Amazon Patents Customized 404 Pages

A patent was awarded to Amazon CEO Jeff Bezos on Tuesday for his "invention" of Error Processing Methods for Providing Responsive Content to a User When a Page Load Error Occurs, which covers displaying alternate web pages in response to HTTP 404 page-not-found errors.

View: Patent Abstract
News source: Slashdot

Report a problem with article
Previous Story

Mozilla ups unpatched Firefox flaw to 'high severity'

Next Story

Copy a CD, owe $1.5 million

18 Comments

Commenting is disabled on this article.

A patent don't grant all rights for generic/common used topics. So, most likely Amazon earned this patent but she is unable to used again other person.

So, i can patent "to breath air" but even when i earn this patent, i cannot earn any money from it, i cannot force to people to pay a royalty every time they breath air but, nobody can grant this patent.


In the end, it covers customized 404 exception handling. Retarded to let someone patent this. I have a few custom ways of security control through web-apps, linking databases and LDAP servers, maybe I should patent that! I mean really, c'mon now, who approves these things? Obviously people who know nothing about design and development. Letting someone patent the custom 404 is like letting some patent port 80... it's just not realistic or acceptable. How many sites are going to be affected by this. Many sites use custom 404 pages, I guess now they'll have to start removing them and just leaving the generic 404 error page showing...

Yes, it can be done on the server so long as your server knows where to go to either put up a custom 404 or other content. This scheme puts the onus on an error processing server that is notified by the client, presumably a browser plug-in. The error processing server handles it by either redirecting the client to substitute content, presenting a cached version of the 404'ed page, or a non-cached page that is similar to, but not substitute for, the originally requested content.

This actually makes sense. It solves a few scenarios of errors that a user may get that amazon would not be able to track otherwise. You have to understand that the pages served by amazon these days implement many 3rd party scripts as well as ajax calls to other services.

some scenarios:

1. A user loads an amazon page that makes an ajax call to another script from the client machine. If the script called by the ajax request was hosted by amazon they may get a 404 error in the logs of the server hosting the script. But that 404 may not give an amazon administrator enough information on the user experience or why that script was called in the first place. Only the clients browser actually knows what happened in this case. The client can report on what it was trying to do when it called that script and what the outcome of that script being missing or error was.

2. Amazon places a 3rd party tracking script or a 3rd party advertisement script on a page of their website. In the case that when a client pulled down that page the script they were instructed to execute were missing it could cause other issues, such as maybe some of their javascript code could fail. Again in this case only the client browser would really understand the user experience.

3. A user has a toolbar, pop up blocker or piece of spyware installed in their browser which alters the way that javascripts act and pages are rendered on websites. This could cause ajax requests to be redirected or could cause scripts to be executed that are not even part of the page that was requested. Only the client browser would know what was installed, what was manipulated and why it caused an error. So if a pop up blocker in a third party toolbar caused an amazon page to error, amazon would now receive a detailed report on that error, where as before they would have to depend on QA for this type of testing.

I can think of more examples but the theme is the same. Since amazon has incorporated much more javascript/ajax code into their site which executes on client machines they are most likely experiencing issues tracking how well this code is actually running on the client browsers.

The correct starbucks analogy would be something like:

1. A customer buys a coffee and itunes download from starbucks.
2. The customer types the code for the download into itunes.com and itunes replies that the code is not valid.
3. Since there is no way for the server at that time to know that that particular customer got the error. the customer would need to walk back to the server and tell him that his itunes download did not work.

That being i think this patent is ridiculously vague. According to this patent if I were to write javascript such as (below) on a web application I would be in violation of the patent.

try
{
//attempt to retrieve data from ajax request
}
catch e
{
//oops there was an error
//report error to authority

try
{
//display alternate content to the user from substitute object X
}
catch e
{
//wow there was another error
//report the error to authority
//display alternate content to user from alternate object y
}
}

So what has to happen now. Do sun micro systems and Microsoft have to remove exception handling from their javascript spec? Truthfully the fact that this patent was recognized was a mistake. If a suit were ever able to go to court i believe the patent would need to be revised and the outcome would simply be a waste of money.

"If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today...A future start-up with no patents of its own will be forced to pay whatever price the giants choose to impose." - Bill Gates ~ 1991 Employee Memo

At one time, Gates agreed with you. Wish he and others still felt this way.

Gates and the other majors agree, they just don't have a choice except to play the system as it is until it changes. It is just good CYA business practices to do so.

before anyone says it, this doesnt include standard fancy 404 pages, it is a technique whereby the browser is served a 404 page, then the BROWSER itself acts on that and notifies the server there is an error and gets a page acknowledging said error and giving details of what to do after back from the server

So, what your saying is that the server needs to be told by the browser that it just served the browser a 404?

That's like me walking into Starbucks, asking for a Double Strawberry Chocolate Coke, the waitress telling me that no such drink exists there, and me then telling the waitress that she just told me that no such drink exists.

Anyone else fail to see the point of this?

(dev said @ #1)
and notifies the server there is an error and gets a page acknowledging said error and giving details of what to do after back from the server

you mean like telling the server to send a ' spam'amazon advertistment?....lol

(Slugsie said @ #1.1)
So, what your saying is that the server needs to be told by the browser that it just served the browser a 404?

That's like me walking into Starbucks, asking for a Double Strawberry Chocolate Coke, the waitress telling me that no such drink exists there, and me then telling the waitress that she just told me that no such drink exists.

Anyone else fail to see the point of this?

it's not meant to be the server that served the 404, it can also be for generic page errors where some code has stopped working and, for example, a blank page is returned

(Slugsie said @ #1.1)
So, what your saying is that the server needs to be told by the browser that it just served the browser a 404?

That's like me walking into Starbucks, asking for a Double Strawberry Chocolate Coke, the waitress telling me that no such drink exists there, and me then telling the waitress that she just told me that no such drink exists.

Anyone else fail to see the point of this?

I really fancy a Double Strawberry Chocolate Coke now...

before anyone says it, this doesnt include standard fancy 404 pages, it is a technique whereby the browser is served a 404 page, then the BROWSER itself acts on that and notifies the server there is an error and gets a page acknowledging said error and giving details of what to do after back from the server

Close, but not quite. It's an add-in/plug-in to a browser that notifies another server, named the error processing server(different than the web server), of the error, and the error processing server determines alternate content to serve up.

Seems like a valid claim to me, even though I disagree with software patents. I've never seen prior art that does this process in particular.