Red Hat Sued Over Hibernate ORM Patent Claim

fmarines writes "Firestar Software has filed a patent claim against Red Hat for infringing on a patent Firestar filed in 2000 covering O/R mapping. The amount of the lawsuit was not disclosed. The complaint centers around JBoss 3, and the patent claims that JBoss was given prior notice that marketing, distribution, and support services violates Firestar's patent, and that Firestar 'has suffered and will continue to suffer substantial damages.' Firestar produces the ObjectSpark, an transactional object mapping engine which appears to not have had a new release since May 2003, according to the Firestar press release page."

Comments

TFA seems confused

[Brian Blessed]

It says:

The complaint centers around JBoss 3

JBoss 3 was released in May 2002.
However, Hibernate wasn't a JBoss project until September 2003.

I'd guess that the claim relates to Hibernate 3, but they are desperate to mention JBoss as much as possible for the FUD value.

Buy a company - get sued!

[Corrado]

This is really low - wait until your software gets in heavy rotation and then go after people using it. My question is why did they go after RedHat/JBoss? Why not go after the big dogs; Oracle (TopLink), BEA (Kodo), and even the JCP (EJB JPA)?

Again, software patents are a bad idea.

Re:before it gets slashdotted...

[dnoyeb]

Im not familiar with what techniques the new EJB is using or KODO. However, I thought the new EJB was going towards JDO technique. The patent seems odd in that it is patenting a technique which would be best called a combination of Hibernate and JDO techniques. AFAIK Hibernate performs its mapping at runtime based on the schema you create. I assume Hibernate uses mediator objects of some sort? JDO modifies its objects at compile time and does not do any runtime modification. This system has a schema, creates an interface object at compile time, and them uses a runtime engine to do somethign else. Overkill.

Maybe thats why nobody uese it?

This is the definition of an obvious patent

[chriseyre2000]

How can you implement an Object Realstional Mapper and not violate this patent?

Re:This is the definition of an obvious patent

[dsurber]

TopLink existed as a Smalltalk product in the early 90's, '93 or '94. I know because my company developed a Smalltalk OR mapping product that competed with TopLink.

It takes more discipline than I have to try to understand a patent, but I'd be astonished if the there isn't a ton of prior art, starting with TopLink.

Re:This is the definition of an obvious patent

[CatOne]

I worked for Persistence in 1996-1997. They had been around for 4+ years at that point, so their O/R mapping stuff dates back to 1993 or before. They had a number of patents on O/R mapping at the time, and later got more patents for their "caching" algorithms which seemed a bit more vague.

They were acquired by Progress Software in 2004 or 2005... I haven't seen any lawsuits from either of them.

But this is really curious, as things like O/R mapping have been around for a very long time... heck the EJB specification itself, which dates back to 1999 or so, includes O/R mapping as a part of it... entity beans must have persistence.

This infringement lawsuit then *must* hinge on some very specific technology pieces that have been violated. I mean, otherwise, as bad as the patent process is, if they missed the fact that the concept has been around with prior art for 10 years, and with specifications that mandate it for 5 years... talk about a joke.

A shot across the bow

[Anonymous Coward]

Patents allow wealthy companies the opportunity to appropriate the profits from the work of others for themselves. I don't trust JBoss and while I don't see conspiracies around every corner, I lack the niavety to believe that such strategies aren't being actioned at arms length by certain tech companies.

I expect this to be the first of many claims :-(

Is Ruby on Rails Affected?

[Corrado]

Just took a quick read of the patent and it looks like this is a pretty wide reaching patent. Anything that maps a database to an object is covered by this one. Does this mean Ruby on Rails is under the gun? Of course, they have no money (I guess) so they wouldn't be a target. But they would still be violating a patent and that could limit their future growth potential.

Re:Is Ruby on Rails Affected?

[Anonymous Coward]

One word Superbase, sold by Digital Research well over twenty years ago for MS&DRDOS based system. It's been updated some since then and is still being sold. At the time I thought it a better program than DB3 and Paradox. I'm not a programmer yet, suberbase appears to incoporate everything this patent was granted for prior to it being granted.

Sue them back into oblivion.

[lowieken]

Firestarter Software is probably not doing very well. Why else would they launch a software patent lawsuit? On top of that, they have actual products in the market. Seems to me that this makes them very vulnerable to countersuits.

Am I right thinking it shouldn't be too difficult to sue Firestarter Software into oblivion?

Prior Art?

[Phil John]

Surely there's a wealth or prior art for this kind of thing, ORM was popular before 2000. What about Next Computer's Enterprise Objects Framework? That's been around since at least 1994 according to WikiPedia - it still lives on as part of Apple's WebObjects system.

Time for Red Hat to leave the USA

[Alain Williams]

Relocate to Europe where we don't have this patent lunacy[**]. These patent trolls would then be limited to trying to steal the USA turnover of Red Hat; Red Hat could perhaps take the option of abandoning the USA market. If Red Hat were to leave the USA it would send a strong message to congress how patents damage the USA economy ... they might even decide that doing right by their country is preferable to accepting the slush funds from the patent lobbyists.

They have some nice offices here, no language problems for existing staff if they move to Guildford (UK).

[**] - OK -- I know that some are trying to introduce it, but the EU seems to not be that stupid (fingers crossed)

Prior Art

[bbroerman]

At my company, we wrote software that basically does the same thing back in the late 80's - early 90's and have been using it for high-profile, high-cost software ever since... While I don't know if we ever applied for a patent on the idea, I would bet that there is a lot of prior art out there... I just hope that Red Hat's lawyers are good enough to find it and use it appropriately.

Re:This kind of "lawsuit inc." business needs to s

[Jerf]

You either license to everyone you intend to allow use of your patent or you lose it.. you should not be allowed to hide in wait and opportunistically/arbitrarily ambush companies and developers.

That would be disasterous. You forget to account for the number of companies who do real work and hold patents for MAD purposes. Give those companies lawyer's a choice between "enforcement" and "losing the patent", and a significant proportion of them will choose "enforcement".

Enterprise Objects (WebObjects)

[Jimithing DMB]

Enterprise Objects certainly implements everything described by that patent and a bit more because it provides a data access controller layer (not just a data model layer). Not to mention I've had some limited experience viewing someone else's hibernate-based code. EO/WO is so much better than Hibernate can ever hope to be.

http://helpredhat.dyndns.org

[Anonymous Coward]

Hi,

I created a small website with MediaWiki, which is dedicated to collect Prior Art against this patent. This will help Red Hat and might prevent the same patent from beeing issued in Europe, Canada, Japan and other countries.

Let's show the world that this wasn't a new invention in 1998 !

http://helpredhat.dyndns.org/

Prior art - Enterprise Objects Foundation

[sp67]

NeXT released Enterprise Objects Foundation (EOF) in 1994 (http://en.wikipedia.org/wiki/Enterprise_Objects_F ramework), and the patent summary doesn't mention any features EOF didn't have.
Isn't this prior art?

Actually, it may be broader than that

[hey!]

The patent either covers hibernate, in which case it covers all object oriented programs which store objects in relational databases, or it covers neither. You can't patent an OR mapping product without making claims on any program that has this functionality, even if it's using a hand coded data access objects (DAOs). You'd be required to hand code database queries into objects that have nothing directly to do with databases, other than being stored in them.

The critical claim, it seems to me, to be this:

10. A computer program fixed on a computer-readable medium and adapted to operate on a computer to provide access to a relational database for an object oriented software application, comprising:

a mapping routine that generates a map of at least some relationships between schema in the database and a selected object model;

a code generator that employs said map to create at least one interface object associated with an object corresponding to a class associated with the object oriented software application; and

a runtime engine that invokes said at least one interface object to access data from the relational database.

This describes to me a functional description like this:

object/data_model -generates-> O-R_mapping -generates-> code_for_interface_object

and

code_for_interface_object -used_by-> runtime_engine

This doesn't seem to fit hibernate at all. You can cobble togther pieces of this, such as automatic generation of the database schema from the object model; code generation from a schema. But hibernate isn't a "runtime engine" for accessing "interface objects". It sits directly between YOUR objects and the database. I'd make the distinction like this:

domain_objects --managed_by--> runtime_engine [--using--> interface objects,--against-->database]

vs.

domain_objects --managed_by--> hibernate --against-->database.

Claiming this is covered by the patent would also require claiming this is infringment

domain-objects --managed_by--> your_code --against-->database.

The only way to avoid this is to hand code database queries into every object:

domain_objects --manage_themselves_against--> database