The development tools in the latter case were Borland's Turbo Pascal for Windows 1.5 + Paradox Engine 3.101 + minor libraries from independent vendors. Currently there is a 32 bit UNICODE compliant version under development, using Borland's Delphi32 Client/Server for increased performance. (Windows'95 and NT native code)
The LexiGraf system operates under Microsoft Windows, allows up to 20 concurrent network users and uses Paradox database files which can accommodate a maximum of 800.000 terms for each language. It supports several independent terminology sets for terminologists working on dictionary development projects evolving in parallel. To function, each LexiGraf workstation requires 4MBytes of RAM and 5MBytes of free hard disk space.
Customised versions of LexiGraf can be delivered -within a 2 month notice- in any language combination for which Ms Windows TrueType fonts and related keyboard drivers are available, ranging to a maximum of 8 languages. Currently this option applies for single byte language sets.
OCR input
Currently, development
work is being performed on combining the LexiGraf tools with optical character
recognition software to recycle terminology material already existing on paper for
updating purposes by the respective copyright holders. With the currently available
technology the process does require human intervention, some programming work
and subsequent proof-reading to be effective; Every existing dictionary presents its
own intricacies and requires individual handling. Nevertheless, a cost reduction of
50%-60% compared to manual data-entry is quite often feasible for good quality
texts. The same applies for the time duration of the data import.
After examining the given text's OCR performance, suggestions are made on the most
cost effective data-import configuration, namely:
- Manual data entry (typing).
- In-house OCR and subsequent proof-reading.
- Assigning specialised third party associates to perform the task; or combinations of
the above.
Final stage of the process is importing the data into LexiGraf databases for
processing, updating, appending and subsequent repagination. Although LexiGraf
supports a high speed internal printing interface and an excellent quality connection
layer to MS WORD, an interface between LexiGraf and Quark XPress is under
development using XTags. This feature enhances the DTP performance of LexiGraf,
due to the flexibility and operation speed it offers. Diagrams and pictures can be easily
imported to Quark XPress and integrated with the dictionary text, a feature useful to
scientific publications. The publisher will have the ability of easily personalising the
dictionary product by altering typefaces, adding footnotes, pictures, comments etc. Quark Xpress will also serve as
a bridge between LexiGraf and the Macintosh systems - which are standard in the
prepress and publishing business. The Xpress documents can be easily ported from
one platform to the other, even in a cross-platform PC - Mac LAN.
ONGOING DEVELOPMENT
- UNICODE support
- CLIENT/SERVER architecture
- Automated electronic dictionary application engine
- SGML import capabilities
