As the 1990s progress, practitioners are beginning to use computerised litigation support to manage documents more efficiently and less expensively.

Computerisation allows litigation teams thoroughly to index, search and retrieve documents, thereby minimising the potential for human error.Although the technological tools of litigation support are continually improving, the basic underlying concepts of computerised document management remain the same: document screening, indexing, searching and retrieving.

The principal electronic tool to achieve these tasks is the database.Computerised litigation support is a process of comb ining the activities of people with the use of technology to acquire, manipulate and present case information in support of winning litigation.

The goal is to conserve staff, time and money while making the lawyer more efficient in retrieving information.

This is accomplished by creating a database to manage, store and/or retrieve information.There are several types of database: image-based, bibliographic, and full-text.

Each database provides a unique type of document management and retrieval capability.

To take advantage of litigation support successfully, it is important to understand the appropriate application for each type of available database.

The goal is to design the most useful database at a cost that is commensurate with its value, with a minimum risk of having to redesign it.

Since each type of database has advantages and limitations, it is essential to know what types of searches will be performed and what information will need to be extracted from the database.Before the information in the case documents is indexed, the litigation team must set up relevancy criteria and screen the documents, culling those pertinent to the case from the larger document population.

Once screening is completed, a master copy of the relevant documents must be created which can be easily and inexpensively duplicated, and from which subsets of documents can be made.

During the 1980s, microfilming largely replaced creating a paper master.

Nowadays, image scanning, in order to create an image database, is beginning to replace microfilming.Image scanning or imaging is simply the process of taking an exact picture of a document and storing it on an electronic medium like an optical disk or CD-ROM.

Having retrieved the document from the database, the lawyer can print it out if necessary.Imaging is supplanting microfilming for several reasons.

Although the cost of creating microfilm or an image database is roughly equivalent, the cost of electronically retrieving documents from an image archive is significantly lower than from a microfilm archive 'blow back' to paper copies.

Perhaps more importantly, an image archive is far more convenient for a lawyer to use.

Documents can be searched for and retrieved instantly using a PC litigation workstation.

Image databases can also be printed to CD-ROM which the lawyer can use on a portable notebook computer.Imaging technology consists of two processes: image scanning and linking the images to a database.

During scanning, each document is converted into a digitised image.

As the documents are 'fed' into the scanner, the imaging process captures an exact replica of the document.

Everything found on the face of the document including typed letters, signatures, graphics and handwritten notations can be viewed or printed.Since imaging is only a picture of the document, information contained within the document cannot be searched for from the image alone.

Bibliographic or full-text databases need to be created to index the documents contained in the image database.

These databases are linked to the images through a common identifier, usually the document number.

This allows the user to retrieve and view the document on a PC screen and print a copy of the actual document after a search is completed.

Since the images are stored electronically, image archives eliminate the need for large storage repositories.The term bibliographic or 'index-style' database indicates that specific pre-determined information is extracted from documents.

The information in the database 'stands in for' the original docume nts, but does not contain all the words found in them.

Litigation management primarily uses a bibliographic database to organise case documents such as reports, invoices and correspondence.

By creating a bibliographic database, lawyers can manipulate the document collection through the use of a computer to locate a document, find the location of all documents pertaining to a witness, or organise specific documents in chronological order - all within minutes of receiving a request.Some standard 'fields' or types of information that are 'coded' or extracted from documents are: document number; document date; document type; title or subject; author names and affiliations; recipient names and affiliations; copyee names and affiliations; and mentioned names and affiliations.A bibliographic database should be created:-- When the litigation team requires field-limited searches.

The ability to search for documents written by a specific person (an example of a field-limited search) requires that a bibliographic database with an 'author' field be created.

A full-text database will report if a person's name appears in a document, but not if the person wrote or received the document.

In field-limited searching, fields can also be cross-referenced to yield focused search results.

For example, by creating a complex search with multiple fields, the database can locate which documents were written by a specific person to a specific recipient within a given data range.-- When summary reports are needed.

If the team wants to create an 'index style' report that displays certain information about the document such as document number, date, document type and title, a bibliographic database is created.-- When the search results need to be sorted in various formats.

If creating a report of the search results in both chronological and document number order is required, a bibliographic database is created.Once the lawer identifies the need for a bibliographic database fields are defined and coding conventions are created.

Each document can be reviewed to capture (code) its relevant information.

Most commonly, coders work from paper copies of the documents, keying the information onto a coding screen on a personal computer.

There is now a movement toward coding directly from electronic images to streamline the process.The term 'full-text database' means that all words, exactly as they appear within the document, are searchable in a database.

A full-text database is primarily used to search for specific statements within the text of documents, for example, trial transcripts to identify the exact location of a statement.

Since transcripts are typically obtained from court reporting companies in a machine-readable format (on diskette), the time and cost to create a full-text transcript database is minimal.When a full-text database is required, but the documents or information cannot be easily obtained in a machine-readable format, the information may need to be keyed into the computer.

The cost to key information is relatively inexpensive if a keying service performs the service rather than a secretary or word processor.

However, since most keying projects are voluminous, the cost just to prepare the data for loading into a full-text database can quickly become expensive.Another method of converting the text of documents into a machine-readable format is optical character recognition (OCR)/intelligent character recognition(ICR).

OCR/ICR is a fairly new technology which is often confused with optical image scanning and misrepresented as the new litigation support tool that will replace the need for bibliographic databases and coding.

This is incorrect.

OCR is the process that converts a scanned document image into a machine-readable format.

In its simplest form, the process looks at each character and makes a best guess judgment as to what character it is.

By doing so, the OCR conversion process creates an 'input file' (the machine-readable data) for a full-text searchable database.Before selecting OCR conversion as the method for creating a full-text database, it is essential to understand the limitations of this technology.

Even though technological advances occur rapidly, there are still many converting limitations with the OCR process.

The majority of OCR applications can only recognise typewritten information.

The quality and condition of the documents also affects the success of the conversion.

Factors that make it difficult or impossible accurately to OCR convert documents to a machine-readable format include:-- Age of the document.

Older documents often have illegible type print and paper discolouring which makes conversion difficult.-- Quality of the document.

Laser print quality or first generation copies typically convert with a greater success rate than documents of lesser quality.-- Handwritten notations.

Not only is handwritten information difficult to convert, but if the handwritten notation 'touches' typed information the typed information now becomes difficult to convert.-- Graphics/charts/tables.

Only the legible typed characters contained within graphics, charts and tables will convert.-- Highlighted or underlined information.

Highlighting or underlining information within the text of the document make it difficult for the OCR conversion to recognise characters.

In many cases, it causes the conversion incorrectly to identify characters.-- Colour of the paper.

Non-white paper prohibits the OCR process from reading the character correctly.Remember, the OCR process can misread the characters in a word, thereby altering the meaning of a word or group of words.

For example, it is possible for an original document to contain the phrase 'waste zone' and have the converted, machine-readable file identify the same phrase as 'was one' because the print on the original document is too light for the OCR conversion to read some of the characters.If the document collection contains some of the above-mentioned difficulty factors, a quality review, or editing process, is performed to ensure the integrity of the converted information.

If the editing process is simply a spell-check, there are few safeguards to prevent incorrect words from appearing.

An 'editor' must compare the converted information to the original document to guarantee the precision of the conversion.

Because of all the conversion limitations, the editing process is often time consuming and, therefore, expensive.

Explore this technology carefully before selecting it.When a full-text database is created, the following factors must be considered: availability of the original text to be in, or inexpensively converted to, a machine-readable format; ability effectively to obtain information without field-limited searches and the extent of need to search all words in the original documents; effectiveness of using a database without report generation capabilities; percentage of 'boilerplate' documents where most words are repetitive and, therefore, not useful as search terms; and amount of data storage capacity on the computer (full-text databases take up more space on the PC than a bibliograp hic database).Once the bibliographic and full-text databases are created, they are linked electronically to the image database.

The desk-top computer becomes the document repository, reading the image from an internal optical disk drive or CD-ROM or using an image file-server.

Legal teams are able to pull, duplicate and refile documents in a keystroke or click of a mouse.

Image databases truly provide instant, fingertip access to documents without the time and cost typically associated with managing small or large document collections.

Practitioners receive the documents instantaneously, giving them more time to analyse them.1995