The only use case I can think of for redo would be for operations like writing to a socket or reading from a database, but if these fail once, subsequent attempts will most likely also fail so it still seems a bit pointless to me and as for retry I can't really think of any case where it would be useful.
This might only seem meaningless to me since I don't know or use Ruby, but I aspire to create an awesome language one day so I would like to at least know the reasoning behind the design of some of the most popular languages out there.
The idea is that you change something before calling redo or retry, in the hopes that the whatever you were doing will work the second time. I don't have an example for redo, but we have found uses for retry in the application I'm working on. Basically, if you have a bit of code that might fail due to something external (e.g. network), but performing a precondition check every time you run the code would be too expensive, you can use retry in a begin...rescue block. Not sure if that was clear, so I'll get right to the example.
Basically, we have some code that accesses a remote directory using Net:SFTP. The directory should exist, but in some exceptional cases it will not have been made yet. If it's not there, we want to try once to make it. But performing the network access to check if the directory exists every time would be too expensive, especially since it's only in exceptional cases that it won't be there. So we do it as follows:
tried_mkdir = false
begin
# Attempt to access remote directory
...
rescue Net::SFTP::StatusException
raise if tried_mkdir
tried_mkdir = true
# Attempt to make the remote directory
...
retry
end