Challenges in Localization

By Alessandra Muzzi (English text edited by Anne Appel)

The software and Website localization market, in spite of the recent economic slow-down, is still active and able to offer interesting opportunities to those translators possessing, or who are willing to acquire and develop, the required technical knowledge. This article is not intended to offer an exhaustive description of localization in all its phases and aspects. Such a task would require more space and is best tackled by some of the resources listed at the end of this article. Instead, its objective is to explain what is meant by localization and to give an idea of what challenges it presents to localizers, particularly Italian localizers.

What is Localization?

When explaining the meaning of localization to newcomers, the aspect of linguistic and cultural adaptation to a local setting is often stressed. As a result, some colleagues conclude that "to localize simply means to translate," as was stated some time ago in a mailing list for Italian translators (Langit, www.vernondata.it/langit/index.html). Actually, a certain degree of adaptation characterizes almost every field of translation, and cannot therefore be considered unique to software localization. When translating a commercial, legal, or advertising text, for instance, a professional translator is fully aware that a superficial, literal translation often makes little sense, and that concepts, formulations, or slogans must be reworked so as to fit a different social, cultural, and/or judicial setting. In the localization field, such interventions must often go beyond the purely linguistic level and involve technical aspects, but the independent translator is not usually concerned with these. What we are talking about here are issues such as date/time format, currencies, fonts, etc., which are either dealt with in-house or managed automatically by the latest versions of the most common operating systems.
The attempt to define localization merely in terms of "cultural adaptation" is perhaps influenced by the word itself and by some official definitions, like the one provided by LISA (Localisation Industry Standards Association), which states: "Localization involves taking a product and making it linguistically and culturally appropriate to the target locale (country/region and language) where it will be used and sold." Two new ideas are introduced here: "product" and "sales." The scope of localization is confined to commercial products, where text is part of the product itself and not merely an accessory. Thus, translating a technical document, such as a user manual for a household appliance, cannot be defined as localization. Although a publication such as a newspaper or magazine might conform to LISA's definition, the term "localization" is currently used mainly to refer to software and Website localization, which has unique characteristics that distinguish it from other kinds of translation. This specialized field presents the translator with a set of technical issues which differ from those arising when translating other types of technical documents, such as automotive or construction texts.
What makes software translation unique is the fact that the texts to be translated do not just describe something external to them, but are an integral part of what is described: they are part of the product's operation and integral components of it. Translators-localizers do not just have a responsibility to create a translation that is clear and readily understandable by the user-reader. They are also responsible for assuring the correct operation of the localized product. Software programs manipulate texts as byte strings, store them on magnetic media, display them, and print them. Software users enter or select text, and text messages interactively explain how the software works or warn about possible errors. Text contained in graphics may also require translation. In a localization project there will not be a continuous flow of text in a single file format, but frequently several chunks of text presented in diverse, difficult to understand file formats. Localizers are often asked to work with specific translation and localization tools and to learn to use new tools, which are sometimes proprietary or even specific to a project or product. A localizer must have the flexibility, confidence, and computing skill needed to quickly and easily comply with these requirements. If mistakes are made when translating these texts or using the localization tools, the program might not compile or work properly. Since so many technical aspects have to be taken care of, specific technical skills are required of translators working in the localization field. However, very few translators will have the know-how needed to complete a localization project from start to finish. Experts in the field of programming, DTP, and computer graphics must usually participate in the project as well.

Internationalization: Making the Localizer's Task Easier

A localization project begins the moment the software product is created. Successful localization requires proper internationalization; that is, designing the software product in such a way as to facilitate and speed up its localization. All the text to be localized must be clearly separated from the software code and made independent from it. This also implies that the code must not deconstruct semantic units within the text. The minimal unit carrying a complete meaning roughly corresponds, in any language, to what we call a "sentence." These are the units that computer-aided translation/translation memory (CAT/TM) tools try to identify when splitting a text up into "segments."
At times, though, a software code might operate on text below the sentence level: on single words or even morphemes. Programmers, not being linguists, are often victims of the common illusion that differences between languages are simply a matter of using different words, but that the meaning behind those words, and the way in which they are put together to make up sentences, are basically the same. You may still find small and medium-sized software companies using "creative" programming methods, which can cause considerable trouble for whoever must localize their products. For instance, a programmer might decide to build GUI messages by putting together chunks of text at runtime. You might have messages like these (this is from a real localization project—the %S stands for an unknown variable portion of text replaced at runtime):

1) File copy complete; %S %S records processed. %S.
2) %S %S Phase %S %S
3) The %S %S record has been reverted to revision %S.

In string #1, the first %S might be a number and the second %S an adjective or qualifying noun. If so, translating "Copia file completata; elaborati %S record %S. %S" might yield a reasonably clear message. Unfortunately, the QA revealed that the last %S is replaced at runtime by a whole English sentence, which was not localized because it was hard-coded.
String #2 is absolutely cryptic. You have four %S's and only one known word. The only thing a translator can do is translate "%S %S Fase %S %S," knowing almost for sure that this won't work. The order of replaced text should most probably be changed in Italian, but there is no way to know how, and even if there were, there is no way to distinguish between the various %S's.
String #3 is more understandable, but there are still problems. The two %S's before "record" might well be part of a noun chain that would need to be inverted in Italian, but again, there is no way to know for sure, nor to operate the inversion. Actually, the whole sentence might need to be inverted, like this: "È stata ripristinata la revisione %S per il record di %S di %S," but variables would be replaced at runtime in the same order as the original English string, and the whole message would make no sense.

Issues Specific to Italian Localizers

The word order issue is particularly relevant to the Italian language (and indeed to other languages as well), since the order in which information is presented often differs from English. In Italian, we tend to convey more general meaning first and more specific information second. It is quite the opposite in English. See the following example:

Assorted sample users are shipped with the software.
The new information is provided first (assorted sample users), followed by the known context. The Italian translation would look like this:

Il software è fornito con diversi utenti di esempio
Here, the known context is provided first, followed by the new information.

This sentence can be correctly translated without problems, but sometimes incorrect text layout and segmentation, or the presence of placeholders (as in the example previously illustrated), can be a hindrance to a proper rendering in the target language.

Another linguistic issue that is actually common to most European languages is that of text length. Translating from English to Italian carries with it the phenomenon of text expansion, which can be a problem in software interfaces due to the limited space available for strings. Internationalization rules prescribe that adequate extra space be allowed by GUI designers to accommodate for text expansion, but these rules are not always adhered to, and even when they are, the space provided might be insufficient in any case. Localizers must learn the skill of concision¾omit articles and prepositions, use short words, omit unnecessary words, and use abbreviations only as a last resort.
Actually, English sentences sometimes present a certain degree of redundancy that can be safely omitted in the Italian translation. Take this example from Microsoft:

Do you want to save the changes you made to "Test document.doc"?
Salvare le modifiche a "Documento di prova.doc"?

Still, in the linguistic field, there is the issue of loanwords. Among FIGS languages, this issue has particular relevance for Italian. The French have adopted a policy of translating foreign words, and the Spanish are on much the same wavelength. German adopts many English words in the software field, but since the two languages share the same roots, German manages to integrate English loanwords in its phonetic, lexical, and syntactical structures quite smoothly. The situation is quite different for Italian. There is no set policy for translating foreign words, and these words, being so different phonetically and morphologically from Italian words, may have a serious impact on the very nature of the language. Because of the lack of a central authority on language, comparable to the Académie Française, engineers, technical journalists, writers, and translators have no clear-cut indication about how to translate a new foreign technical term. Often they will conclude that "no translation exists," and use the foreign term. One Web search technique often used by translators contributes to the adoption of the foreign term: translators will search for the foreign term in Italian Web pages. If they find any occurrences, they will conclude that the foreign term is used in Italian and use it themselves. All this in spite of the fact that a search for an Italian translation of the term might also yield results—if they only knew what to search for.
A fourth, minor linguistic issue, which is nonetheless relevant for the Italian language, is that of capitalization. In Italian schools you are taught to capitalize only proper names, and words at the beginning of sentences. In the English language, capitalization is more widespread than in Italian: there are the well-known examples of weekdays and months which are capitalized in English but not in Italian. This rule is usually taught at the very beginning of English courses in Italy, and therefore few translators would ever fall into such a trap. But there are other insidious traps: words in titles, for instance, which are all capitalized in English, but which should not be in Italian. Then there are the words for appliances, software components, names of interface items, and so forth. English usage acts as a magnetic force on the Italian language, and old certainties about capitalization are being eroded. Even reference materials, which are usually considered the localizers' Bible, such as the Microsoft glossaries, are not always consistent on this issue. In the Italian glossaries, for instance, you will find "accesso remoto," "Accesso remoto," and "Accesso Remoto."
Italian localizers also have to face issues relating to other aspects of their profession, such as the localization market. Unfortunately, Italy seems to be among the European countries with the lowest translation rates, and good Italian translators and localizers find it hard to have their professionalism acknowledged by equitable remuneration. For Italian freelance translators, the situation is worsened by the strong presence of an intermediation layer.
For all localizers, one urgent issue that needs clarifying, in my view, is that of CAT discounts, where complete anarchy reigns, mainly favoring large software houses and large localization companies.
These are complex issues which cannot be thoroughly examined here. I will just express my view that, despite the ever-increasing number of Italian localizers, the market and the profession would still benefit from the injection of quality-conscious professionals. For those colleagues who are interested in delving further into the topic of localization, I suggest these links.

(Article published in the ATA Chronicle – Volume XXX - Number 11 - November/December 2001 – A Publication of the American Translators Association)