New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.
Starred by 113 users

Comments by non-members will not trigger notification emails to users who starred this issue.

Detect cat keyboard input on Chrome and ChromeOS

Project Member Reported by, Jun 29 2012

Issue description

Detect and ignore keyboard input by cats. An extension would be ideal so that it would work in Chrome and ChromeOS alike. 

But maybe we need to focus on ChromeOS first since the detection is likely to be based on specific keyboard layout, which is likely better known on CrOS.

Showing comments 27 - 126 of 126 Older
GAH ^^^^^

Please Fix This NOW!
Regarding #18, we need to also consider cat paws for touch input. There are a variety of games targeted at cats (See which they will likely want to play using their paws. adlr will we detect the paw as 1 large touch point or 5 touch points (1 per pad)?

Comment 30 by, Jul 19 2013

Cats will adapt by unencrypting their communications and just sending messages in plaintext. 
Assigning to James who knows more about plaintext.

Comment 33 by, Jul 19 2013

I think this applies to iOS as well, for touch input.
Owner: ----
Status: Available

Comment 36 by, Jan 12 2014

My name is Butterscotch and I don't appreciate the sarcasm. Cats are real users. Please notify me when you fix this issue. Thank you. >^..^< 
Butters Mouse.jpg
60.5 KB View Download
Status: Assigned
We should think Mobile first here since cats are using phones in addition to Chromebooks. What about virtual keyboards on both platforms? Adding relevant people.
Also adding Albert who has special concerns around cat computer interaction in electric cars and touchscreens.

Comment 39 by Deleted ...@, Feb 12 2014

Please keep in mind that Maine Coons and Norwegian Forest cats have unusually furry paws. As a non-capacitive surface, this might interfere with touch screen function.
Hybrid mutual/self-cap scanning should handle that case just fine.
Vehicle interfaces ARE a concern. Car travel is often traumatic for cats. Allowing them to interact with the vehicle's controls might alleviate some of that anxiety.

My only concern is that touch screens don't usually offer enough access to key vehicle operation to fully make cats feel "in control". Obviously, we'd need to expose access to steering, accelerator, and brakes via the touch screen as well.

There are a number of vehicle manufacturers with in-car touch screens. They might be interested in integrating whatever we come up with.
CCing rsadam, who is working on optimization of the virtual keyboard layout.
See flackr's image in comment #16 for a suggested layout.
Preliminary layouts are ready. Need UX review. Stevet?
74.3 KB View Download
As an enhanced kitty-affordance, we should ensure that the M E O and W keys have a top highlight like the Food key. That way our feline users will know for sure it's a press/clawable key.

Aside from that, LGTM.

Comment 47 by, Feb 12 2014

I am concerned that the paw logos in @c45 look too much like the dogfood logos on android, thus confusing cats and dogs. 

FURthermore, how do we differentiate dogfood cat keyboard layouts vs. released ones.

True. Dog and cat paws are similar (to the untrained eye) and that could lead to confusion. Dogs may also attempt to use this interface* leading to further confusion.

I suggest we retire the term "dogfood" to avoid confusion. Referencing an extinct species with distinctive appendages in our test builds is a good way to avoid getting into this situation again. "Velociraptor-food" would be a much better term and no one will confuse velociraptor claws with cat or dog paws.

*That would not be ideal. While dogs also have 4 paws, any dog-enabled interface would be largely unusable without tongue and nose interaction points and rich smell-based output.
Wait? You're not all dogs, too?
I am not worried about dogs, I filed this issue because keyboards are natural enemies of the cats, as demonstrated here:
Status: Started
Let's get started on an implementation of rsadam@'s mocks from comment 45.

I've lined up a set of trusted testers.
IMG_20120207_221156 (1).jpg
70.9 KB View Download
Those aren't mocks. That's are the real deal. rsadam can give you a live demo (approximated by human typists as we don't have cats available on our team)

I propose we land it behind a flag, "--keyboard-cat"
Right, but I want to flesh it out further and start getting feedback from actual users.
I'll get some input from our CCI (Cat-Computer-Interation) researchers as well.
rsadam@, trusted tester "Arthur" suffers from reduced motor control due to diabetic neuropathy and has had difficulty interacting with some of the keys, particularly the "mouse" glyph. What do you think about offering larger hit targets?
abodenha@, this should be fixed on ToT as part of . We should get Arthur to try the keyboard with the latest version!
Cool. Let me see if I can get him on the latest canary and get feedback. Thanks for the quick action.

Another bit of quick feedback: Several of the testers expect the "mouse" glyph to actually extrude mice somehow. It could be cool, but I definitely see that as V2.
Recommend that Arthur not be given access to canary. Can this wait for beta?
Good point. Canary might be a bit jittery.

We've got a dev build now, I'll get him set up with that.

Comment 62 by Deleted ...@, Mar 10 2014

Thank you @rsadam! The new "dev build" works wonderfully, even with my limited motion.

I can finally see why my pet humans spend so much time with this glowing box. Social media is a revelation! No more relying on scent alone to keep track of the comings and goings of my associates! Thank you! I'm looking forward to sharing this with other cats once it hits stable.

Now to acquire my human's "credit card" so I can learn about online shopping. Does Google Shopping Express stock birds? All this talk of "canary" has made me hungry.
cc'ing myself for Android feature parity.
Blocking: chromium:356242
Status: Fixed
After some extended trusted user testing and extensive Animal-Computer Interaction (ACI) research, this bug is now fixed. Install the kitty keyboard from the Chrome Web Store:

Features include four pawing modalities and mouse mode using your trackpad or the touchscreen, as well as a translation console (beta).
For the future: input research is underway for other pet users, including dogs, fish, hamsters, and dinosaurs. 

Comment 66 by, Apr 1 2014

Opera did similar research, and released the results today in the form of a touch-enabled, cat-optimized mobile browser.

Comment 67 by, Apr 1 2014

I'm really looking forward to the dinosaurs compatibility!
Great to hear that someone is working on that.

Comment 68 Deleted

Don't dinosaurs use Mozilla, on Windows XP? Are we hoping they will switch to ChromeOS?
Status: Available
Unfortunately while that extension is great, it does not address the issue which was filed here. Basically the goal is to *ignore* the cat's input when said cat is walking on my keyboard.

Comment 71 by, Apr 1 2014

My cats are loving the new extension. I am so happy that I can understand them now! It surprises me the things cats think about! Sparkles only wants fish for dinner! How did I not know that? Fluffy says that Sparkles bullies him sometimes. I should have known!

I love communicating with my cats! Now they can be my even-bestester friends! Thanks Google!

Comment 72 by, Apr 1 2014

This is Sparkles. Wow. So red dot. Very chase. FEED ME! Many excitement. Much erratic! NYAN! Tomorrow, and tomorrow, and tomorrow, Creeps in this petty pace from day to day, To the last syllable FEED ME!
this is a much needed feature and not just in chrome os but also in chrome for all other platforms.  All cat lovers need this urgently.

online audience.jpg
175 KB View Download
Owner: ----
re: #70, if the aim is to ignore cat input, it sounds like an event filtering issue rather than a virtual keyboard issue.
re: #74, totally. In comment #15 this bug got swayed into the "cat keyboard" efforts.

One idea: using the device camera might help. First collect image data of both humans typing and cat typing, then develop a classificator. Discriminating between the two in dubious cases (some people type worse than cats, so looking at just keypresses is not enough) would then be just a matter of taking a picture.

This feature could work proactively, not waiting for that cat to appear, but switching on keyboard filtering the moment the human leaves. What to do if both a human and a cat are present remains to be determined.

As much as I hate to admit it as a software person, the ultimate solution may be a hardware based one. Many cats do have a short range radio tags implanted so it is a matter of being to detect a proximity of such tag to the device.

Comment 77 Deleted

(No comment was entered for this change.)
Status: Assigned
Assigning to Santhosh who has shown interest in pawing UX
Ping? Any progress on this?

Comment 83 by, Apr 1 2015

aurimas@ we have a working prototype, but for some reason dogfooders have been unwilling to install the necessary cerebral implants.
This should be a good starter bug.

Comment 87 by, Jun 1 2015

I ran into the same issue. Help! Any progress?
276 KB View Download
288 KB View Download
312 KB View Download
We're considering an approach using the touch sensors to detect fur (with the notable unsupported edge-case of hairless cats)

This really needs PM support to flesh out the details and recruit test subjects.
It is usually done by detecting nonsensical keyboard input: clumps of keys pressed at once, corresponding to the cat putting its paws on them and similar stuff.

I propose detecting not the presence of a cat but instead the absence of the rightful owner. This would turn this into a security feature which should bring more attention to it from more people.

Checking if the owners face detects on the lid camera would be one possibility... (Detecting cat on camera does not work, cat is sneaky and moves outside of the FOV).

(I already proposed adding a RF chip reader to the laptop and detecting the chip many cats have implanted in comment #75)

Comment 90 by, Jun 25 2015

Can't this all be done in software by implementing a Catcha?
#90 "When stepping away from your Chromebook, make sure you lock your screen so no one can mess with your account."
Cats produce a sound called "meow", a meow speech recognizer will be brilliant to add to the project, maybe with pitch interpretation

Comment 93 by Deleted ...@, Jun 27 2015

I have a laptop with Chrome and several cats. Would be willing to test anything if needed. Does it matter if the cats used to be feral, or must they have been domesticated their whole life for this to work?

Comment 94 by Deleted ...@, Jul 1 2015

Oh, I thought cat changes for the mobile platform where already implemented in Android KitKat. Can anyone confirm this?

Maybe detecting the static electricity route is the best?

Furry + static charged + random keypresses => mouseclick to confirm you are human.
To verify the input is from a cat the screen of the computer should display a red big dot moving from left to right edges of the screen. The cat would follow the dot and make keypresses, using JavaScript it listens for key presses during the dot movement and analyzes it so the keys pressed have some sort of pattern (maybe the cat pressed a, g, g, k, l that is the order of some keys in left to right on my laptop)
This was originally closed in #65 with a similar approach, though the fix at the time was to translate the typing into English rather than filter it.
As part of PE initiative assigning this to Kan to bring up with relevant VP and SVPs and report back to us with executive guidance and vision.
Labels: Hotlist-Recharge
This issue likely requires triage.  The current issue owner maybe inactive (i.e. hasn't fixed an issue in the last 30 days).  Thanks for helping out!

Labels: -Hotlist-Recharge
This is totally relevant still.

Comment 100 by, Dec 15 2015

​**​Please note that Review-UI label is deprecating. For all future U​X​​ or user-facing changes​ that need ​a ​review ​to launch, please have a Googler representative send an email to Chrome UI Review with information about the ​change or new feature, any design mockups, screenshots, or instructions to test ​it out.​**​
From a user feedback report (note that any PII has been redacted from this report )

Headline: Too easy for unevolved creatures to wipe the device.

Okay, this has happened twice now.  

In the summer of this year, I spent a bunch of time putting [redacted details of alternate os install] on to my [redacted device name].  At that point in time that was a feat -- I had to [redacted details of efforts required].  Then I come home from brunch one day and as per normal my cat [name redacted] is sitting on the keyboard, and it says like "please insert a valid recovery device", and all that hard work had been wiped.

Second time -- [details of another os install redacted] The baby was banging and banging on the keyboard while it booted back up, as he likes to do.  And ten seconds later: "Congrats, you're no longer in developer mode and the machine is wiped".

Suggestion: Please, please, before wiping the partition, can you require the user to type something like "yes, reset my chromebook" into a text box in order to continue?  I don't think my cats (either one of them) or my baby could manage that.  Or if they did manage to, then that would be so impressive that I wouldn't mind the factory reset.

Original report:
Similar issue to #102 on my Pixel 2015 today: Cat decided that a warm keyboard is a nice place to stay and in the process pressed the power button for a long time.

Result: the system booted up and went down repeatedly, to the point were it claimed that Chrome OS is damaged. Luckily that could be solved by another reset, but it was still unsettling to see the recovery screen just because of temporary feline occupation.
mgreenwald@ your work with follicle protein classification is promising. Any ideas here?

Comment 105 Deleted

Thanks Albert I'm happy to share my progress. At the beginning of this project a meow-tain of data sat before our team. Research has been slow but now litter-ed with potential results.  There is one promising new cat-egory. My friend ksangani@'s research focuses on stopping the main key that is pressed, the Tab-,by cats. ksangani@ tell us more!

Previous solutions were too expensive, and the Fee-line up showed that these costly ones were not practical. However genius efforts by jare@ have solved the cost problem - take it away jare@!
Sadly, tab key research hasn't yielded many results. Tried detecting hairballs landing on the keyboard, but that hasn't coughed up much either. Looking into clawprint detection, but we might just have to start from scratch.
An edge case example of a cat hacking a human device, namely the toilet.  This suggests this same cat could hack keyboard input posing as a human imposter:
Project Member

Comment 110 by, Jan 5 2017

Labels: Hotlist-Recharge-BouncingOwner
Status: Untriaged (was: Assigned)
The assigned owner "" is not able to receive e-mails, please re-triage.

For more details visit - Your friendly Sheriffbot
Status: Assigned (was: Untriaged)
yhanada@ something to consider as we work on the VK.
Chris just mentioned to me this is important for his new keyboard project. He will make this a P0 for the new device.

Comment 113 by, Feb 28 2017

+smbarber who is now my TL on this.
I think this is a prime opportunity to flex our machine learning muscles, but as with all machine learning ventures we'll need a lot of data.

I'll allocate some lab space and acquire the cats. igo@ if you can find a representative set of Chromebooks (different keyboard layouts, sizes, etc) then we can start gathering some useful data.

Comment 115 by, Jul 24 2017

Owner: ----
Components: -UI
Cats are similar to teddy bears, right?

Comment 117 by, Jul 24 2017

Shouldn't we be making Chromebooks for "everyone"?  Cats can't type, so +zalcorn for cat voice interaction and assistant PRD.
Language translation shipped in an extension on April 1, 2014. This bug is specifically about supporting paw -> keyboard interaction.
@omrilio as the early promising progress for virtual keyboard in #45 apparently weren't followed up on. Reports were that the trusted testers objected to being called "dogfooders."
@alberto, since cats only go on computers as a heat source, should we put an air conditioner in each Chromebook?
I don't think the statement in #120 is correct. Cats have a wide range of motivations, only some of which (warmth, a desire to annoy their pet humans, etc) are fully understood.  Thus the need for a fully fleshed out UX and translation system.

Comment 122 Deleted

Houston, we have a problem:
Chrome Unboxed would like to submit Chrome the Cat as a test subject for any upcoming case studies.
1.9 MB View Download
Since our devices mostly use trackpads - not mice - cat usage is likely low. 
Showing comments 27 - 126 of 126 Older

Sign in to add a comment