> Well, this is an interesting question, although I don't know what is
> the difference between BINARY and BLOB...

probably, the BLOB has infinite capacity, while BINARY has some limit
specified during schema definition.

> The interesting thing is the following: if you look at the page 181
> and 182, you will see that BLOB should not be accessed by neither
> getString nor getBytes, moreover you should not pass a String or a
> byte[] to a Blob. So the current behaviour of the FBBlobField is
> much more similar to a BINARY field rather than a BLOB field with
> the only exception that it returns a Blob instead of byte[] ;)
> Actually I can say in the current implementation the FBBlobField
> *does* implement the BINARY (VARBINARY, etc...) type, not the BLOB
> type, and in my opinion the driver should be consistent in this
> question.

I completely agree with you that driver has to be consistent in this
question. I will try to fix this problem.

Best regards,
Roman