I'd like to create simple bash script that, given a directory, compresses each directory by name, e.g.:

Contents of ~/Documents
Folder1
Folder2
Folder3

compress-subdirectoies.sh ~/Documents

Results:
Folder1.[some kind of compression like bzip2 or zip]
Folder2.[some kind of compression like bzip2 or zip]
Folder2.[some kind of compression like bzip2 or zip]

Any advice would be appreciated

Since you have BASH, I presume you have linux.

find . -type d -mindepth 1 -maxdepth 1 | while read DIR
do
        echo tar -zcf "$DIR.tar.gz" "${DIR}"
done

Try this...

find . -type d | xargs -I {} tar -cf {}.tar {}

That make zips for folders inside folders, not just zips for the outside folders.

mmm... maxdepth and mindepth is missing in my code...
I just created one level of folders and tested it... didn't realize that there can be subfolders too...

find . -mindepth 1 -maxdepth 1 -type d | xargs -I {} tar -cf {}.tar {}

Wow, thank you both. Gosh, I love the "little pieces" way Unix works. Here's what I'd change:
1. move from plain text to XML as the transfer method (hidden from user) as the default communication method;
2. tell the Java people they've got the right idea but they didn't implement it right - building their apps is composed of lots of pieces (good) that are almost impossible to configure to work correctly directly (bad).

We have a single column of data here. What value does XML add? Imagine a pipe chain failing because you got the name of a field slightly wrong in one of your commands. xml isn't appropriate for streaming, besides.

I can only see the value of XML for data which isn't easily stored in a simple, tabular format.