A typographical error (often shortened to typo) is a mistake made in the typing process (such as spelling, or leaving out a word). As a linguist, you must have encountered these basic issues, but to avoid them is another question.
We work on the English to Indonesian translation project recently and we found out that any spell-checking machine including CAT Tools with their advanced spell-checking features cannot detect these typos accurately, even the most sophisticated ones. The human resource still plays a very important role to prevent the typos from occurring in the translation. We are sure this does not happen in the language pairs we work on only. Other languages may have the same issues.
Ideally, translation should be final-checked after editing and proofreading to avoid these typos as we found that typos cannot always be detected by the CAT Tool we used. We would like to share the case study below:
A translation company assigns an English to Indonesian project and it is related to website localization. The word count is around 5,000 words and a specific Cat Tool has to be used. Assume SDL Trados 2011 as this CAT Tool has good spell-checking feature for Indonesian; at least it is based on our 2-year experience of using this CAT Tool. The translation process is completed and it now goes to editing and proofreading process and to final checking. In spell-checking the translation, we found some typos that we can detect earlier using this CAT Tool;
- Need More Talk Time for Your Mobile Phone? –> Butu Lebih Banyak Waktu Bicara untuk Ponsel Anda?
- Not yet a member? –> Belum jadi angota?
- Sign up for free at –> Daftar gatis di
Trados 2011 and other CAT Tools may be able to detect these clear typos (Butu, angota and gatis, each should be Butuh, anggota, and gratis), but they will be powerless when they meet the ‘typos in context’. What we mean by ‘typos in context’ is typos that are actually not typos as they are accepted and found in the official Indonesian dictionary. You see these words are not mistyped, rather than you see that they have nothing to do with the context. If you use them, the translation will be absolutely wrong. This kind of typo cannot be detected by any machine, but you as linguist can detect it. It is the reason why manual final-checking is always necessary. Trados 2001 and other CAT Tools only detect the typos on the surface level, rather than the context level; thus machine can never replace the competent human resource to date.
Below are typos in context we mean:
- We collect information from you in various ways when you use the Service and Apps. –> Kami mengumpulkan informasi dari Anda dengan berbagi cara saat Anda menggunakan Layanan dan Aplikasi.
- Web beacons are small, invisible graphic images that may be used on the Service or in emails relating to the Service to collect certain information and monitor user activity on the Service, such as to count visits, understand usage effectiveness and to tell if an email has been opened and acted upon. –> Web beacon adalah gambar grafik kecil dan tidak terlihat yang dapat digunakan pada Layanan atau di email yang terkait dengan Layanan untuk mengumpulkan informasi tertentu dan memantau aktivitas pengguna pada Layanan, seperti untuk menghitung kunjungan, memahami keefektifan penggunaan, dan memberi tahun apakah email telah dibuka dan ditindaklanjuti.
Berbagi, tahun and pertanyaan are listed in the Indonesian dictionary and they are Indonesian standard words. Any Cat Tools can never detect them as we are sure they are listed in their local language database. If those three typos are missed out, the back-translation will be:
- We collect information from you in sharing ways when you use the Service and Apps.
- Web beacons are small, invisible graphic images that may be used on the Service or in emails relating to the Service to collect certain information and monitor user activity on the Service, such as to count visits, understand usage effectiveness and to give you year if an email has been opened and acted upon.
The source would actually say in various ways instead of in sharing ways, to tell instead of to give you year and privacy statement instead of privacy question; it is absolutely a big mistake and the only solution is a thorough proofreading by you. You cannot absolutely depend on CAT Tools you use. There is specific portion that needs your extra care.
For more info about typos, please visit http://en.wikipedia.org/wiki/Typographical_error. Please also note that we have no business correlation with any CAT Tool providers, but so far we use SDL Trados Studio for our trusted spell-checking machine and we are satisfied with the feature. If you have more trusted and better spell-checking machine for Bahasa Indonesia, please share here.