| { |
| "paper_id": "1981", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T07:47:03.614773Z" |
| }, |
| "title": "DISCUSSION ON SESSION 2: Aspects of Editing Rapporteur", |
| "authors": [ |
| { |
| "first": "Fran\u00e7oise", |
| "middle": [ |
| "Moore" |
| ], |
| "last": "London", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "", |
| "pdf_parse": { |
| "paper_id": "1981", |
| "_pdf_hash": "", |
| "abstract": [], |
| "body_text": [ |
| { |
| "text": "In response to surprise at the distinct gap between programmers and posteditors it was stated that European Commission translators were being brought increasingly into areas needing improvements, thanks to the Cambridge Language Research Unit's work on the annotation of programs.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.", |
| "sec_num": null |
| }, |
| { |
| "text": "As a criterion for use in the evaluation of MT, accuracy had superseded intelligibility, since an intelligible translation was not necessarily accurate and intelligibility varied with readers.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "2.", |
| "sec_num": null |
| }, |
| { |
| "text": "Public access to the European Commission's Systran was restricted for reasons of system rights and ownership. However, MT could be made available to EEC government agencies, and a pilot study in 1982 would concern the MT of abstracts retrieved from the Euronet-Diane information network.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "3.", |
| "sec_num": null |
| }, |
| { |
| "text": "The reason why posteditors had more difficulties with terminology than with syntax was that in technical texts the syntax could be modified easily once the terminology was correct, whereas the reverse was not true. The posteditor must of course know and understand the terminology to know whether the MT output was correct.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "4.", |
| "sec_num": null |
| }, |
| { |
| "text": "It was suggested that errors might be harder to spot when postediting is done on a screen, as opposed to hard copy. However, the screen was said to have the advantages of immediate correction and of access to more facilities (such as term banks). Horizontal screens would not necessarily be preferable, being perhaps an attempt by the new technology to ape the old.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "5.", |
| "sec_num": null |
| }, |
| { |
| "text": "Concern was expressed that high technology might lend authority to the text even more than does the printed word, and that inexperienced posteditors might therefore leave errors uncorrected. However, undue respect for machine-generated text, whether screened or printed, was said to be quickly unlearnt. The risk of error is eliminated if the posteditor, as in human translation, checks with the original. The Xerox speakers explained that the use of Multinational Customized English eliminated ambiguities in the source text, so outweighing any tendency of translators to miss errors in the machine translation. Without such controlled input, MT to a productive level was said to be impossible.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "6.", |
| "sec_num": null |
| }, |
| { |
| "text": "Another concern was that MT might impoverish the use of language and rob us of a great deal of language power. MT with restricted input, however, was said to increase clarity at Rank Xerox. With very diverse natural-language input, on the other hand, it was suggested that to postedit MT up to the standard of human translation was a false objective, since the quality of postedited MT was inevitably considerably lower.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "7.", |
| "sec_num": null |
| }, |
| { |
| "text": "The idea that MT might get rid of boring work was questioned on the grounds that revisers of human translations would often prefer to be translating; postediting should therefore be mixed with translating proper. However, it was suggested that SESSION 2 / SUMMARY OF DISCUSSION the aim was not to eliminate boring translations in order to do postediting, but to do interesting translations. Again it was stressed that postediting was a separate skill, which it took time to acquire.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "8.", |
| "sec_num": null |
| }, |
| { |
| "text": "The advent of MT service bureaux before long would bring various new possibilities, including postediting by customer or bureau (often using freelance posteditors). One such experiment in Canada was said to have been stopped for political reasons, another in Luxembourg due to problems of system ownership. Service bureaux would enable investment costs to be shared, not only for MT, but for documentation systems and terminology; a participant about to become the first Weidner customer in the UK was prepared to discuss a bureau service with others interested in productrelated vocabularies. Moreover, customers unwilling to purchase systems outright could now lease them.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "9.", |
| "sec_num": null |
| }, |
| { |
| "text": "The difference in translation costs between public institutions and the private sector was explored. In response to a suggestion that the difference was less marked in North America, it was reported that a 1979 study had found the difference between in-house and freelance translations in the UN even greater than that quoted for the EEC, possibly owing to poor-quality originals by non-native speakers. In the EEC study samples had been either human-translated and revised, or machine-translated, postedited and revised; the ordinary translator, posteditor and reviser computed their own time, and the cost was based on their respective salaries at the time. The figures now would be much higher. Surprise was expressed at the fact that figures quoted for private translation agencies did not include the cost of revising human translations, although some agencies did revise; inclusion of this would narrow the gap between public institutions and translation agencies.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "10.", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": {}, |
| "ref_entries": {} |
| } |
| } |
