[MUSIC] In this lecture, we'll talk about translation quality blocks. Basically, this is the type of mistakes that translators can make in their work. Why is this important? We talked about content classification in our previous lecture. It will help you understand which errors are critical in any specific project and which are not. You will see later that each quality block corresponds to a specific task that is part of the translation process. In our further lectures, we will learn how to predict which quality blocks are the most important using content classification. Based on these priorities, we will determine the best sequence of tasks in any given job. But first, we need to figure out what types of quality blocks or errors may be made by a translator. Mistakes can be classified differently and we will talk about this in more detail in the module dedicated to translation quality assessment. For the sake of convenience, let's now start on single model of error classification. The one that we use at Palex. In our company, we highlight six quality blocks, subject matter, accuracy, compliance, language, readability, and formal. You can read the detailed description in the additional materials for this lecture. Let's take a look at each block in detail. What is subject matter? We talked about these mistakes in our previous lecture. They can be seen in the cases when a translator did not have deep knowledge of the topics evoked on. Usually there are two key types of errors in this block. The first is subject matter mistranslation. This error can be seen if the translator didn't correctly understand ambiguous text. Here, you see an example. If you don't know how a device functions, descendants may carry both of the minions for you. But someone who works with the device on a daily basis will see the problem with the translation right away. The second subject matter mistake is the wrong terminology usage. If the translator doesn't know the subject matter well, they will not likely know how specialists in the field cause specific object to function. For example, what do you think this is? Is it a screw, a bolt, or a being, maybe spindle? In such cases, the reader will usually complain saying that the translators used wrong terminology. The second block is closely related to the subject matter, but has one fundamental difference. This block is accuracy and it includes all semantical errors with which the meaning of the text is lost or is relayed incorrectly. Accuracy errors include mistranslations, which is when the meaning is changed, omissions when the translator didn't convey the meaning fully and addition when the translator edited part of the meaning of the text when this was not necessary. It's important to understand the difference between accuracy and subject matter. Accuracy errors may be seen when the translator didn't clearly understand the source text, or couldn't convey its meaning adequately in the target language. For example, in this sentence, the grammar could be deciphered correctly only by 10% of the candidates that were tested. But to translate correctly, you don't need to be a specialist in the subject matter. All you need to know Is how English works. Compliance, this block includes all the errors that are related to the translator not following the client's requirements. Such requirements can be small such as a couple of lines in a mail or big multi page style guides that regulate even how ashes and commas need to be used in the text. Would indentation and spaces should be within paragraphs and much, much more. Glossaries can be another way for clients to convey their requirements. As we discussed this error, subject matter block may include errors as a type of way to incorrect terminology usage. But these are different. Not following the glossary is a compliance issue. Meaning that the translator has either forgotten about the existence of the glossary or didn't want to spend the time looking at it. A translator doesn't need to study how a machine functions in detail to choose a proper term for one of its path. They just need to follow the glossary. But if the translator wasn't given access to a glossary, then all the terminology errors are automatically placed into a subject matter block. Let's move on to the readability block. On the one hand, this block is about how easy it is to understand a translation and how pleasant it is to read the target text. The following errors fall into this category, redundancy, officialese, use of sentences that are too long and stringing structures. If you start falling asleep before finishing the sentence as is shown in the example here, then you see a clear readability error. On the other hand, this block reflects how much the style of the translation meets a readers expectation of the language used. This means that this block includes errors that are related to changes in the narrative within one text. Choice of figures of speech and phrases that don't correspond with the purpose of the text and don't appeal correctly to the reader. As you might guess, this quality block is closely related to the application of the text. Meaning that the requirements for readability will be different depending on how the text will be used. The last two blocks are simple and are usually checked automatically or semi automatically. Language includes all spelling punctuation and grammatical errors. Almost everything that's underlined by Word or CAT tools falls under this category. However, no tool out there can completely automate the language quality control for any and all languages. This is why people still need to be involved in proofreading. We have finally gotten to the formal section which includes errors which can almost always be fixed automatically. There are special quality assurance or great tools in the localization business. Which help to quickly find and correct cases of double spaces, wrong punctuation usage at the end of sentences, and correct formatting of numbers, dates, units and so on. For example at Palex, we use our own query tool called Verifika, which has over over 50 similar checks. We learned how to classify the target text and went through the errors that can be made during a translation. In the following two lectures, we'll go through how content classification can be used to accurately determine the translation requirements. Even if the client didn't provide them. You'll see how to predict what could go wrong during the translation process. Then we will try using contact classification to determine which quality blocks will be the most important during the translation and which can be ignored. [MUSIC]
