Re: How to make lock escalation look good...

norman.dunbar_at_lfs.co.uk (Norman Dunbar) wrote in message news:<3078e2fe.0305210552.a8b3ee7_at_posting.google.com>...

> And another thing, this locking mechanism is patented - which makes me
> wonder how exactly they managed that. My Sinclair QL uses exactly the
> same linked list system when multiple jobs on the system need access
> to a resource. One job has it and when done, flags the next job that
> it can have it.
>
> This, in a home computer built in 1982 for crying out load. Can
> American companies patent stuff that has been around for years then ?
>

Hehehe! What you have to understand is that DB2's locking mechanism was "cooked" back in the early 70's, when the product first saw the light of day. It used EXACTLY the same locking mechanism as page-based Codasyl or hierarchical mainframe databases, such as IMS. Those had their lock system patented a long time ago. I'll bet you anything it was the same code. They even shared the same VSAM PIOCS file structure.

The "row-locking" they now claim as "patented" is just an extension of that old mechanism. That's why the thing is half-baked and needs the lock escalation to survive: it was never designed from the start as a true row level locking system.

As for patents: did you know the simple bitmap-XOR operation that allows a mouse to appear to move across the pixels in a screen is actually patented? There is an ex-Park guy making a motza out of that one...

Talk about a deranged patent system!

Cheers
Nuno Souto
wizofoz2k_at_yahoo.com.au.nospam Received on Wed May 21 2003 - 20:41:52 CDT