|
Well ... It was filed for in 1991. At the time there was no prior art as far as he could tell (certainly the U.S. patent office has been known to make mistakes though). My dad talked to Sun about it in the summer of 1992, and claims that this had an effect on Java as we know it today. You can see the patent at: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=36&f=G&l=50&co1=AND&d=ptxt&s1=pickett.INZZ.&OS=IN/pickett&RS=IN/pickett I do think though that there's a difference between this patent and what Java does. Java has Remote Method Invocation (RMI), which lets you execute methods in classfiles remotely, running on a different virtual machine. Does Java also allow you to start executing a program even if all the classfiles are not present on your machine (or LAN), and there are no other VM's running? (I should be embarrassed not to know these things, working on a VM and all, but I'm not doing this kind of stuff) Anyway, I'm not claiming that nobody else does this. Certainly Microsoft appears to, but it's difficult to get details out of them. Java maybe, still not sure (who can I ask?). I've not seen any projects that do this for compiled code. I don't really know what to tell you much more than what I've already said and what the patent description says. I'd be grateful for evidence (patents, projects) showing that it's not unique. I suppose what I would say is most unique is the application of the patent claims. Imagine a network of code servers, and clients and "sources" connected to this network. Code servers are also running on the clients and sources. When you start your application, your local code server checks your machine to see if the module / shared object / library / DLL / classfile / whatever is there. If it cannot find it, it asks the closest code server on the network. These requests keep happening until the source machines are reached, and the code gets distributed over the network and finally ends up on your machine (much like how you ask your ISP for a web page). When an update is available for code, any code servers that have requested this code are notified, and next time you start your application the new modules are automatically downloaded for you and installed. You could also use it in a subscription service (I realize this is probably not of interest to you folks) whereby a certain piece of your program expires monthly and you have to pay to get a new one. The above may sound a little naive (I am not a networking or installation person) but I still think it holds water. It stops you from having to download binaries by hand and install them. I even think you could use it for uncompiled source somehow. Redhat's up2date or whatever it's called seems similar but is different because it is not invoked every time you start an application. Okay, that's all for now, any further leads on this would be great (or you can just tell me it's rubbish and I'll go elsewhere..............) Cheers, Chris Mark Mielke wrote: How about a query regarding the validity of the patent? Do you think people haven't done this sort of thing for decades now? How many other 'remarkably similar' patents do you suppose might exist? What is unique about this patent? mark On Wed, Sep 10, 2003 at 03:51:08PM -0400, Chris Pickett wrote:Hello, I have some questions about a code server patent belonging to REC Software Inc. (my dad's company) and I was wondering if somebody here can help me out. It's a patent that describes dynamic (and automatic) software installation over a network. You start executing some module, and it goes out and finds the other pieces you need (based on the existing "check every piece is there before executing" process) and installs them in the background. It can also be used for software updates and software subscriptions. This isn't remote execution we're talking about, the code actually runs on the client. My questions are: 1) Given that you guys are installation people does this sound at all useful? 2) Who can we talk to in more detail about this? I know software patents are considered evil, the possibility of giving this thing away exists, but we'd need some convincing that it would be worthwhile (in a specific context, not general "it's worthwhile to give any patent away" stuff). Feel free to send me a personal email or respond to this list -- either is fine. I thought rpm would be a good place to start asking questions. I have some familiarity with GNU/Linux (I work on an open source Java VM called SableVM), my father has until now (!) been a Windows developer. I have no familiarity with the technical details of rpm, but have used it on occasion (Debian is the distribution of choice for my university). Cheers, Chris Pickett _______________________________________________ Rpm-list mailing list Rpm-list@redhat.com https://www.redhat.com/mailman/listinfo/rpm-list |