symbols beyond the 26-letter alphabet) to represent the handwritten document as faithfully as possible. The bullets have not been inserted into the source text: they are automatically created by the browser program every time the page is loaded. The only change has been to rearrange the text into aligned blocks on the page for greater clarity: every "Item..." was enclosed in <LI> tags; the legacies section as a whole was defined as a <UL> element.
Here is a possible set of styles for such a list. The most important features are (1) to include both left margin and left padding values to satisfy the different needs of MS Internet Explorer and the Mozilla/Netscape browsers, and (2) a specification of a list-style-type that can be carried over to sub-lists.
margin: 0px 0px 0px 3em;
padding-left: 3em;
list-style-position: inside;
list-style-type: disc;}
By the 19th century, it was no longer usual among English legal draftsmen to iterate the legacies with the item notation, but the 21st-century typographer can nonetheless unite wills from different historical periods by treating the core part of the will, the disposition of the estate, as a list, or <UL> element in HTML.
In 20th-century wills, this core list is often obscured by the fact that the document is presented as a sequence of numbered paragraphs, but the item-oriented typography can be imposed over this without falsifying the original. Here is the fictitious modern-style example in which the bullet draws the eye to the bequest. If only one bequest is made, even this is, for mark-up purposes, a list: it consists of just one item.
The job of the information architect is to devise an anatomy for such documents that leads to better screen presentation than the paper originals could achieve. No one anatomy is the "best" or "right": which one you use depends on your purpose.
We can illustrate this by considering the different methods of human anatomy. An artist is likely to use an anatomical system that divides the human figure into head, trunk and limbs. A dermatologist's anatomy might divide the body into closed spaces and surfaces (which could be further subdivided into skin and mucous membrane). The two anatomies have nothing in common, but each serve their own purpose.
The basic anatomy of every European will is fairly plain: (1) the testator is identified, (2) listed property is distributed and (3) the document is validated with a signature. Here is a more detailed anatomy, specifically for an official English record of probate, of which the will is a sub-part:
These segments are easy enough to represent as divisions in the HTML structure of a document, and I have already suggested above how some of these parts should be marked off by sub-headings.
A more finely detailed anatomy that is true for most examples is difficult to discern in an English will. This is not only because a will is free text, but also because these documents have many different uses for a modern researcher. Wills are variously of interest to social historians, economists, anthropologists, genealogists and of course lawyers, and ways of using them are for example discussed in the journal History and Computing (volume 7 issue 3 (1995)) (Order)
Since the uses vary so much, there will probably never be a standard way of (re)structuring wills. One group of researchers is interested in the economic values, another in the historical persons, and the legal scholars in the legal mechanisms they contain. Numerically, the largest group of researchers using wills are probably those reconstructing families, so I will propose an anatomy and a macro-typographical design tailored to that area of study.
The principal focus when using a will for family reconstruction is the set of names and family relationships in the document. Often more than one person is mentioned in connection with a single bequest. The gift may go to two or more persons jointly, or to one person for life with reversion to another, as in the first legacy of the Robert Tyler example seen above.
Sometimes the persons mentioned are not legatees at all: the will of Alice Alefounder († 1801) made a gift of a copperplate picture showing a beneficiary "with his cousin Anna Maria Jane Alefounder", although the latter woman is not a legatee.
This is quite a challenge in marking up, because all these personal names are embedded in the flow of text. The natural divisions in the text are the sentences. So what kind of signals do we use to make the personal names more visible through the typography? One way is to simply highlight the supplementary names with a different text colour, or with a marker-pen-style background colour, like this:
to my nephew Stephen Hempsted my silver tankard to my grandson George Alefounder four large silver tablespoons, my metal watch and the copperplate picture of him with his cousin Anna Maria Jane Alefounder Also I give to my said Grandson George Alefounder all my household goods ...
A principal disadvantage with this is that the colouring may not show up on many printers, where the drivers are set to suppress background colour.
An alternative is to interrupt the flow of the sentence every time a new name comes up, and set that name indented and bulleted. This helps the eye skip quickly through the salient points of the text, for example:
This is desirable because not only is it easier to read, but the subsection (about Anna Maria Jane) could also be given a documentary ID and quoted entire by other software functions.
The convenient structure in HTML for creating these break points is the <UL>. Obviously these asides in the will are only snippets of text, not true sub-lists in the semantic sense, but it is legitimate to adapt HTML's limited family of elements to other uses. The <UL> elements should be given a class name that makes it easy to convert them to some more fitting XML element at a later date.
The <UL> (usually consisting of only one item) should be nested within the bequest-level, next-higher <LI> as in this example:
<LI> Stephen </LI>
<LI> George <UL> <LI> Anna Maria Jane </LI> </UL> </LI>
<LI> George </LI>
Notice that the subcategory is best slotted inside the list item, not in the "empty space" between two list items.
One peculiarity of HTML that frequently causes difficulties in both information design and macro-typography is that lists cannot be nested inside paragraphs, despite the fact that many lists occur in mid-sentence. This situation was remedied in 2002 with XHTML 2, but that will not be a usable mark-up standard until it is embraced by all the main browsers.
The traditional way of coping with this was to chop a sentence up into multiple <P> elements like this:
<P>Take a bowl and mix</P>
<LI><UL>two eggs</LI>
<LI>one cup of flour</LI></UL>
<P>with a wooden spoon.</P>
That is structural nonsense. For will transcripts using the HTML 4 and XHTML 1 standards it will generally be more convenient to dispense with the <P> elements altogether. Divide the whole will text into ad-hoc <DIV> sections and fit the headings, plain text and lists inside these.
In the example, each subsidiary <DIV> has the class name "deedsection". The list of bequests is a subset of one of these <DIV> sections, and begins directly, without any top margin. There is not a single <P> in the entire transcript.
So far the discussion here has covered complete, word-for-word transcriptions of old documents. The reality however is that most online wills collections are the work of teams without consistent standards, are usually incomplete and mostly consist of a mixture of brief outlines, longer summaries and only the occasional full transcript. Kathie Weigel's remarkable collection of Cornish wills consists mainly of abstracts. Nick Hidden's collection of 1,000 Berkshire wills is a mixture of long and short versions (and, by the way, badly needs a typographical makeover). There is generally little motivation to provide complete transcripts, especially when it comes to wordy and legally complex 19th century wills.
Abstracts greatly simplify the task of publishing any deeds collection. Eve McLaughlin's booklet Wills before 1858 (Solihull, FFHS, 1995) suggests this comprehensive list of essential details to be noted from English wills:
Nick Hidden suggests that at the very minimum, an abstract should contain:
Not only are such abstracts quicker for a transcriber to write. They are also easier for an end-user to read. And as we have already noted, a good transcription should not only expand what has been unduly abbreviated: it should also compress down to what the modern reader really wants to know. So why not offer readers a first view of all the deeds in the collection as abstracts rather than full text? And wouldn't it be even better if only a single source file had to be put on the Internet, but could be visible in two ways, expanded and collapsed!
Collapsed and expanded views of the files and directories on your computer are one of the most powerful features of the "graphic user interface" (the way a Windows or Mac computer makes information visual). Collapse-expand is an example of a pattern of interaction between the user and the information that guru Jenifer Tidwell calls "optional detail on demand". It's as easy as switching on and switching off a light: one click and you have all the detail. Click back and you have just the essentials.
Miraculous as it may sound, it is quite easy to make a document expand and collapse in a visual browser. The dull two-thirds of a document can be concealed and only the essential one-third displayed until the reader says, "I want to see it all now."
The secret is a technology called style-sheet switching and requires just a little snippet of code which I will share with you below. The only hard work is the initial structuring of the information: the editor must go through the text and choose the essential items in the document worth abstracting, and mark off all the rest.
Here is how this is done. Sometimes it will only be a few words or phrases of waffle in mid-sentence that are non-essential and can be hidden in the collapsed view. In the case of a will, we may only want to see the names and nothing else. The unwanted, so-called inline sections of the text should be enclosed in <SPAN> tags. The span can be given some kind of descriptive class name, for example "vanish". Here is an example:
<li class="person"><span class="vanish">And I give to </span>my Nephew Stephen Hempsted of Market Ilstey in the County of Berks Surgeon<span class="vanish"> my Silver Tankard (being Family Plate)</span></li>
Now we can define what happens to any text of class "vanish". CSS allows us to style it with the display attribute set to "none". This means it will not appear on the web page. Non-display means all the other material closes up to fill the gap as if the non-displayed text had never existed. And that is precisely what the abstract view of the transcript is. The style code looks like this:
.vanish {display: none;}
It is not just inline sections that can be hidden. Entire block-level elements can be set to non-display too: the browser can be forced to not display a list item, including its bullet, if that element is placed in the "vanish" class. Suppose for example that we have the <LI> elements marked up like this:
<LI class="person">I give to my said Grandson George Alefounder all the rest and residue of my Personal Estate</LI>
To remove any of the LI elements from the flow, simply add another class after a space, like this:
<LI class="person vanish">I give to my said Grandson George Alefounder all the rest and residue of my Personal Estate</LI>
Revealing the vanished text is also an easy matter: the trick is to create a special-purpose style sheet to control the concealing/expanding. "Sheet" sounds portentous, but an extra line of code in the head area of a document fits the bill just as well. Just tack this additional <STYLE> element on below any existing style information:
<style type="text/css" media="all">
.vanish {display: none;}
</style>
Documents can have multiple <STYLE> elements, and one of the greatest features of today's browsers is that these, and all the external style sheets, can be enabled or disabled singly. If the browser employs the W3C Document Object Model, then it accesses all the style sheets as a numbered "collection": the first in the collection is numbered [0], the second [1] and so on.
Now consider what would happen if we disabled the style sheet above, the one that does the concealing. All the other style sheets would stay in operation, but the concealing mechanism would become ineffective. And, logically enough, the concealed material would come out of hiding and the entire document would be "painted" on the screen or put on paper by a printer.
Here's how. Count how many style sheets (external or embedded) are referred to in your document and ascertain the index number of the one that does the concealing. This is the item you want to switch on and off. Suppose it is the second style sheet (index number = [1]). Then you will have to insert the following code somewhere in the text of the document:
<input type="button" value="Show Complete Text" onclick="document.styleSheets[1].disabled = true">
To make the whole business user-friendly, add another button to undo the effect:
<input type="button" value="Hide Non-Essential Info" onclick="document.styleSheets[1].disabled = false">
Have a look at an example. Don't forget to View ... Source. This should work in most browsers apart from Netscape 4.x.
If you prefer the document to load at the very beginning in the expanded view, add the word "disabled" to the style tag as follows:
<style type="text/css" media="all" disabled>
This will however only be honoured by Internet Explorer browsers.
And should you be wondering why the attribute media="all" was used in the <STYLE> tag above, study an explanation by Liam Quinn. The value "all" forces the under-endowed, 1997-vintage browser Netscape 4.x to display the expanded document with no collapse mechanism.
The online view of the document can be further compacted using a macro-typographical device, the run-in heading. This eliminates some unneeded white space from the interior of the document, and uses colour instead to break up the greyness of the text. Here is the CSS code that goes with the H element:
{display: inline;
background-color: lightblue;
margin-right: 1em;}
The first property-value pair, display:inline, not only ensures that the text following will appear on the same line as the sub-heading itself, but also eliminates top and bottom space around the heading in the major browsers. The background colour creates a highlighter effect around the text. The margin-right property is there to ensure that there is some empty space to the right of the heading before the text begins.
To keep this structure flexible, I dispense entirely with P elements. Instead I tag the entire unit— the heading and the text— as a single DIV. The heading is a sub-unit within the DIV. The text might also be book-ended between its own ad-hoc set of DIV tags to mark it off as an inner unit, but often this is not necessary. Leaving such sub-units unmarked is often called leaving them "anonymous".
Note that the preamble at the start of wills does not have its own sub-heading in the model texts, as the start of the text is clear to the eye and no added macro-typographical device is necessary here. Nor has any sub-heading been put over the disposition section, since the bulleted list makes this area a typographically distinct one.
A trick with the margin property can be used to add space to the left and right of the grant date in the heading:
.grantdate {margin-left: 1em; margin-right: 1em;}
This is to be preferred to methods that add superfluous content, such as empty spaces, to the source text. See the Alice Alefounder will after incorporation of these changes.
Note another place in a deed where it is also appropriate to use the UL element: when displaying the names of the two or more witnesses to the document. These persons are not usually members of the family being reconstructed, but their identities may well be of interest to other researchers, and their names should be shown in an easily readable fashion.
To demonstrate all the techniques, here is a long document combining wills from different periods and languages in a standardized mark-up that finds expression through the macro-typography. As a demonstration of the liberty this gives to change the page's appearance, the same document can also be seen re-styled in a different format.
Here is a demonstration of how semantic mark-up can illuminate the meaning of a handwritten document from the pre-computer age. Unlike many marking-up schemes, this one is designed with clear graphic presentation and simplicity as the two top priorities. The hierarchy below is one possible anatomy (intended for genealogists) of an English probate record, and demonstrates how each element can be appropriately styled in HTML.
The principal topic of interest in any testament will be the names of the beneficiaries.
| DIV.deed | H2 | testator | first name | |
| surname | ||||
| abode | ||||
| H4 | deed type (will or administration) | |||
| grant date | ||||
| jurisdiction | ||||
| DIV preamble | P deedsection (optional) | SPAN vanishing text | ||
| SPAN testator's condition | ||||
| DIV disposition | UL legacies | LI person | UL extra person | |
| DIV executor | H5 | |||
| SPAN vanishing text | ||||
| SPAN executor | ||||
| DIV signature | H5 with date * | |||
| DIV witnesses | H5 | |||
| UL the group | LI person | |||
| DIV codicil | DIV preamble | H5 with date | ||
| DIV disposition | UL legacies | LI person | ||
| DIV signature | ||||
| DIV witnesses | UL present | LI person | ||
| DIV grant | H5 with date | |||
| SPAN grantee name | ||||
| DIV bond | H5 with value | |||
| P deedsection (optional) | ||||
| DIV inventory | H5 with value | |||
| UL property | LI article | |||
| UL inventarizers | LI person | |||
| DIV archive | H5 | |||
* Additionally it might be desirable to include an image of the autograph signature, or at least some meta-information about the testator's autogram (shaky— indicates infirmity; a cross— indicates illiteracy; etc.)
Note in this analysis how the bullet-list presentation that has already been recommended for legacy listings can also be employed in supplementary documents such as codicils (later additions to wills) and in inventories of estate property.
If you do not have access to dedicated software to mark up transcripts, it is possible to fashion your own tools using editor software such as NoteTab for MS operating systems. Fookes Software of Switzerland provides a free version, NoteTab Light, which you can download at www.notetab.com. Users can then fashion their own specialized tools, or "libraries", within NoteTab. To get you started, I have provided a free sample containing most of the tags from the table above. After you have installed NoteTab, download the library by right-clicking on the hyperlink to this file. Select "Save target as ..." Save it to the following directory on your computer: NoteTab Light\Libraries\. Now load a file in NoteTab to be marked up, click on the DeedMark library from the libraries bar, select the text snippet to be tagged and double-click on the appropriate "clip". The result: instant tagging as you go. You'll find it easy to modify such a library for your own purposes.
A good transcription is more than just a faithful translation of pen-strokes to a digital, print-style alphabet. It has to expand whatever has been unduly abbreviated, convert shorthand to conventional orthography, provide guesses for illegible text (with the reasoning for the guess explained) and smooth over obvious errors - such as repeated words - by the scribe.
The typography should distinguish the words and letters that were unarguably in the original document from those that the transcriber has "read into" the document. Traditionally this is an either-or affair.
Major transcription projects such as the Papers of George Washington encourage their transcribers to "silently" expand shorthands that indicate a double letter, an incomplete word, a contraction or a misspelled word, but to enclose doubtful words in brackets. This practice— the two-way choice between original running text or editorial inserts— is dictated by the print form. It is a practice that makes the work of transcribing harder than it should be, because many transcribers are not so confident about their translations.
A digital transcript allows a wider range of presentations and therefore greater leeway. Suppose, for example, that a transcriber wants, for the sake of veracity, to also record the precise abbreviations found in the manuscript. The Text Encoding Initiative's recommendation on transcriptions suggests enclosing an expanded word in tags like this:
<expan>defuncti</expan>
with the abbreviation added as an attribute like this:
<expan abbr="defti">defuncti</expan>
This can be mapped quite well onto an HTML tag, for example:
<span class="expan" title="defti">defuncti</span>
The hidden text will be visible as a tool-tip if the web user places the mouse cursor over the word, as in this example:
defuncti. Try it.
This simple technique is explained in detail in the section of this style guide dealing with glosses. If the tool-tip has to contain unusual or differently sized characters, or a translation of a longer passage in Latin, use this style guide's technique for pop-up footnotes.
Examples: • Headings • Two Sub-Headings • Three Sub-Headings • The Core as List • Single-Item List • Expanding Text • Run-In Headings • Compilation • Variation
© 2002-2007 Jean-Baptiste Piggin. All rights reserved.