Slashdot Mirror
Microsoft Patents 'Phone-Home' Failure Reporting
theodp writes "On Tuesday, the USPTO awarded a patent to Microsoft for its Method and system for reporting a program failure, although a much more sophisticated version of this technology has been standard on IBM mainframes for years. Maybe prior art searches will improve once the USPTO moves into the new two million square foot USPTO campus, which includes five interconnected buildings, a twelve-story atrium, a landscaped two-acre park, and a museum."
Netscape has done this for years.
The difference between Canada and the USA is that in Canada healthcare is a right and gun ownership is a privilege.
The abstract is not the real patent. it is only there to make skimming them easier, and judging the value of a patent by its abstract is like only looking at the title of a book before judging it
Netscape is mentioned in the text of the patent:
To gather more information about a crash, different approaches have been taken. For example, America Online has the ability to determine the location of a crash of Microsoft's "INTERNET EXPLORER" web browser and report this information to Microsoft. However, other information regarding the state of a user's machine at the time of the crash is not known and it is difficult to distinguish between different crashes. Without this valuable information, not much can be done to determine whether there is a bug and, if so, correct the bug. Other approaches have been taken to tackle the problems of failures. For example, Netscape's "COMMUNICATOR" web browser includes a quality feedback agent to report a crash to Netscape, although it is not known what information is reported to Netscape. Moreover, both of these approaches apply to fatal crashes, i.e., when the operating system has decided to kill the application. Thus, there is a need for a method and system for reporting non-fatal crashes, such as when the operating system continues executing the application's code.
Patents are not about who is right, or who is first; patents are about who will sue.
The US PTO is a money-making service for the government, and this fact is why it operates as it does.
There is a misconception that it is the central duty of the PTO to form a blockade against granting patents. The PTO can and will block applications where there's heavy similarity with prior art or existing patents, but that's really just a guideline to using the service, not the core function.
The meager regulatory behavior also weakens further in tough economies, because Big Business believes that having patents, even if they are untenable, will generate revenue; the administrations can open the floodgates at will.
The PTO's purpose is to grant patents for a fee, and it's wholly suited to do so.
The application vetting process of the PTO is a cost center for the operation of the PTO. This is akin to saying that customer service is a cost center for the operation of AT&T. It is required, but they'll cut costs as much as they can get away with.
To fix the patent application vetting process, two things must happen:
As of 15 March 2001, the USPTO has changed their policies to solve that second problem. They can now publish patent applications before the patent itself is awarded to the applicant. Third parties may now submit "helpful" arguments against controversial applications. The USPTO can then weigh obviousness against challenges without incurring the costs of doing all the searching themselves.
Breaking patents by finding simple prior art is not enough for most cases. Patents already granted are almost never cracked, certainly not by someone using an independent third party's prior art. In the famous Heinlein/Waterbed case, the patent was denied before it was ever granted by the Patent Office. Once a patent has been granted, the Patent Office rarely will get involved in disputes; that is a matter for the courts.
[/RANT]
[
I know for a fact that on the pSeries (Unix) and zSeries (Mainframe), the system includes a separate service processor that does just exactly that. So even if there is some sort of catastrophic failure of the main system, the service processor sees it, phones IBM, and their service guy shows up in an hour with the exact part that needs to be replaced. I think I would call this prior art.
Here is the list of all withdrawn patents since 1790: withdrwn.txt
Shouldn't the Amazon one-click shopping patent (# 5960411) be on this list if it had been thrown out?
Three words: IBM AS/400 ECS
I should say iSeries ofcourse. I'm not sure when ECS was added, but it was in atleast OS/400 V3R4 and later. It does differ from the MS system though, it detects hardware faults too.
Detects hardware/software faults
Calls IBM
Downloads/orders PTFs
How is that NOT prior art?
"...process will try various methods to push failure to server with the first being dialing out to a end user paid 1-900 number..."
I found this document about MS Patents.
Phone home itself is an obvious application.
MS was granted a patent on a very specific phone home -implementation-.
I hate bad software patents as much as the next guy, and that's why I took it upon myself to get educated on the process and language; and that's why I -read- these patents.
Their patent covers a phone-home architecutre that:
. detects the failure (via exception handling)
. locates the source of the failure
. -asks the user if they want to allow phone home-
. phones home to a repository (if allowed)
. looks up the failure in the repository to determine if there is any request for additional information to gather for that particular type of failure (or by particular application that failed)
. gathers the requested additional information from the failed machine
. transmits all the desired information to the repository.
. searches the repository for any existing fix for the problem
. transmitting the fix to the failed machine and applying it
. if no preexisting instance of the failure exists, creating a new instance
further, the repository can be local or remote, as specified in a setting located on the client machine.
If IBM's system has done all this prior to microsoft's - then fine, it's prior art. But in my experience, that is well beyond the capabilities of IBM's 'phone home' system.
If you make a phone home system, so long as you don't -exactly- copy their implementation it's fine.
If you even didn't offer the user a choice to phone home, that would be enough of a deviation from their patent to protect you. Their patent is that narrow.
It's that easy to work around it. They're just protecting themselves from direct copycatting as a matter of routine.
// "Can't clowns and pirates just -try- to get along?"
Crashes that are reported to Microsoft via this system (known internally as Watson, not Dr. Watson, BTW) are "bucketized" by dll, version, module, offset. No PII is collected in the basic crash reporting.
Developers often use the crash bucket analysis to focus on high-hit buckets, but other kinds of analysis can be (crashes per module, etc).
A developer or tester can ask for regkeys to be sent back, files off the machine, or a minidump. If those things are asked for, then a different, more detailed user information dialog ("additional information is required..." is presented for ok by the user.
A survey can also be requiested, which navigates the user's default browser to a survey that asks "what where you doing", asks for (optional) phone number, email name, etc.
A bucket can also be can also return an OCA (Online Crash Analysis) URL back to the user, when a known crash cause is found. This fires up the default browser to that URL, which usually has something like "The cause has been identified. Install Windows XP SP1" or the like.
Given the millions of buckets, it's not surprising that you've never seen an OCA response. But thousands of buckets are set to return them.
PTO rarely gets involved because there are rarely any requests for re-examination (I think only a handful so far this year).
Legislation was passed last year to encourage more non litigation challenges to existing patents. Here is an informative piece detailing what more needs to be done.
The specific innovation in this patent is that after the client "phones home", the server can request additional data. I didn't see that in a cursory scan of the linked IBM paper.
As usual, slashdot is making a patent sound broader than it is.
I am one of the two people who inveted Talkback (aka the Netscape Quality Feedback Agent).
We showed this technology to Microsoft early in 1998. We outlined the whole process, and nothing in their abstract describes any innovation over what we showed them.
The discussions in that meeting were held under NDA. As a small company, we had little choice but to use Microsoft's NDA. Their version allows them to use anything they "remember" as long as they don't disclose our specific confidential information.
We did get a patent for Talkback, one of several that were in the works before we sold the technology. Supportsoft now owns the technology and the patent.
I'm not saying that this is just another example of Microsoft having to rely on screwing small companies and coopting their innovations. You can decide that for yourself.
Haven't Netscape and Real Player been done this for several years now?
Fred
"A fool and his freedom are soon parted"
-RMS