Another issue with SCSI tape interfaces are in the details. Al has pointed out the lack of data returned with error conditions. There are also niggling little differences between implementations that don't exactly follow ANSI X3T10. For example, is a tape mark considered a block? I've got drives that go both ways--one says it isn't and so doesn't increment the position (I'd call that an error) and other says it is.

It's mostly the problem of adding another layer of someone's interpretation to another older protocol and hoping for something coherent to come out.

--Chuck

Reply via email to