Workinprogress Enduser Requirements for MachineAided Translation Tools Hendrik

  • Slides: 47
Download presentation
Work-in-progress: End-user Requirements for Machine-Aided Translation Tools Hendrik J Groenewald and Gerhard B van

Work-in-progress: End-user Requirements for Machine-Aided Translation Tools Hendrik J Groenewald and Gerhard B van Huyssteen Centre for Text Technology (CTex. T™) Research Unit: Languages and Literature in the South African Context North-West University, Potchefstroom Campus (PUK) South Africa E-mail: Handre. [email protected] ac. za; Gerhard. van. [email protected] ac. za 5 July 2007; Potchefstroom

Overview • • • Project Status & Problem Statement Background: End-user Requirements Background: The

Overview • • • Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Development Model Current Phases Problem Statement Project Status • Development model of Whitten, Bentley & Dittman (2004) 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Introduction The Client Problem Statement Results Status of the Project Preliminary Recommendations Compiling End-user

Introduction The Client Problem Statement Results Status of the Project Preliminary Recommendations Compiling End-user Requirements Introduction Sub-heading 2 25% Grey/Bold Sub-heading 3 25% Grey/Bold Status of the project 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Development Model Current Phases Problem Statement Project Status • Development model of Whitten, Bentley & Dittman (2004) • Current phases: • Statement of Work • Business Requirements Statement 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Development Model Current Phases Problem Statement Current Phases 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Development Model Current Phases Problem Statement Current Phases 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Development Model Current Phases Problem Statement • Common misconception during the requirements analysis phase is that developers must determine what the client wants • Real objective: Determining what the client needs • This project is unique in the sense that the client does not have a preconceived idea of the functionality of the software that is to be developed • The development team has freedom in proposing a system or solution in the genre of machine-aided translation tools • How do we go about this? 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Background: End-user Requirements 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Compiling End-user Requirements 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Needs Analysis I • Why perform needs analysis? – To avoid the common scenario of engineers and programmers developing a system that they hope fits the requirements of the users • Success factors: – On time – Within budget • Critical success factor: – Software that meets the requirements of the user 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Needs Analysis II • First step for success is analysing the client’s situation as precisely as possible • Analyse the software that is currently being used • Determine the root of the problem • Goal: – To answer the question: What must the new software be able to do? 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques JRP • What is JRP? – Joint planning sessions of the intended software users and developers with the intent of eliciting user requirements • Goal of JRP: – Providing an open environment for people to discuss their requirements • Success depends on: – Availability of facilitator with superior mediation and negotiation skills, – to ensure that all parties receive equal opportunities to contribute to the development of the system 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Interviewing • Individual interviews are held with the intended users by the development team • Start with a list of prepared questions with subsequent questions posed in response to the answers provided by the person being interviewed • Listen to what the client has to say • Important: – Interviewer must suppress any preconceived notions regarding the needs of the client 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Questionnaire • Interviews may be more effective than questionnaires – new questions can be formulated based on the response to previous questions • Questionnaires are more effective in terms of cost and time when the input of a large number of individuals is required • Advantage: – Carefully though-out written answers from clients may be more accurate than verbal responses • Disadvantage – Low response 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Introduction Needs Analysis Fact-finding Techniques Research • It is important that the development team must be familiar with the application domain • You cannot ask meaningful questions to a specialist without having knowledge about his/her field of specialisation • Using the proper jargon is important while communicating with the client 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Structure NLS and Translation The Client 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Structure NLS and Translation The NLS and Translation • The NLS translates documents for government • A large number of translators and terminologists are employed by the NLS to comply with the large volumes of translation work 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research The Process: Capturing Methods • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research The Process • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research JRP • JRP was held at the NLS in Pretoria • The development team collaborated with translators from African Languages, as well as FLEA • A limited number of terminologists also attended the meeting 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research The Process • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research Interviews • Individual interviews were held during the course of the day • Obtain technical information – e. g. infrastructure, data management, etc. 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research The Process • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research Questionnaire • A questionnaire was distributed among the people attending a meeting of the National Language Forum • The National Language Forum has the goal of advising DAC on language policy matters • 15 Respondents completed the questionnaire 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research The Process • • 5 July 2007; Potchefstroom Joint Requirements Planning (JRP) Interviewing Questionnaires Research Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research • Currently a large number of research projects addressing the topic of machine translation, are ongoing • Very little information on end-user requirements • Exception: Metis-II project 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research METIS-II • Successor of METIS-I, a successful MT project completed in 2003 • Aim of METIS-II: Constructing free text translations through pattern matching/statistical techniques and large monolingual corpora • Partners in the METIS-II project: – Institute for Language and Speech Processing (ILSP) – Katholieke Universiteit Leuven (KUL) – Institut der Gesellschaft zur Forderurg der Angewandten Informationforschung (GFAI) – Fundació Universitat Pompeu Fabra (FUPF) 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations The Process JRP Interviewing Questionnaires Research End-User Requirements in METIS-II • Requirements were obtained by means of a questionnaire • Remarkable similarities between the questions in the METIS-II questionnaire and our own • Both the questionnaire and results are available at Available: www. ilsp. gr/metis 2/ 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Towards a Proof-of-Concept 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations PIECES • The results of the JRP and interviews are presented in the PIECES framework (Wetherbe & Vitalari, 1994) • The PIECES framework is a method for identifying and classifying problems within an existing system according to certain categories • PIECES is widely used in the systems development process • The six categories are: – – – Performance Information Economics Control Efficiency Service 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Performance • Work-overload: – The NLS currently receives more work than their translators can handle. – This overload of translation work has a negative impact of the throughput and response times of the NLS • Work-overload is the essence of this project – Served as stimulus for funding by DAC 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Information I • Lack of electronic resources: – Electronic dictionaries: limited access by employees of the NLS • FLEA has access to adequate dictionaries and reference materials, although not in electronic format • The personnel of African languages generally do not have access to sufficient dictionaries (electronic and paper) • Paper-based resources such as dictionaries and terminology lists are widely used – Online resources: • FLEA uses online resources such as freedictionary. com, Google and Wikipedia, which they found to be very helpful. Access to online resources is restricted because some of the translators do not have internet access 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Service Information II • Lack of electronic resources (cont. ): – Electronic tools (such as translation tools, thesauri, spelling and grammar checkers) • Spelling checkers used by most of the respondents. • Limited use of other tools • Online machine translation systems such as Google Translate are not widely used by the translators. – The translators at FLEA uses online translation systems such as Google Translate and Altavista Babelfish for translating between English and foreign languages such as German and French 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Information III • Lack of electronic resources (cont. ): – Translation/editing environment • Trados Workbench, a computer-aided translation system, was used in the past by the translators. Trados is not used anymore, due to software compatibility and licensing issues • Hence, translators make no use of previously translated documents • Most of the translation work is done with MS Word as the editing environment • MS Word is currently the most popular format in which documentation is received for translation. PDF (4. 87%) and Open. Office (2. 03%) documents are not very popular 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Information IV • Limited interaction between terminologists and translators – Translators do not receive direct help from terminologists and are compelled to rely on other translators for help with terminology issues • No terminology management solution – The NLS has a terminology-database, but this is not electronically accessible by the translators due to license restrictions 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Economics • Outsourcing of work is expensive – Since there is too much work for the number of employed translators, some of the translation work is outsourced – Outsourcing is more expensive than in-house translations and should therefore be limited – Translation work is also outsourced because the NLS does not have translators for certain foreign languages 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Control • Inefficient document management solution – Document tracking – A document tracking system was developed in-house, but is not used by translators. An electronic register containing information such as the title of the document, data on which it was received and the cost of translation is used by FLEA in the place of a proper document tracking system – African languages do not use an electronic register. They keep their own records on their personal computers, while their register is kept by admin support – The naming system of the files used in the translation process varies between FLEA and African languages – Data redundancy – African Languages receives the documents that need to be translated via e-mail and hard copy 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency • Unbalanced administrative support – FLEA does not get the same amount of support from admin as African languages • Inadequate infrastructure – Hardware: Most translators do not have dual monitors connected to the PCs – OS: MS Windows is the only operating system on the computers of the respondents, despite governments plans to move to Open. Source. This might cause portability issues in future. – Internet Access: not all respondents have internet access at work, while very few have internet access at home 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Service I • Quality of Service: – The quality of language used in some of the received documents is not very good – Inaccurate translations – The overload of work that the NLS receives has a negative influence on the quality of translations – Translators do not receive any feedback from users on the quality of their translations. The danger of this is that the same translation mistakes are repeatedly made – The NLS does not require their translators to be accredited with SATI (South African Translator's Institute) 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Service II • Limited control over quality – The quality of the translations are not evaluated or monitored. The result of this is that the quality of the translation is solely dependent on the proficiency of the translator – Since the translations are performed by a large number of translators, the quality of the translations are also inconsistent – The NLS does not have control over the quality of outsourced translations. Outsourced translations often need to be edited by the NLS before sending it off to the client 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Performance Information Economics Control Efficiency Service III • Wide range of services – The translation and terminology work performed by the respondents are more or less equally distributed between different government departments – The translation work is performed on a variety of document types. Reports, publications, legislation, manuals and regulations are the most popular types – Wide variety of languages • Source language: mostly English (some Afrikaans) • Target language: – All SA languages, with tendency toward isi. Zulu, Sesotho sa Leboa, Setswana and Xitsonga [might be skewed due to small group of respondents] – Other foreign languages (mainly French, Spanish, Portuguese, German) 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald Service

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations General Findings • Negative Attitude towards MT – People are afraid of losing their jobs – METIS-II: Low quality of the available MT systems 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Next Steps Preliminary Recommendations • More time and effort should be invested in fact finding techniques for compiling user requirements • Making recommendations regarding possible solutions at this early stage of the project is not ideal, but some preliminary solutions could be provided • PIECES identified diverse problems at the NLS – We focus on problems directly related to the translation process – Other problems such as document tracking, administrative support, business process management etc. are of no interest • It seems as if there is a need for an integrated translation environment, since translators use a variety of software and resources in their work 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Next Steps Preliminary Recommendations • Integrated translation system: • • • Text editor Translation Memory Terminology Manager Machine-aided Translation Tools Electronic Dictionaries • These tools are readily available for language such as English, but do not exist for other official SA languages • Not sensible to developing an integrated environment if we do not have access to the required modules • Recommendation • Firstly concentrate on developing the various tools • The integrated translation environment 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods

Project Status & Problem Statement Background: End-user Requirements Background: The Client Process: Capturing Methods Findings: PIECES Preliminary Recommendations Next Steps • Evaluate/Investigate other similar software • Up to now: avoided contact with other software, because we were afraid of obtaining preconceived ideas about possible solutions to the client’s needs • Fact finding • More fact-finding sessions should be organised • Proof of Concept • Proposing possible solutions to the client 5 July 2007; Potchefstroom Gerhard B van Huyssteen and Hendrik J Groenewald