Over the past several weeks, Tom Lane has been working on replacing our old Cache Management Alorithm (ARC) with a new, patent free one (2Q). In order to reduce the number of 8.x deployments out there that are using the old manager, they have just released 8.0.2, and encourage adminstrators to upgrade at their earliest convience.
PostgreSQL 8.0.2 released with patent fix
2005-04-12 Databases 10 Comments
Why is this done for the 8.0.2 release, was it not knows when they released 8.0 I’m surprised something like this could happen, is it not ‘dangerous’ for companies using 8.0, they’d be liable for using this code?
The patent in question has not been granted to IBM as of yet, it’s still pending, so they can’t really do any legal action.
So it is really not dangerous, they just did it as a pre-emptive measure.
Which was a nice move if you ask me.
The code that was newly introduced in 8.0.2 performs on the same level as the IBM patented code and well, it is likely to be dropped out of 8.1 as a completely new locking mechanism will be introduced.
Which begs the question: If the patent has not yet been granted, doesn’t it follow that it should not be given the obvious prior art?
Welcome to Earth, you must be new here.
If the patent has not yet been granted, doesn’t it follow that it should not be given the obvious prior art?
prior art only counts if it’s before the application, not if it’s between the application and the grant …
Two totally independant parties came up with the one concept, and because of the patent system IBM are the winners?
What a joke.
I’m not entirely sure that’s what happened here, but yes, that is how patents generally work.
However, if the `second’ party published their concept before the `first’ party applied for the patent, the patent would not be granted due to `prior art’.
I believe in America there is a 1-year `grace period’ in which an inventor may publish his own works without that counting as `prior art’. In Europe, as far as I know, that is not the case.
read all about it at http://www.varlena.com/varlena/GeneralBits/96.php
Even if IBM does receive a patent, it’s highly likely that they’d grant perpetual license to PostgreSQL anyways — there are a number of very good legal and PR reasons for that to be the case.
Kudos to the PostgreSQL team not only for due dilligence in avoinding patent issues, but also to a damn fine database product. I’ve worked with Sybase, Oracle, MySQL, Access, MS-SQL, and PostgreSQL over the years and I feel that they’ve hit that sweet-spot of performance, functionality, and ease of use that makes it the best general-purpose database out there. Obviously, there are situations where it’s not called for, but absent special requirements, I find PostgreSQL to be best-of-breed.