Permuted Title or KWIC Hyperindexes
Christians. // All who profess and call themselves 389:16
less. // They politics like ours profess, / The greater prey upon the 231:26
of a dedication is flattery: it professes to flatter. // The known style 276:8
-akers; they hold up Adam's profession. // There is no ancient g- 437:5
that are contrary to their profession. // Those things 389:36
man a debtor to his profession. // I hold every 25:14
-tons---I panted for a liberal profession. // My father was an emi- 154:14
the head of the literary profession. // Your Majesty is 181:24
An ornament to her profession. 99:33
almost charmed me from my profession, by persuading me to it. // 480:29
Parentage is a very important profession; but no test of fitness for 489:23
member of the most ancient profession in the world. // Lalun is a 304:40
friend he drops into poetry. // Professionally he declines and falls, 178:3
against the laity. // All professions are conspiracies 489:20
-oing-good, that is one of the professions which are full. Moreover 546:40

Introduction

The hyperindex shown here is as it might appear in the window of a display screen, but without the usual decorations and accessories. It is adapted from the index of the 1941 edition of the Oxford Dictionary of Quotations to suggest how a hyperindex might be used interactively. Because excerpted quotations are being indexed, this example is more in the style of a permuted title index than of a full text KWIC index, so that end of excerpt indications (//) are prominent. Also, some of the lines are not filled with text as they would be for a full text index. The end of line indicator (/) in the first index line is such as would be used for line separation in poetry, or for paragraphing in plain text.

The background to this explanation of hyperindexes is given in the essay "The KWIC and the Dead: A Lesson in Computing History" which appears in the January 2001 issue of Computer. In that essay it is explained that, to my knowledge, hyperindexes have not been implemented, though static KWIC indexes are provided occasionally as an option for viewing search results made available through the World Wide Web, for example by Lexis.

Therefore the following description should only be taken as suggestive of how a hyperindex might appear to a user. In particular, the colours used should be able to be changed to suit the needs or tastes of a user, especially one who has some form of colour blindness.

The Display

The text displayed should be as dense as possible to allow browsing and association of ideas to be as free as possible, to which end all the capabilities of hyperindexing software must also be designed. The example as given here is less dense than a hyperindex of plain text might be, and, because of the HTML encoding used, more spaced out vertically than it should be.

The column to the right of the display provides an indication of, and a link to, the item from which the text of the line is drawn. The link could be used to bring a display of the text of the item, or of a hyperindex of it, into a separate window.

The middle line of the display is highlighted to indicate that it is at the moment the main line, and as such will provide different options than the secondary lines. In effect, the cursor of the display is on the keyword of the main line, and moving the cursor vertically will cause scrolling which changes the main line appropriately.

A sequence of lines on the display, rather than a single line, could be highlighted as main lines as a result of selection and as a preliminary to the lines being copied or otherwise manipulated as a group. However, cursor movement commands or keytaps will result in the normal, single, main line being highlighted.

Keywords and Sequence

The keywords of all the text lines, and there is only one in each line, are left aligned vertically, and define the primary sequencing of the lines displayed. The cursor of the display being on the keyword of the main line, moving the cursor horizontally will change the keyword in the main line. The effect is to move to another section of the hyperindex without changing the main line.

Where several consecutive lines have the same keyword, their sequence is secondarily determined. Unless the keyword is at the end of the excerpt, the secondary sequence is according to the go-words (see below) following the keyword in the line. If the keyword is at the end of excerpt, as several lines in the example are, then the go-words preceding the keyword are used for sequencing.

There are many options applicable to sequencing, such as whether case or punctuation or stop-words are ignored, and whether secondary sequencing is by preceding or following words or by some combination of them. For English text there are also optional possibilities in respect of the handling of prefixes and suffices. All these options should be provided for the user to control.

Main Line Operations

A main line, or a main line cluster, can be operated on in many ways as a line or line cluster, primarily by copying or deleting.

A user may maintain a collection of hyperindexes, and copy from Web or other hyperindexes into these. A main line may be copied as a single line, or as all the lines of the hyperindex in which the item it indexes appears. Copied lines are copied into the sequence of the recipient hyperindex, not merely appended.

When there is a single main line, another go-word in the main line may be highlighted, as head is in the example above. This highlighting marks the previous main line keyword, which makes return to the previous display easy, and which enables selection by combination of the main line's keyword and the previous main line keyword. This selection can be additive (or) or reductive (and) with respect to the overall hyperindex at which the original enquiry was directed.

Stop Words and Go Words

The light grey or pink words are stop words, words that are displayed but are not indexed and are not usually considered in sequencing of index lines. A stop word cannot be a keyword in a hyperindex display for that reason.

Stop words are words like and and the which either are so common as to clutter up an index if they're not ignored, or which are deemed to have little independent meaning, or both. A word that is not a stop word is called a go word.

For practical purposes, there need to be two kinds of stop words: global and local. Global stop words are designated as stop words when an index is being constructed, and cannot be switched by the user to become go words. On the other hand, global go words can be switched by the user to become stop words locally.

Word operations would flip the selected word (or all such words) between stop and go, move the view to the section of the hyperindex where the selected word is vertically aligned, or remove or add index lines for Web pages containing or not containing the selected word at various locations---adjacent or near to some other degree, and left or right or either side.

The Term Hyperindex

The term hyperindex was used in the original Computer essay as a derivative of index analogous to the derivation of hypertext from text.

The term has had other uses, for example applied to or within a domain name, a search engine, a geospatial database, a structured summary, an index of indexes.