Big file systems make computers run more slowly. This problem is most noticeable for laptops. On a project that I am currently working on, I am using multiple versions of three frameworks: The Flex SDK (source and ASDoc), the MicroStrategy Flex SDK and the BlazeDS source tree. The result is thousands of files, which take hours to copy and slows down disk access even when I’m not programming.
Because these files are only read and never written, the solution is to not unzip them, and to only access them from within the zip files that they were provided. This facility is called a “local loopback filesystem” on Linux, and “accessing compressed folders” under Windows.
Eclipse/Flex Builder has native support for accessing compressed folders as if they were regular folders, but that functionality is not fully exploited. Ant tasks need O/S level support to enjoy this feature.
Results are dramatic. Usage is very simple. What’s not to love?
A client's IT department told me that they did not support computers used in that fashion. My initial response was somewhat uncomplimentary, and included a statement to the effect that I did not need their help anyway. Later I modified my statement to say that so long as they attempted to enforce idiotic rules which prevented me from working effectively, that no deadlines would likely be met from that point forward, and costs would be without limit.
On Windows, I use the free Pismo File Mount Audit Package. It provides the ability for zip files to be mounted as virtual drives.
On Linux, use fuse-zip. It provides the ability for zip files to be mounted on any directory mount point.