9

This is a recently issued patent from Twitter

Device independent message distribution platform "Device independent message distribution platform"

This recently issued patent describes the basic mechanics and functionality of Twitter.

BASIC PATENT DATA:

  • Patent Number: 8,401,009
  • Priority Number: US2008177589
  • Issue Date: March 19, 2013
  • Priority Date: July 22, 2008 (provisional filed July 23, 2007)
  • Assignee: Twitter, Inc.
  • Inventor: Jack Dorsey, San Francisco and Christopher Isaac Stone

PATENT OVERVIEW: This is the patent that describes the basic mechanics and functionality of Twitter, summarized roughly as:

  • A method of communicating among users
  • One user (the "first" user) can "follow" another user (the "second" user).
  • The "second" user can post a message. The message is not posted "to" anyone, it is just posted.
  • All followers of the second user receive the posted message.
  • The first user can choose from a set of possible methods for receiving messages.

The first claim contains more steps such as, if the first user tells the system whom they wish to follow, the system records that fact. And when the second user posts a message, the system looks up the list of all users previously recorded as following that user so it knows who to send copies of the message to. And so on.

Claim 1 requires, specifically:

  1. A method for device-independent point to multipoint communication, the method comprising:

    Receiving from a first computing device of a first user a selection of one or more endpoints for receiving update messages;

    Receiving, from the first computing device, a request to follow a second user;

    Designating, by a computer processor, the first user as a follower of the second user in response to the request, wherein designating the first user comprises configuring an account of the first user to reference update messages broadcasted by the second user;

    Receiving, from a computing device of the second user, a broadcast request to broadcast an update message in a first format, wherein the update message lacks identification of the first user as an intended recipient, and wherein the update message includes an identification of the second user as a sender of the update message;

    Identifying, by the computer processor, a plurality of followers of the second user in response to the broadcast request, wherein the first user is among the plurality of followers;

    Determining addressing information of each of the plurality of followers, wherein the addressing information of the first user identifies the endpoints for receiving messages;

    Applying, for each of the plurality of followers, rules to the update message based on the addressing information;

    Translating the update message into an appropriate format for each of the endpoints; and

    Broadcasting the update message to each of the endpoints in the appropriate format.

Other claims are for the computer system that does this.

WHY IT MATTERS: This patent covers the basic functionality of a Twitter-like service.

Are you familiar with technology similar to that described in this patent application?

If so, please submit evidence of that prior art as an answer below. Please submit only one piece of prior art per answer below. We welcome multiple prior art proposals from the same individual; please create separate answers for each one. This is so the community can vet each individual piece of prior art independently.

For details about what makes good prior art, please see our FAQ. If you'd like to contribute in another way, please vote or comment on submissions made below. And we welcome you to post your own request for prior art if you know of another questionable patent or patent application. Thanks for participating!

3
  • Hi Philip, welcome! Would you mind editing your question to follow our format for prior art requests? I know you intend to answer this yourself, but following a consistent format would also help others if they had any knowledge to share. Thanks!
    – Adam Lear
    Mar 21, 2013 at 17:31
  • Your paper seems very relevant to this patent's claims. I wanted to let you know that it only needs to be early enough to qualify as prior art with no extra points for how early it was.
    – George White
    Mar 21, 2013 at 19:07
  • Philip, Can you please add the suggested prior art you found: "web.archive.org/web/20010223204516/http://miski.sourceforge.net/…. This describes a somewhat similar messaging system with follower/following mechanics. Note that Feb 2001 is not only earlier than the application date for 8401009, it is also earlier than all but 2 of the patents in the "Referenced" list (and those two don't have anything to do with "follower/following", as far as I can tell)" as an answer to your own question. Thank you. Mar 22, 2013 at 23:49

6 Answers 6

3

Miski: A White Paper

This document, which originally existed as a documentation web page for an unimplemented project on SourceForge, now only exists as a series of snapshots on Wayback.

The document describes a messaging system with follower/following mechanics, very similar to that described in the independent claims of the patent. Note that Feb 2001 is not only earlier than the application date for 8401009, it is also earlier than all but two of the patents in the "Referenced" list, and those two don't have anything to do with "follower/following". The reason I assert priority over the other references is to point out that one can determine the relevance of Miski as prior art without reading all the other references dated later.

To match Miski with the Twitter-like system defined in the patent's first claim, one must translate or relate concepts as follows:

  • The patent's "first user" is the subscribing user in the Miski document (which I also referred to as the "second user" at one point).
  • The patent's "second user" is the posting user in the Miski document (which I also referred to as the "first user" in some places).
  • There are no "subjects" in the patent. Miski would be equivalent in functionality to Twitter if everyone had a default subject called "Everything" (this scenario is not considered in the Miski document, but the users of an actual implementation of Miski would always be able to choose to define such a subject within their own subject hierarchy).
  • In the Miski document, I suggested that a message always consists of a URL which points to the content. As with the patent, actual delivery of the message is by whatever means is agreed between the subscriber and the system.
  • Within the 9 steps detailed in the patent's first claim, the Miski document does not explicitly list steps 5 to 9 as separate steps, but it does clearly state that a message is delivered from poster to poster's server to subscriber's server to subscriber, as a consequence of the subscriber subscribing to the poster and the poster then posting a message. At the time of writing, I would have regarded the detailed computation steps required as fairly "obvious" (but I could discuss this in further detail if desired).
  • In the Miski document I described a system of unique IDs, where each ID consists of a unique name within a given domain plus a domain name, with a similar structure to that of email addresses. The patent claims do not seem to mention assignment of IDs at all, although it is hard to imagine how a Twitter-like system could be implemented without explicit IDs.
  • "Reposting" is like "retweeting" (but there is no "retweeting" in the patent anyway, so that concept is not relevant to the prior art issue).

To assist with reading the Miski document, so that anyone can skip those parts not relevant to the issue of prior art, I have identified three selections within the document that contain all the prior art for this patent:

  • From "A brief summary of the features that define the Miski system is as follows -" to "The headers include a web URL pointing to message bodies for those messages for which the receiving users decide they want to receive the full contents."
  • From "The following scenario demonstrates the basic features of the system- " to "Her web browser retrieves the message contents from Fred's server. "
  • The section under the heading "Delivery Methods" (which ends "as a moderated chat system.")
2

This patent is an obvious application of the "publish-subscribe" pattern. An earlier implementation of this pattern was the usenet system in 1979 which allowed messages to be posted (published) to newsgroups without any predetermined recipient. Other users were able to follow (subscribe) to newsgroups so that the newsnet client software could track which messages they had previously read. A more detailed explanation of publish-subscribe and usenet can be found simply by Googling those terms.

3
  • AKA Observer pattern, first implemented in SmallTalk en.wikipedia.org/wiki/Observer_pattern Jul 23, 2013 at 10:12
  • Also: "Exploiting virtual synchrony in distributed systems" 1987 (i don't actually have access to this paper, but it should be relevant) Jul 23, 2013 at 10:16
  • I was able to view the "Exploiting virtual synchrony" paper ... section 3 discusses "tools" that can use the proposed method (as proposed in the paper, not the patent), and section 3.9 (page 130) singles out a "News Service". Quoting: "Each subscriber receives a copy of any messages having a "subject" for which it has enrolled in the order they were posted."
    – user4889
    Jul 23, 2013 at 17:08
1

Interesting reference, but I think one main difference is that Miski is based on users subscribing to arbitrary subjects, whereas the Twitter patent is (partly) based on users subscribing to other users. Now this might seem like an obvious variation ("let subject = user id"), but:

  1. it changes the functionality in subtle but significant ways (for instance, only one user can make posts on a "subject", and a user can only make posts on one "subject"); and
  2. the claims have many other elements in addition to followed/follower mechanics, and this is related to only one of those (and even that one it doesn't directly address).

When it comes to subject-based publish/subscribe systems, there are many, many other systems that may predate even Miski, such as mailing lists, group chatrooms, and message-oriented middleware such as MQSeries/MQ, Tibco, and so on. A quick glance at the cited references found quite a few about forwarding messages to relevant subscribers (although I didn't find any where the subscription is for a user).

Additionally, it's important to realize that the claim has many requirements other than following posts. The first claim is as follows:

A method for device-independent point to multipoint communication, the method comprising:

receiving from a first computing device of a first user a selection of one or more endpoints for receiving update messages;

receiving, from the first computing device, a request to follow a second user;

designating, by a computer processor, the first user as a follower of the second user in response to the request, wherein designating the first user comprises configuring an account of the first user to reference update messages broadcasted by the second user;

receiving, from a computing device of the second user, a broadcast request to broadcast an update message in a first format, wherein the update message lacks identification of the first user as an intended recipient, and wherein the update message includes an identification of the second user as a sender of the update message;

identifying, by the computer processor, a plurality of followers of the second user in response to the broadcast request, wherein the first user is among the plurality of followers;

determining addressing information of each of the plurality of followers, wherein the addressing information of the first user identifies the endpoints for receiving messages;

applying, for each of the plurality of followers, rules to the update message based on the addressing information;

translating the update message into an appropriate format for each of the endpoints; and

broadcasting the update message to each of the endpoints in the appropriate format.

As such, the claim is not only about following users and broadcasting posts to followers, but also about "applying rules" based on "addressing information" to those posts and "translating" them to "appropriate formats" based on the devices associated with the followers. (Plenty of prior art around for those aspects as well, including in the cited references, but the combination of all these aspects is what made the claim novel and non-obvious. Conversely, one must perform all the elements in the claim to infringe.)

3
  • 1
    Miski is users first and subjects second. If all Miski users created a default subject called "Everything", and only subscribed to the "Everything" subject of other users, then the mechanics would be identical to that of Twitter. The only other two significant differences are - 1. that Miski is explicitly multi-server, whereas Twitter is implicitly single-server (implicit in the sense of the implicit subject of all the claim steps). And 2. Miski always delivers messages as URLs pointing to actual content. Which with 8401009 is given as a possible method of delivery. Mar 22, 2013 at 3:41
  • Ah, I see, users subscribe to a combination of user + subject. That certainly is more Twitter-like, though it is still not exactly the same (i.e. requiring the creation of a "*" subject). Again, seems pedantic, but really, even tiny differences matter in interpreting claims. If the writeup had explicitly mentioned this possibility, it would make a stronger case for obviousness. However, from a quick trip to USPTO PAIR, it seems the followed/follower mechanic elements were what finally differentiated the claim from all the prior art, so Miski might yet prove to be very relevant.
    – kinkfisher
    Mar 22, 2013 at 20:09
  • As an aside, the multi-server vs server difference is not important, because multiple machines are still involved: the implicitly mentioned server / "computer processor", the "first user's" and the "second user's" devices. It doesn't really matter if they are servers or smartphones because they're all simply "computing devices", so in that sense, Miski does not differ significantly. The client vs. server distinction might only matter in Miski when it comes to re-posting, which these claims do not address anyway. The difference in delivery methods is also not relevant.
    – kinkfisher
    Mar 22, 2013 at 20:26
1

How about US Patent No. 6,389,455 (1998 filing)?

Col. 8:4 references corporate accounts where the regional manager can be on the receiving end of sale rep communications. All contacts go through a dotbounce hub so the message is not sent to anyone. Rules of the forward impact formatting.


EDIT

For those that did not read the '009 patent -

"Disclosed is a system (and/or method) that includes, for example, a routing engine that receives a message from any of various entry points, including e-mail, .... The routing engine determines the identities of the destination users to receive the message, possibly by expanding destination groups. The routing engine determines the endpoints on which the destination users wish to receive the message, the endpoints can be one or more of e-mail,.... The destination endpoints are independent of the source entry points, and the message sender does not need to have knowledge of the endpoints, or endpoint-specific user addresses."

The '009 is premised on the belief that as of its filing -

"Additionally, there is no general method of sending a message to multiple users who may be receiving the same message using different devices or interfaces. For example, a user may want to send a message to three SMS recipients (all using different cell phone service providers), two e-mail recipients, and a web interface recipient."

Hate to tell you but that is what the '455 patent is. The big question is who is Meyer Cordless LLC?

__ Edit - answered my own question - Looks like it is affiliated somehow with IV.

http://webcache.googleusercontent.com/search?q=cache:kJPfdZL-WgQJ:www.plainsite.org/flashlight/index.html%3Fid%3D3245967+&cd=4&hl=en&ct=clnk&gl=us&client=firefox-a

__ EDIT - We've overcome the "'009 isn't directed to email" and are now on the "receiver choosing" is different, well to the extent you think "receiver choosing" is in the '009 claim -

'455 discloses the receiver controlling the flow of the email - Non-sender notifies and controls routing of the incoming message.

'455 - "For example, a user who is a frequent flier on a particular airline may wish to receive airborne e-mails care of the airborne e-mail service provider. Such airborne e-mail service provider notifies the bounce system of the presence of a particular User and inform the bounce system of the applicable routing address and duration of the validity of such routing address."

Is there is a particular element in the '009 claim that is lacking, point it out. Please reference the claim.

4
  • I don't have enough "reputation" to downvote this, but that is a patent about forwarding your own emails to another email account which is yours (or perhaps, your manager's). That is quite distinct from broadcasting messages publicly to everyone else in the world who wants to read all the messages that you post. Mar 25, 2013 at 20:48
  • Phil - entry point 102 of the '009 patent can be an email. The exit point can be anything, including an email (see 114 above). You see the '009 patent as excluding "email" as incoming message formats? I'd also note that the concept of "broadcasting to everyone is not what the claim is directed to. If there is a limitation in there, please ID it.
    – user3359
    Mar 25, 2013 at 21:02
  • I actually said "broadcasting messages to everyone ... who want to read all the messages that you post" And steps 4 and 5 of claim 1 of patent US8401009 pretty much do address that, and step 4 has the word "broadcast" in it. Whereas email is definitely not a technology for broadcasting messages to non-specified recipients. Mar 29, 2013 at 2:49
  • The word "broadcast" is NOT in the '009 spec or in the claims as filed. It was inserted in the final amendment before allowance. If you want to ascribe a special meaning to the term "broadcast" that is not supported by the application, the '009 patent would have 112 issues. Non-device dependent entry message is a broadcast which is what the '455 has.
    – user3359
    Mar 29, 2013 at 11:53
1

Some of this patent's claims also hinge on the fact that a Tweet is processed and pushed out to multiple transports (email, SMS, web, IM...) and users may consume these messages with multiple clients.

At least one open-source system had these capabilities for almost a decade prior: XMPP, the Extensible Messaging and Presence Protocol, also known as "Jabber". XMPP was created in roughly 1999, and IETF standards were produced in 2004 (http://xmpp.org/xmpp-protocols/rfcs/). There have been many implementations of XMPP servers over the years, especially when interoperability was a concern. Until recently the chat system within Google's Gmail, Hangouts, and Talk system interoperated with other chat systems via XMPP.

Admittedly, XMPP's communication model is very different from Twitter. XMPP is based on person-to-person instant messaging, although a variant of the protocol could be used for chat rooms. XMPP has no notion of blogging and passive reception, aka the "message without a recipient" model that Twitter touts as its unique feature. XMPP is an XML-based protocol, designed to 'tunnel' over anything that can handle a long burst of text. Twitter is not a protocol at all, but a single service.

I only mention this system to make it plain that these claims are not unique to Twitter:

  • defining a person as an 'end point', with a unique address. In Twitter, a unique account is conventionally represented with the "at" sign, like @jack for Jack Dorsey. In XMPP, the form is similar to an email address, [email protected].

  • sending a message out to multiple 'end points', aka people

  • messages are routed to these end-pointy people depending on what transport they are using (email, IM, SMS, web...)

1
0

None of these are really patentable. While Twitter created a brilliant product with brilliant execution, it was a blending of two already-existing ideas:

  1. Blogging with followers lists
  2. SMS with distribution lists

One of the creators of Twitter, Evan "Rabble" Henshaw-Plath, confirmed as much in a recent talk given at the OHM 2013 conference, on August 2, 2013, where he named many precedents for Twitter as direct influences.

The innovation here was not technical innovation so much, but in product design. In previous broadcast SMS systems, one had to manually define a list. With Twitter, it was defined by the other users who followed you. And the blogging sites which had a concept of 'followers' typically focused on long-form diary entries, unsuitable for Twitter's vision of a stream of instant, short messages, shareable to mobile devices. But great product design, however praiseworthy, doesn't deserve a patent.

As for blogging with followers lists, Henshaw-Plath named Facebook. But I would go back a bit further, to the blogging platform LiveJournal, which had almost all of these features, almost from its inception, in 1999. It began as another blogging platform but rapidly became about the network of users who "followed" each others' blogs.

LiveJournal is not well-known today, but in the early 2000s it was one of the leading social networks, especially for teens and young people. It even briefly appears in the movie The Social Network, since the real-life Mark Zuckerberg was a user before he created Facebook.

As for SMS with distribution lists, there were many similar services that predated Twitter. Some of them were explicitly for social sharing, like Dodgeball (founded 2000). Henshaw-Plath cited TXTmob, which was used for protest actions during the 2004 US Elections, as one of the major inspirations for Twitter. Henshaw-Plath also related the same story about TXTmob in comments to a chapter of a book about Twitter. Those comments seem to be removed from the site today, but they are still available on the Internet Archive's cached copy from 2009. You can also find a description of a system created by Microsoft Research in 2005 (http://research.microsoft.com/~counts/pubs/slam_jcscw_leisure2.doc)‎ that allowed for the creation of such distribution lists on mobile devices. Other sites like 3Jam and Yahoo's experimental service MixD offered similar services in the mid-2000s, contemporaneous with early Twitter (http://www.minger.net/2006/11/30/yahoo-launches-group-sms-service-mixd/). To be honest, you can probably find all kinds of papers and products by just searching for "group SMS text 200x" and replace X with a digit. It was a hot area of research at the time, and many group SMS services were sprouting up.

Point by point, here we go:

A method of communicating among users

Well, pretty much everything on the Internet is. But these systems described above are very much like Twitter, used primarily for self-expression, and sharing with one's friends.

One user (the "first" user) can "follow" another user (the "second" user).

LiveJournal had that, although it was typically called "friending".

The "second" user can post a message. The message is not posted "to" anyone, it is just posted.

LiveJournal qualifies. While one could read the LiveJournal of a user as if it was a blog, that was just one way of consuming the stream of messages.

All followers of the second user receive the posted message.

The notion of 'receive' is a little bit fuzzy here. Even on Twitter, users do not necessarily receive the message. Instead, they have an account which is configured to 'follow' certain Twitter accounts, and their devices may request portions of the chronologically-ordered combined stream. When Twitter is used in that way, through its web interface, it's analogous to LiveJournal.

However, in its earliest form, Twitter also instantly pushed messages out to SMS and continues to provide similar alerts to with its mobile apps. LiveJournal never did anything like that. To find an analog you have to look to the group SMS distribution systems I referred to above.

The first user can choose from a set of possible methods for receiving messages.

LiveJournal had that early on. One primarily read through the web, but an API as well as RSS feeds were added, and soon clients for many different platforms appeared.

1
  • Your point by point analysis is helpful but would be more helpful applied to the claim wording itself rather than the OP's simplification of it. Also, something does not need to be difficult to implement to be an "invention". Pointing out that new thing X is just old thing Y and Z combined doesn't make X not new but it might make X obvious. If X was what the world needed and takes off like crazy based on being Y + Z ( rather than a big marketing budget) that can be evidence that X was not obvious to do until it was done.
    – George White
    Aug 4, 2013 at 23:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .