8 I have used both cluster3.0 and pycluster. Overall, it's easy to use, provides basic functionalities for data analysis and cluster3.0 can also let you visualize the results. A few drawbacks: speed wise, it's fairly slow for mid to large data set; visualization isn't great for multi dimension data sets; limited parameter settings in each machine learning algorithm. I would recommend to use it if your data set isn't too large and complicated.

5 It's a good tool to tell apart some western languages like latin etc. However the accuracy is fairly low when classify Asian, Arabic, Russian, and eatern european languages. So if you need a tool to distinguish between English and Spanish, it's okay. But when it comes to Asian languages, for example, it's not a reliable tool at all.

10 Weka is a great data mining tool. It has various data mining algorithms implemented and fairly easy and intuitive to use. For results analysis, it provides evaluations from different aspects. It can also let you visualize the results, which is great. Java API makes it easy to incorporate into people's projects. Highly recommend it for data miner!

5 I feel it's not very good to use. First, on the buddy list, it doesn't differentiate which IM the buddy's on. Sometimes it causes confusion. Second, it doesn't give you the option to have new messages coming in as separate remind window. So for some not important messages, you don't have to switch to the actual window to see what it is (you can just peek it when it shows up at the corner). I still like trillian more in that regard, however, trillian doesn't support gtalk, which sucks.

8 It's a great tool to have asterisk working with ruby on rails. It simplifies the process of integrate a free pbx with a web application framework. The only drawback is that in some cases, debugging is not that easy.