How-to use the dataset

I'm Hoang Thanh Lam - a Vietnamese student at Moscow State University Russia. I'm going to apply particle swarm optimization to clustering data problems. I found that your datasets are very useful for my project. But I have some questions concerned with how-to use these datasets, hope you answer soon, thank you very much:
1/ The 15 datasets described in you article: "Criterion Functions for Document Clustering
Experiments and Analysis" are quite different from the datasets downloaded from your website at URL: http://glaros.dtc.umn.edu/gkhome/fetch/sw/cluto/datasets.tar.gz. For example the Fbis in the article was described with 12674 terms while in file fbis.mat there are only 2000 columns, this means, there are only 2000 terms in all documents. Why are there different between them?
2/ In my knowledge, an element of a sparse matrix is a pair,in which the first one is a column's number (term's index?) and the second one is a floating point value. You didn't say in detail what the second value is, but I guess that It is the number of appearance of a term in a document, isn't it? I need this information because I want to build a TFIDF vector for each document.
3/ Were all the documents preprocessed to remove the common words and to stem by the Poster's stemming algorithm?

I'm looking forward for your answers. Thanks again!