The version I am using is uncompressed. I wouldn't use the compressed one since it only saves 2k over the uncompressed version and it would be a pain to modify. The uncompressed one is still half the size of the current one and much faster.
Yes, we could try fixing what we have. It was easier though to get an updated script that works and is effectively a drop-in replacement of the old one. That's because I don't know enough about the subtle differences between the browsers to figure out why it works in some but not others.
I'll see if they're willing to let us include it in OpenACS.