The METAL System. Status 1991

609

Authors

nan

Abstract

AbstractThe METAL system which originally evolved from a cooperation between the University of Texas and Siemens became a product in 1988. METAL is implemented on multi-user worksta- tions with a LISP server in the background. It is integrated into the office environment and permits automatic deformatting and reformatting of documents. METAL is characterized by recursive grammars, best paths parsing and a modular lexicon structure. Recent changes in system design have focussed both on internal structure and on user interface. Experiences with productive use have proven METAL’s cost-effectiveness but have also shown the need for increased cooperation between developers and end-users.

Keywords

nan

Full Content

[MT Summit III, Washington, DC, USA, July 1991] The METAL System. Status 1991 Thomas Schneider Siemens Nixdorf Information Systems Otto-Hahn-Ring 6 D-8000 Munich 83, Germany Abstract The METAL system which originally evolved from a cooperation between the University of Texas and Siemens became a product in 1988. METAL is implemented on multi-user worksta- tions with a LISP server in the background. It is integrated into the office environment and permits automatic deformatting and reformat- ting of documents. METAL is characterized by recursive grammars, best paths parsing and a modular lexicon structure. Recent changes in system design have focussed both on internal structure and on user interface. Experiences with productive use have proven METAL’s cost-effectiveness but have also shown the need for increased cooperation between developers and end-users. 1 Introduction The METAL machine translation system, originally evolved from a cooperation of the University of Texas at Austin and Siemens, existed as a prototype for German to English in 1978. After another ten years of develop- ment it finally reached the level of a viable product in 1988, As any machine translation system of today's state of the art, METAL is not intended for the impeccable translation of literary texts but for the rapid dissemina- tion of information of a factual nature, i.e. the types of text in which style is of less significance than the ex- plicit information content, which have to be translated fast and are of sufficient volume. METAL is used for two distinct applications: the translation of documents for publication, and the translation of documents for pur- poses of information gathering. In both cases, METAL has been used successfully. 2 General System Description Since the METAL system structure has been described in more detail previously ([Schneider 87], [Schneider 89]) the following outline will just summarize the main fea- tures. 2.1 Hardware METAL is implemented on a hardware package consist- ing of a number of translator work stations (Siemens MX-300or MX-500) and a translation server in the back- ground (Symbolics 36xx or XL). Translators work at a menu-driven terminal which supports various tasks: im- port of texts from other DP systems, automatic defor- matting, administration of translation jobs, postediting, automatic reformatting etc. The actual translation pro- cess runs on the server in the background and does not interfere with the translator's work. 2.2 Grammar METAL uses an augmented phrase structure grammar. The application of the individual rules is restricted by tests on features and values of the constituents and on their interaction. As they are recursively applied a lim- ited number of these rules is able to analyze surface structures not explicitly stated in the grammar. This has proven to be a viable alternative to the (mistaken) assumption that a large number of single rules, each de- scribing a specific surface structure, would eventually cover all legal utterances of a given natural language. Natural languages are infinite systems, and grammars have to allow for an infinite number of surface struc- tures. The prioritized chart parser operates on an agenda. …

Content truncated to 500 words

Labeled Text

AbstractThe METAL system which originally evolved from a cooperation between the University of Texas and Siemens became a product in 1988. METAL is implemented on multi-user worksta- tions with a LISP server in the background. It is integrated into the office environment and permits automatic deformatting and reformatting of documents. METAL is characterized by recursive grammars, best paths parsing and a modular lexicon structure. Recent changes in system design have focussed both on internal structure and on user interface. Experiences with productive use have proven METAL’s cost-effectiveness but have also shown the need for increased cooperation between developers and end-users.

Publication Details

Paper ID:
609
Year:
1991
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan