Can you rotate your screen without infringing a patent? A case study on reducing the cost of patent searching

最終更新日

In this post we review a patent claiming the rotating screens used on the smartphones or tablets we almost all carry – and consider the most cost effective means of finding the key patent that invalidated this patent when it was asserted against Rackspace.

 

Can you rotate your screen without infringing a patent?

Can you rotate the screen on your phone without infringing a patent? Thanks to the efforts of Rackspace, you (probably) can. Rackspace had been sued by a company called Rotating Technologies over US patent 6326978, with a filing and priority data of April 1999. This patent claims a Display method for selectively rotating windows on a computer display.

Rotating Technologies had asked for $75,000 to go away. Rackspace declined the offer and instead went after the Rotating Technologie and tried to invalidate their patent. Even after Rotating Technologies offered to walk away, Rackspace kept going, eventually invalidating the patent in an Inter Partes Review held before the USPTO.

The just published decision itself can be downloaded from the USPTO website here, after typing US6326978 into the search bar. The opinion found the Robbins patent invalid in the light of three main documents, being in order of importance

  • US6137468, filed by IBM in 1996 for a Method and apparatus for altering a display in response to changes in attitude relative to a plane
  • US5345543, filed by Apple in 1992 for a Method for manipulating objects on a computer display
  • The Adobe Photoshop 5.0 User Guide

And the asserted patent? Its first claim read:

1. A computer display window comprising:
     a display portion;
     a frame surrounding the display portion; and
     means for selectively rotating the window about a rotation point at the discretion of the user;
     wherein the plane of the window, the plane of rotation, and the rotation point are coplanar.

Or in other words, a screen that rotates such as found on almost any mobile device. You may say that as a user, as opposed to manufacturer of devices, you should be free to ignore this patent. However in the past other NPEs have not been slow to chase end users of alleged infringed patents, such as the users of a scanner to send a document who were (and maybe still are) being sued in relation to patent formerly owned by Project Paperless. 

 

How to find a knock-out in 10 minutes. And, what is the cost of patent searching?

Many of you reading this blog will be very familiar or practiced with patent searching. A number of free and paid patent search services are available, while there are a range of companies offering patent searches from $400 upwards.

However the cost you pay to access the patent search services is only part of the cost of doing the search. The real cost in patent searching may be your time, including the time you or your clients may spend reviewing the results. I have been told that a patent examiner can take 2 to 3 days to look for prior art for a patent. The cost of this would only be known to the managers of these patent offices, but we can presume that it would be significant. The same would apply to any highly paid professional reviewing patents, and this would describe many people in the patent profession.

So clearly the cost of patent searching is a lot more than the cost you may pay or not pay for access to a patent database or patent search results.  

To demonstrate this, it may help to consider how you might search for prior art for the Robbins patent. You could, for example, type in ‘rotating screen’ into the search box for Google patent – which will show you 39,600,000 results, not all of them 100% relevant, such as patents for rotating screens used for screening rocks in mining:

GooglePatentSearch.jpg

 

Of course this is a very simplistic search strategy, even if it was the first that reasonably came to mind. Experienced searchers would then fine tune and optimise this search strategy, often using the powerful filters found in some patent search engines. But this would take time, and this could lead to a few hundred similar prior art patents to search through. These could take hours of valuable and expensive time to review.

 

How to search when you don’t have several hours or days, or the money to pay for several days of professional patent searching?

The need to speed up patent searching is what led Ambercite to develop AmberScope. For example, if we were asked to find prior art for the Robbin’s patent, we might start by entering the Robbins patent number (US6326978) into the AmberScope search bar. The Robbins patent then becomes the ‘seed patent’ for this particular network. The result might look like this (click on the network image to see a fully interactive version of the same network). 

US6326978-as-seed-patent.jpg

 

Each dot represents either a forward or backward citation patent, which should be reasonably similar to the Robbins patent. Since we are mainly interested in prior art, we have the option of hiding all patents filed in the year 2000 and later using the year filter (the Robbins patent was filed in 1999). 

 

US6326978-year-filter.jpg

Not many patents remain when we remove the forward citations. Some of these are solid dots (representing direct citations) and some are translucent dots, representing what we call ‘ghost patents’, which are patents that still my be relevant but are not directly connected to the seed patent.

To quickly review any of the patents, we can simply mouse over them. If we do this, one of the patents we find US5949408, for a Dual orientation display handheld computer device, filed by Hewlett Packard in 1997 and with a priority date of Sep 1995.

 

HP-patent-found.jpg

 

This discloses, among other things:

A computer device, for instance a palmtop personal organizer, has a dual-orientation display device with both portrait and landscape modes of viewing the display.

Besides a simple review of each patent shown, there are a number of options for further filtering patents in AmberScope – perhaps my favourite is the Similarity filter, which can be used to progressively hide less similar patents as its value is reduced from 100% down to close to zero % (the value on the filter shown the proportion of the patents remaining as the filter is used – and the less similar patents are removed in this process). 

For example, if we were to apply the similarity filter to exclude all but the most similar prior art patent, we would end up with US5329289, filed in 1992 by Sharp for a Data processor with rotatable display. In the image below, note the dark blue around the Similarity filter. This filter has been reduced to a value of ‘3’, which means that all but the most similar, top 3%, of patents connected to the Robbins patent are shown. 

 

US6326978-year-and-similarity--filter.jpg

 

This is a potentially relevant patent as it refers to: 

A data processor with rotatable display includes a display unit having a rectangular display surface rotatable to either a vertically elongated position or a laterally elongated position.

Admittedly this refers to a conventional computer screen, but the concept is the same as for a mobile device. However perhaps the real value with the Sharp patent is that the summary box refers to 85 further patents, not directly connected to the Robbins patent but which still may be similar to it. ,

85more.jpg

 

By selecting this ’85 more’ button, the patent network will be refocused onto this patent;

Sharp-patent-as-focus-patent.jpg

 

Again we can apply an 1999 and earlier filing year filter.  Since we might aim to find something connected to this patent that was not cited by the examiner for the Robbins patent, we will use the similarity filter to show the top 10 most similar patents, the top 19% most similar patents, as shown below.

 

IBM-patent-found.jpg

 

And voila! one of those 10 most similar patents turns out to be the IBM patent. In the words of the invalidity judgement on the Robbins patent, this IBM patent:

discloses rotating a laptop around one axis that results in a change in orientation in another axis, and maintaining windows level with respect to a preselected reference plane during such a rotation.

This appears to be the key patent in the invalidation of the Robbins patent, even if applied in combination with the Apple patent (which claims more of a method of manipulating objects in a display, rather than the whole display).

Of note, the IBM patent was not directly linked to the Robbins patent – which indicate that the examiner for the Robbins patent did not appreciate its importance. This in turn makes it more suitable than perhaps the directly connected patents such as the Sharp patent, as the owner of the Robbins patent cannot claim file wrapper estoppel for a patent that has not been cited yet (file wrapper estoppel is the principle where an already cited patent cannot be raised during an reexamination process as the examiner should have already considered this during examination. However this does not apply to federal court invalidation).

So in just a few steps in AmberScope we have found a key prior art document that has not been cited by the examiner. And without having to spend many valuable hours having to review hundreds of patents…

 

For those who truly want to speed up patent searching:

Ambercite has also developed Automated Patent Searching. These fully objective searches list and most importantly rank the most likely prior art or later patents that are:

a) potentially similar to your patent of interest

b) either directly linked to your patent of interest, and perhaps of most interest, not directly linked to your patent of interest by citation linkages. In other words, these are patents whose relevance to your patent of interest has either been missed or ignored by patent examiners. However as the above Rackspace example has shown, this does not mean that these indirectly linked patents are not relevant. So the information found can be very valuable – and also missed by other patent searching techniques.

So what did our Automatic patent search (“Prior Art Finder”) find in this case? The full report is available upon request, but in short, the three most similar directly connected prior art patents are shown below. We of course could show more than the top three, but we wanted to keep this short

 

Predicted Similarity Rank

Patent number (Filing year)

Listed patent owner

Patent title

Amberscore  (links to AmberScope)

1

US5329289  (1992)

SHARP

Data processor with rotatable display

7.4

2

US6115025 (1997)

SILICON GRAPHICS

System for maintaining orientation of a user interface as a display changes orientation

6.6

3

US5949408 (1997)

HEWLETT PACKARD

Dual orientation display handheld computer devices

9.5

 

What is more interesting to many people, and particularly patent lawyers wanting to invalidate patents in reexamination processes, is the ten most similar indirect (not previously cited) prior art. In the table below, I have highlighted all patents that refer to a rotating screen in green, while the pink patent is the IBM patent that played such a key role in invalidating the Robbins patent.

Predicted similarity Rank

Patent number (filing year)

Listed patent owner

Patent title

Amberscore  (links to AmberScope)

1

US5973664 (1998)

PORTRAIT DISPLAYS INC

Parameterized image orientation for computer displays

11.4

2

US4542377 (1982)

IBM

Rotatable display work station

3.7

3

US5774233 (1994)

MATSUSHITA (PANASONIC)

Document image processing system

2.8

4

US5708561 (1996)

HUILGOL VIVEK R

Portable computer having display slidably and rotatably mounted for movement between landscape and portrait orientations and to open and close speaker ports

6.2

5

US4831368 (1987)

HITACHI

Display apparatus with rotatable display screen

4.0

6

US5577187 (1994)

MICROSOFT

Method and system for tiling windows based on previous position and size

5.9

7

US5134390 (1989)

HITACHI

Method and apparatus for rotatable display

4.7

8

US5596699 (1994)

DRISKELL; STANLEY W.

Linear-viewing/radial-selection graphic for menu display

6.2

9

US6137468 (1996)

IBM

Method and apparatus for altering a display in response to changes in attitude relative to a plane

6.4

10

US6189850 (1998)

MITAC INT CORP

Rotatable lcd screen device

5.9

 

So in just one Automated Prior Art search, we have identified the knockout prior art patent for a previously granted US patent.

 

How much do Ambercite tools cost to use?

Access to the AmberScope is now charged, but with pricing plans ranging from $40 for a ‘Short’ plan, which includes 10 searches (10 seed patents) . Considering that the above sample search was able to locate the IBM patent with just 2 seed patents (being the Robbins and Sharp patents), we think that the 10 seed patents should be more than sufficient for most standard search projects, particularly considering the productivity of AmberScope searching compared to other patent searching processes.

Alternatively we can offer an ongoing subscription at $200 (plus local taxes) for one month, or $2000 for one year. Again there are limits on the number of seed patents, but we think that 100 seed patents in a month is plenty enough for most searchers, patent attorneys and IP managers. Higher volume plans are also available for those whom really love their searching.

An Automated Patent Search costs from $500 and can be delivered in 1 to 2 working days.

 

Will Ambercite tools save you money?

Again, when considering the total cost of patent searching, we think that our prices are very good value. Most importantly, they can increase searching productivity and so help to lower the total cost of patent searching for our users.

In this case we were able to show how quickly Ambercite or Automated Patent Searching could find the knockout patent – and a patent not cited by the USPTO when they examined the Robbins patent despite the 2 to 3 days they might have spent searching for prior art for the Robbins patent.

If nothing else, this helps disprove an argument that ‘anybody could have found the IBM patent’, because in this case both the patent examiner and the patent applicant failed to find or appreciate the significance of this patent in their searches, presumably using more conventional approaches.

Read more