I thought I would try something different here. What I’m going to do is use Snowball, which I mentioned in my last post, as a new CAT tool, for the very first time, and I am going to blog about it as I go.
So here I go over to the snowball website to download the program and set it up. It is now 3:20 PM.
It is now 3:23 PM and the software is downloaded and installed and registered. Now I am going to read their quick start guide from the website.
It is now 3:26 PM and I have read the quick start and the cheat sheet and, after making myself a coffee, I will come back and start using the program.
It is now 3:40 PM and I have figured out Snowball, been impressed by it, and realized that I cannot use it. I must say, that it was extremely cool to see that it offered useful translations of the headings in the patent right out of the box. It took no time whatsoever to be able to start using this application and for it to be useful. I would definitely recommend it to anyone who was going to translate a European language. Unfortunately, it was not possible to segment Japanese. In the first sentence, it went along until it came to the first katakana “bo” and I could extend the segment to the other side of that “bo” or to the end of the sentence, but not to individual words. Snowball was obviously looking for the spaces between or what it considered to be punctuation and is not capable of dealing with languages that do not use spaces. But perhaps in the future this problem will be overcome (all they have to do is get it to recognize a switch from kana to kanji, which is what MS Word does and is pretty simple from a programming point of view) in which case I will gladly give it another try.
There was, however, another reason why I decided not to go any further with my explanation of Snowball and that is that the source text is entirely overwritten, and disappears. That means that, when you go back to check over your translation, or when, on page 3, you change your mind about the way page 1 should have been translated, you would have to go back to the source document to find the source text. Side-by-side presentation of the source and target texts, and the ability to search for both source and target in the translation is 75% of the utility of a CAT program, and Snowball, not having this, is not really going to cut it.
UPDATE: I sent a link to this post to the folks at Snowball and immediately received a very friendly reply. A fix to support Japanese seems to be in the works and there is also a possibility that source text access may be handled differently in future versions. It looks promising. If anyone is using this tool and has blogged about it, please leave the links as comments or otherwise just let me know. I can’t help feeling this one has a lot of potential and I am eager to keep abreast of it.