VICTOR Translator is a new English-to-Spanish machine translator developed by MTC Soft, a small software company in Granada, Spain. The new translator was commissioned by BERCA Translator, a Granada translation company. This new product, which has just come on the market, promises to revolutionize the world of automatic translation. What does this revolution consist of? The answer is simple: VICTOR Translator makes good translations, something that no other machine translator had yet achieved with this language pair.


Objetive
When BERCA Translator commissioned the development of VICTOR to MTC Soft, they were seeking one objective: to achieve an automatic translation into Spanish which was so nearly perfect that the translated text would require a minimum of retouching in order to be apt for professional use.

The Challenge
Until now no machine translation program from English to Spanish had achieved satisfactory results for professional work. There are dictionaries of words, including short expressions, which are useful and in popular use, but the translation of paragraphs of certain length and complexity had not been resolved until now. This is due to several factors:

  • The complexity of the English language, particularly the great number  of words with multiple meanings
  • The different roots of both languages, which lead to great differences  between English and Spanish. For example:
    • Between English and Spanish syntax
    • The use of expressions which utilize the  same words combined in different forms

While the automatic translation from French, Italian and Catalan into Spanish were resolved years ago, the machine translators from English to Spanish were incapable of delivering reliable results.
Why hadn’t anyone succeeded in programming a reliable translator from English to Spanish until now?
Here at MTC Soft we believe that their failure was due to the fact that the companies who tried it wanted to apply the same basic analysis for different combinations of languages, even though the process is totally different depending upon the different combinations both to and from different languages. All previous attempts suffered from an inadequate process of analysis.

Now  we know that the key to the solution lay in a new approach from the very beginning, a revolutionary linguistic analysis which is systematic and exhaustive, dedicating special attention to the errors committed by previous programs, errors which seemed impossible to resolve.

MTC Soft benefitted from a double experience when it came to focusing on the problem and then solving it. In the first place we counted on the services of a professional translator with more than 20 years’ experience in the European Commission, a person who has worked with all of the principal translation programs and was well versed in their faults and limitations.  This was the person responsible for the
linguistic analysis, for the translation and for the elaboration of the system of identification and labelling with their linguistic attributes of all the words in a text.

Once the linguistic analysis was done in natural language, the MTC Soft programmers took over and “translated” this language into a computer program capable of creating, totally automatically, an “enriched text,” the greatest challenge in achieving a correct translation was practically overcome.

How does VICTOR Translator function in relation to the user?
VICTOR starts with an English text in MS Word (or compatible) and creates an analogous text in Spanish, also in Word. VICTOR integrates with MS Word as an additional “macro,” accessible to the user as a new menu option  on the toolbar, and also as a simple keyboard shortcut. A person who is working with a document in English can obtain a copy of the same document translated into Spanish with a simple click of the mouse on a menu selection or icon, or by using a combination of keys.

Could we see a sample of translations made by VICTOR compared to those of another traslator, Google, for example?
Yes, of course. The first of the following paragraphs is the original in English. The second is the translation made by the Google translator. And the third is the translation done by VICTOR. We think te difference in quality is notable.
Original text in English:
“Eurostat, in its press release, stated that it was not in a position to certify the figures included in the notification of Portugal, due, among other reasons, to shortage of information on capital injections from the Portuguese government to public corporations, which had been treated as acquisition of shares and other equities with no effect on the government deficit.

The same text translated by El Mundo.es:
” Eurostat, en su comunicado de prensa, declaró que no estaba en una posición para certificar las figuras incluidas en la notificación de Portugal, deuda, entre otros motivos, a la escasez de información sobre inyecciones de capital del gobierno portugués a empresas públicas, que habían sido tratadas como la adquisición de partes y otras acciones ordinarias sin el efecto sobre el déficit de gobierno. ”
The same text translated by VICTOR:
“Eurostat, en su comunicado de prensa , afirmó que no estaba en condiciones de certificar las cifras incluidas en la notificación de Portugal, debido, entre otras razones, a la escasez de la información sobre las inyecciones de capital del gobierno portugués para las empresas públicas, que había sido tratada como la adquisición de las participaciones y otros valores sin efecto alguno en el déficit público.”

General considerations about the quality of translations

VICTOR Translator is a software that translates formal, well-written texts from English to Spanish in automated way. VICTOR’s goal is not to reach the perfect translation of any text; this is in theory, but not in practice.

VICTOR is a useful tool to provide a translation of sufficient quality to be used as a draft for final editing – by a human translator – or directly to understand the content of a text in English.

To achieve this level of quality, VICTOR was developed for a large but limited type of documents.

VICTOR translates correctly:

- Business documents: brochures, manuals, guides or descriptions of products, services or companies, etc.

- Legal documents: contracts, conditions of use, safety, quality, etc.

- Technical articles, theses, educational material, etc.

- Articles and press releases.

In general, descriptive texts and complete sentences – with subject, verb and predicate.

VICTOR does not translate correctly or translates inaccurately:

- Poetry and literary works in general.

- Very colloquial texts with incomplete sentences such as chats, forums, etc.

- Sentences which contain spelling or grammatical errors.

- Incomplete sentences: newspaper headlines, succinct indexes, enumerations, etc.

Restrictions because of the size of the dictionary

There is another important restriction because of the development of the dictionary (it has nothing to do with the design of VICTOR):

The texts with words and expressions that are not previously inserted in the dictionary mislead in identifying the grammatical function, and this can affect the meaning of the whole sentence.

VICTOR’s dictionary has more than 300,000 words and all kind of expressions. Furthermore, it is already far advanced in some thematic or glossaries that are available on the test page in VICTOR’s web.

However, the language is too large and there are many words – and especially expressions – that are not inserted yet. For example, a nuclear physics text with very specialized terminology gives a rather poor result.

The dictionary – and the thematics that it includes – has a steady increase. If you need a large number of translations of a specific thematic you can consult us to incorporate this terminology. This operation takes no more than a few days in most cases.

Restrictions of the current version

By definition, VICTOR Translator is a never-ending job. It is made up by the variety of the language and the huge amount of words and expressions that make the improvement process of the program and the enlargement of the dictionary almost unlimited in time.

In addition to design restrictions above mentioned, the current version has some shortcomings that will be completed over the coming months. Most of them are only related to some syntactic structures. Others are negligible and do not change the meaning of the sentence, although some fine-tunings are required in order to get a perfect final translation.

Definitely, VICTOR achieves translations of the large majority of documents with a good quality and it is an useful tool for a human translator or the final user to understand the content of a text in English.

But it is interesting to know the relationship of changes and improvements that will be incorporated into expected versions and will amend known errors and shortcomings of the current version.

Outstanding improvements to include in the program:

- Names: the well-known are properly treated, although not all names in the world.

- Interrogative sentences. Some syntactic structures are not translated correctly.

- The plural conjugation of verbs with double subject or with direct or indirect object is not always successful.

- The gender and number of adjectives fail in certain circumstances, such as the connection of an adjective with multiple nouns.

- The “determination” of nouns – placement of the proper determinant – is not as accurate as it ought to be.

- Capitalization and punctuation are sometimes problematic, but nevertheless they change the meaning of the sentence.

Others improvements of minor importance are envisaged:

- Format: bold, underline, font, etc. Currently, VICTOR only holds some characteristics of the texts. Its main objective is the quality of the text rather than the format, but it is envisaged to improve this aspect and keep most of the attributes of the texts.

- File Format. The commercial version of VICTOR supports documents in several formats: Word, Excel, Txt, HTML, among others. Also, this list of supported formats will be expanded over time.

- Treatment of acronyms. It works correctly in simple formats but, when the structure of the sentence is more complex, it becomes less accurate.
Glossaries

At the moment, the major limitation of VICTOR is the size of its dictionary, despite it includes more than 400,000 entries. Most of them correspond to common language, although there are several glossaries regarding to specific thematics – you can see the available glossaries on the test page in VICTOR’s web.

When sentences include words or expressions that are not inserted in the dictionary yet, the result is not so good. Notwithstanding this, the solution is very simple: we insert this word or expression into the dictionary and now the translation of the whole sentence has been greatly improved.

Due to the richness of the language, the current number of specialized glossaries is not too high. In a short period of time, we can create a new dictionary for each customer with a significant volume of documents inserting the specific terminology and improving notoriously the quality of translations.

Paid version features

VICTOR Translator is already a commercial product and you can order its translation service for a very low rate per word.

Order it now and begin to translate your documents from English to Spanish!

Nowadays, the process consists in the reception of the documents by sending an email to BERCA Translator. Then, the document will be translated by VICTOR in a short period of time and, finally, we will return to you the resulting translation by email.

Several formats of documents are supported. We can also offer you the possibility of a custom-made dictionary. Contact with us to consider your specific case.