Hacker News Reader

Git doesn't violate this in any way. Read the claims. They're talking about distributing data throughout a network of servers, using an hash as a key to know from which servers to request the data. It's more similar to database sharding.

So Memcached then...

> They're talking about distributing data throughout a network of servers

Hmm... We used that to shard the MySQL databases for our WordPress setup too...

> This sounds alot like hashing the contents of a file to get and identifier for it

Oops... I think the image asset deduplication method I used at the portal I worked for infringes on that. It's a Brazilian company, so, good luck for them.

Yep... They still use it:

http://i0.ig.com/bancodeimagens/02/9a/e2/029ae2zxd9u8jflt2le...

Document and image management systems have been doing this sort of thing for decades. Form a unique identifier for an image by hashing, split the bytes out into a path name by hex converting, and you have a file system path.

I'd say that this particular troll is going to get stepped on pretty hard, and all of these patents are going to be invalidated.

I never claimed I invented it ;-) It just seemed like the obvious way to prevent duplication when someone would rather update the same file again than search for a previously existing version.

In the end, when a dupe was uploaded for the second time, the uploader would be rewarded by not having to edit any metadata - as they would be redirected to the original asset page.

If someone ever uploads a second, valid, JPEG that collides in md5-space with a previously uploaded one, the sysadmins in the app group will receive an e-mail commemorating the fact someone defeated our deduper.

Does it know from which of a pool of servers should it get the data, based on that identifier? If not, it doesn't violate the patent.

Isn't it pretty obvious that they would avoid suing a larger company like Amazon because they would be able to afford a good defense. A startup, on the other hand, is much more likely to settle to avoid spending a lot of money, even when it may seem that eventually they would win.

Rackspace is no startup. They had more than a billion in revenue last year and they've been around since 1998.

I doubt jurisdiction was a deciding factor in who to sue. But I hated Civil Procedure so I could be wrong...

Jurisdiction is a huge factor when filing suits. IP lawyers do their best to move cases into Tyler, TX (or other backwoods cities) where the judges don't understand technology whatsoever. A popular tactic is to shake down porn companies for money using nuisance patent suits.

Do you really think it is because they supposedly don't understand technology? I've read that they are simply particularly "sympathetic" to patents over there, which is why there are a host of patent trolls / NPEs that have their mail address over there. This American Life had a good episode about it.

A close friend of mine defends major smartphone companies as well as other tech companies. He's told me the average presiding judge has no idea how to connect to the Internet, much less understand JPG compression (an actual case my friend worked on).

It's probably both.

I don't think the judges over there are substantially more ignorant about technology than the average judge in Kentucky or Montana, etc. They can call in experts to advise on technical questions. I'm afraid technical knowledge wouldn't help much, they have this cottage industry of patent lawsuits and they'd probably prefer things to stay as they are.

Another popular tactic: Suing porn companies to be granted standing (all the conservative judges in the South will gladly hear cases that cost porn companies money), then add defendants that you want to shake down. In the past, patent trolls have done this with Google/Microsoft/etc, but apparently these large companies have been fighting back or (more commonly) lobbying to get patent troll applications violated, so the nuisance patent trolls tend to stick to small companies they can easily shake down for a $20,000 "licensing fee."

And you think that jurisdiction is the reason they chose to sue rackspace instead of github?

I've head from lawyers that East Texas became the place to patent lawsuits mostly because the War on Drugs. It was much easier to get the courts' time for civil trials or motions before they were overloaded with constant criminal prosecution. East Texas happened to have a lower load at the beginning or some townships spotted an opportunity to bring in some more business by making themselves attractive to litigators. Either way, the reasoning things are filed in East Texas has more to do with it being the Silicon Valley of patent trolling than any kind of venue shopping.

> If anything it sounds like maybe Git itself violates this, but I thought that Git hashed the difference between all the changes of a commit to get the hash. Some of the others sounds equally obvious at first glance.

Git creates hashed objects for individual file revisions, as well as tree objects referencing many individual file objects and commit objects which reference those tree objects.