Translator 2.0: Translators and their tools

(Click here to see this article in Spanish)

When I made my first dives at this craft and this art that is translation my only tools were a pencil, an eraser, a lined notepad and three heavy dictionaries, one bilingual, and two monolingual, one for the source language, one for the target language. Back then computers were extremely heavy, slow and expensive, hard and boring to use for anything that was not video-games (which were very primitive by then, anyway), the Internet was still more a promise than a reality.

Translators may be classified within the category of “knowledge workers”. A knowledge worker is one who thinks for a living, and as thinking is nothing but a very complex kind of information processing, it is normal that the best tools for thinking are those from information technology. Nowadays the speed of information technology innovation is dazzling, and there is no dearth of available tools to ease the tasks of translators. Therefore this occupation is constantly changing, along with this innovation, to co-adapt ever more deeply with an ecology of tools that have been collectively called CAT tools (Computer Assisted Translation). These include programs that are needed for any task that implies information processing, such as digital reference books (with their respective reader software), data search interfaces, email and text processors. But within CAT we may also find more specialized tools.

The most specialized CAT software for translation is the inaptly called translation memory software. In fact, many people really refer to this kind of software when they mention CAT software. Some authors propose to call this kind of specialized software I am referring to as “Translation Environment Tools” (TEnT for short). These TEnT perform many functions to streamline translation; among them, the translation memory function. Thus, we have people who denominate TEnTs by their hypernim (“CAT”, which is a broader category in which they are included) and we have people who denominate them by their hyponim (“translation memory”, which is only a subset of their functions).

In a translation environment software both the source and target texts are simultaneously visible, in separate but adjacent text fields, and the parts of the source text that the translator may have already translated, show up in the target text field for the translator to use them or discard them at will when composing the target text, thus acting as an extended memory assistant. It also allows to manage terminology, whether it is gathered by the selfsame translator, or accessed from a collaborative project, as well as to manage projects, among other more specialized functions, although these tasks may also be performed with the use of other tools, as I will mention further ahead.

This kind of software has its advantages and disadvantages, so I personally consider it best to use it only when its positive aspects can be maximized. There are many options within the class of TEnTs, but the one we use currently at Translatare is Trados, which is also the most popular one, as it consists of Macros that integrate with Microsoft Office, generating within the Word document itself the two text fields, distinguished by pastel colors, and by their position (source text up, target text down).

Trados also allows to use memory fragments that are not exactly identical to the current segment, but which are similar enough to be useful. In this case, Trados indicates how much divergence there is between the remembered fragment (that had been translated in a previous occasion) and the current one, indicating a percentage of similarity. When they are exactly the same, this indication is 100%. One of its limitations is that it only works with source texts that come in formats compatible with Microsoft Office and with pdf texts when they have OCR (Optical Character Recognition), allowing for the digitalization of text, which is not always the case. Indeed, the pdf format has several other handling complications, as I will show later. Another limiting factor for Trados is its high price, which makes it difficult to acquire and update it.

Another specialized tool we have is AnyCount, a counter of words in files. With AnyCount a more precise word count can be performed within a variety of document formats, including non-OCR pdf files. Although in Microsoft Word it is easy to count words, its options are limited and in fact it does not have that same option in a native manner for other popular document formats, such as Excel and Powerpoint. AnyCount can count words for most common formats. Also, it has an option to automatically draft an invoice from the selected documents, which is very useful when you want to produce estimates of translation costs (per word) for any document.

It could be said that the rudimentary and solitary labor of the translator is being both technified and collectivized at the same time. The digital era allows us not just to better process information, but also to share it better, in order to coordinate tasks collaboratively. The translator 2.0 must conflate not just with the machine, but with other translators, with real or potential clients, and with the market and society. In this respect, the Internet is invaluable, and above all having a web address and an email inbox which can serve as a platform for contacts, to show one's work to the world, and to attract more and better jobs.

As is well known, natural language, as opposed to the formal language that rules the world of mathematics and informatics, is ambiguous. Actually, its power resides in this ambiguity. Nevertheless, this ambiguity, whose source is the multiplicity of ways of saying things, also becomes a linguistic barrier when translating as a work team.

Within the discourse of a single individual a certain amount of variability in diction is allowed as long as the register and coherence of tone are kept. But the technical terminology is plagued with synonyms, homophones, homonyms, hypernyms and hyponyms which may be freely exchanged in a literary context, where the entertainment factor has a greater relative importance. Nevertheless, in the technical context, where the instructive function has the greater relative importance, clarity mandates a terminological coherence, which is already hard to find in a single individual, but attains a volatile variability between different translators.

The tool that solves this problem of a lack of terminological coherence in the most cost-effective manner is the online application Google Docs. Through this application our network of collaborators keeps a series of glossaries that we all refer to and edit simultaneously in real time via Internet, without an editorial hierarchy, but with a clear system of rules that allows us to unify terms on the one hand, and solve quickly and resolutely any conflict that may arise as to "le mot juste", the just word. Furthermore, it comes with an option to review and undo a history of edits by each collaborator, which allows to perform a centralized unification in case it may be needed.

Google Docs is useful not just for terminology unification but for project management with multiple collaborators, allowing to set up an organizational chart more or less openly accessible to collaborators, containing the relevant data and metadata of a project, appointment of specific tasks to specific persons, and indications on general instructions. This Google tool allows a more dynamic flow of information at the level of internal operations, as well as at the level of customer relations management, and at the level of financial accounting.

In the year 2012 Google launched Google Drive, its own version of a kind of service that had recently begun to expand, globally recognized by the expression “cloud computing”, in which data and file storage, transfer and processing services were provided over Internet. At the same time that Google began to expand Google Docs to include not just text documents and data sheets, but forms, presentations, drawings, graphic calculations, and electronic signatures as well, it merged it with Google Drive.

Google Drive offers the same service as Dropbox, a major player in the cloud computing market: online storage of many kinds of files with readers for visualizing them, and featuring the possibility to access them in different places and devices as long as there is presence of Internet and the corresponding device application. In both cases, under a certain storage limit, the service is free, and above that, it is offered under some payment model.

The great problem with Google Docs is that in order to make the documents editable in its own browser-based platform (a capability which Dropbox does not offer, as it specializes more in the sharing of the files, without editing them) it requires the use of formats which are very similar but still slightly different from the most popular formats used by Microsoft Office. This implies that for very simple documents there is no difference, but for documents which use the most advanced properties of Microsoft Office, when Google Drive reconverts these document for transfer they lose part of the information or formatting of the original. On the other hand, although Microsoft Office is still the standard, that might change with the advent of free and open source, robust and agile word processing software, such as LibreOffice, although it still lacks some of the most advanced options in Microsoft Office.

The great advantage of Google Docs is that it is a part of the Google universe, Google being a constantly growing company dedicated to Internet solutions. Microsoft Office, on the other hand, lacks the capability to collaborate online with files, except with the use of absurdly tedious software such as Microsoft Groove, which is complex enough that an improper use of it may easily cause loss of important information, and especially of precious time, even when it works properly. Therefore, barring other solutions outside of Microsoft's most popular and reliable tools, the documents still need to be printed in order to be shared effectively. Google Docs, on its part, is in constant development and expansion, approaching the realization of the paradigm of the paperless office: a work environment where the use of paper is reduced or eliminated, with productive, economic and ecological benefits.

A positive step in this direction is offered by the recent fusion of the company HelloSign with Google Drive, allowing the electronic signature of contracts through documents on the cloud. The legal recourse of electronic signatures is in constant expansion all over the world. These are made even saver by means of the implementation of digital certificates, which can be obtained for a fee from third party services in digital certification.

The technosphere of computer assisted translation is constantly evolving, including innovations that a few years ago were still considered to be in the realm of science fiction, but which now are here to stay, and whose many possible uses are still being discovered, among them, and of importance to us, translation assistance. Between these are services of automatic translation, portable electronic devices and speech recognition software, which I will briefly describe in the remainder of this article.

The most popular automatic translation tool is the Google Translate website, in which a language pair and translation direction may be chosen, a text may be entered as input, and a scarcely usable target language text may be produced automatically. Google Translate sets misleading traps with an alarming frequency, returning apparently fluid and coherent text that is really a senseless gobbledygook, or even worse, a perfectly semantical and grammatical text that bears no relation to the content of the source text. This makes translation enormously difficult, because if at the beginning there was a source text that by itself may be complex and laden with booby traps, now there are two texts, the source, and an intermediary that, if the translator is not careful, will become an obstacle and confuse them, leading them astray.

The next innovation, much more useful and versatile, are portable electronic devices, whose main usefulness for a translator is acting as source that is more or less immediate (depending on device and application used) for access to digital reference material (such as handbooks, encyclopedias, glossaries, dictionaries or term databases, whether online or offline, whether monolingual, bilingual or multilingual). Although this may be done in the same computer used by the translator as workstation, the proliferation of windows and applications may dampen the workflow somewhat. On the other hand, resorting to a different device is at least an excuse to change body position, if only slightly and briefly, which is already a luxury for a translator who sits for hours in front of a computer. And of course, these devices, whether they be smartphones, tablets or ebook readers, have an invaluable usefulness when translating outside of the workstation, or interpreting (performing simultaneous oral translation), although the act of consulting them still gets a little on the way, interfering with the flow of translation or interpretation.

The barrier posed by current reference tools on translation and interpretation will, nevertheless, be overcome in a short time, especially with the advent of two new tools that will revolutionize not just translation, but all other aspects of our lives, if not directly by being quickly assimilated, at least indirectly by sparking innovative competition for a market in the kinds of tools these new products may implement. One of these tools is Google Glass, a hardware consisting of a head-mounted monitor. The other tool is the main product from the company Dragon NaturallySpeaking, a dictation writing software, and its variations (for different OS for mobile, tablet or computer).

Both of these tools will bring us closer to the paradigm of the cyborg, which means cybernetic organism, that is, one in which the control regulation systems have expanded beyond the body, into technological devices which expand the cognitive, perceptive and acting capabilities of the human body. And at the same time they bring us closer to another paradigm, one less anticipated, but perhaps more revolutionary: wearable computing. This concept refers to computing devices which, instead of being carried in bags or pockets, are worn in the same way as clothing items such as glasses, wristbands, earrings or belts. This tendency, allowed by the unstoppable miniaturization of chips and batteries, enables a greater mobility which in turn has an effect in the proliferation of the phenomenon of telework, which is also called work from a distance, or work from home.

Dragon NaturallySpeaking, which is already being marketed, is a speech recognition software that may be used to make dictation by converting natural human voice automatically into digital text, as well as to convert text into synthetic voice, and to execute voice commands. All of this promises to be a much more efficient and intuitive way of interacting with computers and robots, but is especially an opportunity for people who work with texts, among them, of course, translators. There are cases when the text to be translated is simple and voluminous enough that it is worth it to invest in a human assistant that takes a dictation, whether in person or from a digital recording which may be sent over email. This suffices when those conditions are met only occasionally, but when the bulk of the work is of that nature, it starts to become an attractive option to invest in the automation of the dictation process.

Google Glass, which at the moment of completing this manuscript is still in the development stage, will be the first massive commercialization of a head-mounted display device for augmented reality, that is, it allows to see more information about the surrounding environment than is visible by the naked eye, and even consult ancillary information without loosing visual orientation of the surroundings. That is, through the combined use of the head-mounted display and speech recognition it will be possible to compose text messages while crossing a street without danger, and of course, consult reference material without losing sight of the working text or the clients in an interpreting job.

Before I even thought I would ever be a translator I read a lot about Artificial Intelligence and I was very eager to be able to see Turing's test be passed by a machine. Then when I became a translator I started to fear that moment, because I thought the day a computer could generate natural language, translators would be left out of a job. But now I'm a rational optimist: robots and artificial intelligence are just the last expression of an ancient, non-stopping trend, and in fact, all of our lives improve when they improve, even if a few of us have to find new ways to earn a living. I know I'd like to earn a living at a café with a view to the sea, with a Google Glass on, taking my dictation.