Steve Gillmor chats with Chris Messina about GNIP and the politics of micro-objects with Doc Searls and Robert W. Anderson. Recorded Wednesday, July 2, 2008.

[display_podcast]

[music]

Steve Gillmor: Hi, this is Steve Gillmor, welcome to the Gillmor Gang. Today is Wednesday and we are going to have a conversation with Chris Messina and whoever else shows up from the gang, which right now means Robert Anderson. Say hello folks.

Chris Messina: Hello.

Robert W. Anderson: Hello.

Gillmor: Chris, tell us about this GNIP situation that you apparently are involved with?

Messina: Sure. Well, GNIP is a service that essentially launched I think, yesterday or something. There was some news on TechCrunch and Plaxo and elsewhere. And the idea effectively is a pretty simple one, obviously has relevance to sort of like Twitter’s ongoing availability concerns, but the service essentially is a way of collecting an essential place activities that people are taking and then pushing them out to essentially recipients.

So it is kind of using the pub/sub model of content distribution to enable much more efficient delivery of what people are up to. And so, they basically launched a service on an API that currently – let me see right now – they are working with Ask500People, del.icio.us, Digg, Urban Dictionary, Flickr, GetSatisfaction, MyBlogLog, Six Apart and I guess Twitter eventually.

And the other thing that might be useful to know is that the founders of this project came from MyBlogLog before. So, they have kind of built this type of distributed system previously. Of course, MyBlogLog is a service that allowed you to see other friends or folks who are actually visiting blogs and websites anywhere.

Gillmor: Yeah, even if you are unaware of the fact that your attention metadata is being stolen, but I always wanted to…

Messina: That is not true.

Gillmor: I will be here all week.

Messina: [laughs].

Gillmor: But, I think, we had a conversation – Nik Cubrilovic and I had a conversation with the principles of GNIP yesterday. And it is on TechCrunchIT by the way, so you should go take a look at that, it is a great video.

Messina: OK.

Anderson: Should we go do that right now?

Gillmor: Yeah, I will just sit here for 20 minutes in silence.

Anderson: Oh, we will catch up.

Gillmor: Good idea. So, what exactly is your involvement with the project because it is a little different than what you were just talking about?

Messina: Yeah. I mean I don’t want to overstate my involvement too much, but I have had ongoing conversations with Judd and Eric for sometime more or less around the actual activity streams themselves and around representing them in a way that provides some sort of semantic cook or whatever, so that we could start to get a sense for what activities are being taken by individual users. And this of course is coming out of the work that I am doing on the Dezo project.

So, it isn’t so much, from my perspective, working on the infrastructure of scaling opportunity, which clearly is where they are focusing, as well as I think, there is room for plenty of providers, whether it is Amazon eventually or whoever wants to tackle this kind of real time stuff or anybody does XMPP and pop/sub. But, I sort of wrote a post a little while ago about adding richness to activity streams and of course activity streams are becoming quite popular. They are the foundation of FriendFeed. Twitter itself is kind of just status update activities, the news feed in Facebook and so on and so forth.

So, we started a conversation around how we can figure out how to make activity streams much more portable, so that I can be publishing, let’s say, an activity stream on my own blog and then other people could subscribe to portions of that using pub/sub. For example, if they are interested in my photos, but not my tweats, there should be a way to doing that. That gives the subscriber a lot more control over what kind of activities they are getting and how they are being received, so if that makes any sense.

Gillmor: Yeah, it certainly does make sense. The questions about the technology are one thing and the other issue of course is politics. So, let us separate the two and we will talk about both of them. The technology, how does this get implemented; does this require a new standard? One of the things that Eric and Judd were talking about yesterday was the opportunity for them to essentially be an intermediary and create a sort of a Grand Central Station as somebody in the Chatroom pointed out, through which a standard might emerge. Do you think that is a realistic goal?

Messina: Yeah, I think, what is nice about what they have done with their API so far, that it is very generic actually and there is a real I think, opportunity to take the work they have done, just sort of saying, you know, activities are going to be a very important part of the web, just like we have items in our assess feeds, the activities that people are doing and then the links to those social objects at the other end of those activities are also valuable.

And so, insomuch as this is a recognition that the activity streams and what people are up to, can actually serve as a foundation of a centralized business, I think that that is good thing. From that, we start to actually see, well what are all the different sites and providers that are publishing activities and what types of activities are actually being done, and furthermore what classes of activities are being created.

So, an example here in terms of standardization is around the ability to group related or similar activities. So, it is one thing to say, “Oh Chris, post a new blog entry,” or “Chris, bookmark something on del.icio.us.” It is an entirely different thing to start looking at well, Chris is uploading photos to Flickr and SmugMug and those should actually be merged into sort of a micro gallery of like the last five photos.

And furthermore, if I have uploaded 40 photos from let’s say a trip, all those photos should necessarily be shown in that activity stream, so how do we actually condense some of these things. And I think that once we start figuring out the standard object types, the media types that are coming through activity streams, we can then start figuring what are templates that can be applied against those various generic media objects. And that starts to actually add a lot of richness I think, between sites, so you can build services on top of them.

Gillmor: All right, let’s bookmark that and come back to the technology discussion, but from a political perspective, what’s in it for any of the large players?

Messina: In terms of large players, give me a sense for who you are talking about here?

Gillmor: Well, you mentioned them all, Twitter…

Messina: OK, Twitter and Digg and all those guys? I think, the biggest thing right now is that there is a lot of expenses being applied to all the different, either API consumers or subscribers who are polling constantly for notifications of updates for essentially RSS or add-on updates. So, as Joseph Smarr wrote about in his blog post.

Right now, the state of the art is essentially to do polling, which means a web service or a feed reader will continually ask, “Hey, you got anything new, got anything new, got anything new,” and this will be on various intervals. So, it will be 10 minutes, 20 minutes, an hour, two hours, three hours.

And because of the scale in which that type of polling is happening, there is a lot of either API providers or service providers, del.icio.us in particular and Twitter and so forth that are having to push back and say, all right you can’t poll us that often and if you poll us as often as you have been polling us and you are really aggressive in this, we are going to shut you off.

And that is really a terrible situation for everybody involved. It is bad for the people who want to build core applications on top of these services and it is also bad for the people that want to create value in terms of the service that they are building.

So, in terms of what this gets them, GNIP basically will become kind of the aggregator of all these different activities and services and so forth and will put out standardized feeds, and I think, will bear the brunt of the infrastructure cost in two ways.

One is that for those savvy audiences or let’s say application builders that understand the pub/sub model of information consumption. They will be in a much better position to get near time or almost real time updates about what’s going on. So, in terms of, let’s see, Placta’s blog post, they said something like with Diggs, for example, previously in Flaxo pulse you might see Chris dug something an hour after I had updated, whereas with GNIP, they’re saying that they would be able to actually see that update within 60 seconds, because of this new type of, the new type of architecture for getting those updates.

Gillmor: That’s assuming that Digg is using Website XMPP to send information to GNIP, right?

Messina: Well, I’m making two assumptions. One is that either Digg is going to be using the GNIP API, which currently is not based on XMPP, or eventually will be moving to XMPP before the stuff. Essentially, if you do Hub Sub either the XMPP or via some other service and you’re not unfortunately technically enough not to bring in all the details of this, the updates are still going to be pushed out much more quickly as they happen, as opposed to having both services continually pull.

Gillmor: Basically, what you are doing is you are aggregating all of the pings into a single or into a hub, which allows the speed. You can have one rapid update as opposed to 100 very small…

Messina: This aggregated model where again Digg is being hit thousands and thousands of times per second, from all these other providers.

Gillmor: OK, so, now that we understand what the architecture looks like, at least the theory, what is going to engage the people who aren’t part of this announcement – For example – Twitter.

Messina: Did you say Twitter? God, you know.

Anderson: Can I answer that? I mean, Twitter doesn’t have to do anything and if you look at GNIP central web page, GNIP can or will be able to just take XMPP right now.

Gillmor: Yeah, exactly. But, as we will find out, I’ve had some long discussions about this with various parties and I’m going to be releasing a report on Tech Crunch IT later today, which describes why at least at this point, Twitter is either unable or perhaps unwilling to release that kind of activity stream via XMPT in a syndicate-able fashion. So, let’s just say that for now, Twitter is… although Eric… I’ve forgotten how you say his last name.

Messina: Marcoulea.

Gillmor: Thank you. Marcoulea indicates that they have been working with Twitter all along and that they are all good friends. I’m sure that’s true. That they don’t have any clear understanding yet with Twitter about them being at either end of this pipe. And I would suggest that that is probably due to the fact that currently Twitter is under such pressure to just keep their service up and running that there’s going to be some interval at which we have to basically give them the opportunity just to go away and be left alone while they try to fix things.

At some point, though, that’s going to be over and more importantly, I think, something that Chris alluded to earlier, which is the developer community and how they interact with these different nodes. This is starting to become a pressure, which I think, will not be as easy to avoid a resolution.

I’m trying to keep my words very careful here, but let’s just say that Twitter doesn’t really have at the moment any need to employ such a service except from a, say, technical survivability, availability perspective. In other words, politically they don’t want to do this. What would that imply to you, Chris, in terms of your perception of what’s going on there?

Messina: Well, you know, I thought a bit about this and I thought about strategy and tactics and so forth. I’m typically a big proponent of finding new technologies and then trying to get a bunch of people around them, especially if there is good people behind them. Actually Bruno Pedro is on the chat and he’s got a service called Tar Pipe, which I’m also pretty excited about.

In terms of whether or not these services actually get adopted, I think, it comes down to whether or not business or an organization or anybody like this, wants to see that control or the architecture of something, in a sense, give up some portion or some iota of control. In so much as Twitter already has a backing in XMPP and they are trying to figure out, I presume, how to make that scale and how to make that work and so forth. And they have had some issues with that.

It might be that there’s an opportunity for them to actually get this stuff right. I would think that working with GNIP maybe there’s a way of moving things forward together, but it’s unclear to me.

Unfortunately, and this is what I was going to say before, is that since Blaine has left, I’ve had very little transparency into what’s up with Twitter. I’ve got good friends there but I haven’t talked to them recently. So, I’ve love to hear what their take is on it is. How they expect to do some scaling things. How they are going to make stuff work.

Obviously, if Twitter today: hey, we’re moving to GNIP. We’re going to work with them. Obviously, we’re still going to have our own infrastructure but as of today, we are also going to side load and put a bunch of the activities streams into their system. We believe that their architecture is sound and we have been working with them to build this out.

That can have a lot of really positive ramifications, of course, for GNIP. As for where this goes with Twitter, it’s a little bit like Facebook continuing to build out their own mechanisms and formats and things like that. For example, building their own chat protocol, when clearly they also could have gone the route of doing XMPP from the start.

Gillmor: Yeah, I don’t think that’s a little bit similar. I think, it’s identical.

Messina: Well, right.

Gillmor: Like what happened with FriendConnect and Facebook chatting.

Messina: That’s what I’m saying, both Facebook and Twitter, in these cases, have so far decided to go it alone and to weather the scaling challenges. Facebook hasn’t had quite as many, but they’re not getting the pick as Microsoft IM or AOL IM and stuff like that. So, it hasn’t been that big of a problem.

Gillmor: Well, as we discussed at Super Nova, I recall you were standing there at the back of the room. I don’t know if you jumped in on this particular part of the conversation with David (I’m having another senior moment here).

Messina: David Moran or…

Gillmor: Yeah, Dave Moran. Thank you. They were willing to discuss this issue of FriendConnect right up to the moment where it became clear that this wasn’t a technology problem. This was a political or a legal…

Messina: Legal or political,-huh.

Gillmor: And I’m not suggest that that’s where we are with Twitter yet, but I think, the existence of a company or an enterprise like GNIP raises the question and, forget the user perspective, since this isn’t really user facing, except for it might actually work.

From the developer perspective, there’s going to come a point in time where third party developers are going to either sink or swim with basically the licensing architecture of how to access. as Robert put it, an XMTP feed.

That, from my perspective, seems to be the intellectual property that Twitter has. And they’re going to be very careful, as they should be, to provide some balance between the rights of their third party developers, which I think, is the fundamental secret to their success, in terms of grabbing a large community and their own business model of how they move forward.

So, the question is, it may well be that they will be able to go collect Eyeballs or Twitters for some period of time before they have to confront this issue with users, but with developers I think, we may be very close to some sort of a meeting in the center of town where some cards are played.

Anderson: Well, you know, it is not clear to me right now what Twitter wants to be when it grows up.

Messina: Yeah.

Anderson: And I think that we all thought we knew where it was going, but then it also started to fail, and then these services have gotten shut off. And you know, we are assured that, oh yeah they are working on it and we will have some of the stuff back, but it is still not clear what it is going to be.

Gillmor: You know, I don’t agree, yes all these services have started to fail, but they all fail in a certain pattern and all roads lead back to Track – they lead back to XMPP and Track. And I think that the GNIP announcements speak to this issue, which is there has got to be a way of being able to maximize the flow of this kind of data and do it in a way that isn’t quite so verbose or noisy or redundant, which is what is going on right now.

There is just a whole bunch of people banging away at their servers. They can’t sustain it. They shut it down. The services that are banging away at them get under pressure for what exactly is their value proposition. The whole FriendFeed transition that has been built by some of the more rabid FriendFeeders as moving away from, I think, as Dave Slusher calls it a, it is not part of his world anymore, Twitter.

I mean he overstates the case in order to represent the value proposition of FriendFeed, which is that it continues to be up, it has got a mobile client now, so it is easier to use. This morning there was an outage on Twitter and so I was over in FriendFeed for a few minutes using the web client on the iPhone and it was working pretty well. On the other hand, actually talking into that cloud and using it as the main thoroughfare, that is a very long road that people are going to have to go down before they fill FriendFeed with the same kind of interactive real time value proposition that Twitter has when it works.

And the real cusp of that is going to be – in my opinion, it is going to be what developer Fidelity is to Twitter. FriendFeed, on one level, is a third party Twitter app. Over a quarter of the data that goes into FriendFeed is Twitter.

Messina: Well, I mean, there is a number of things I think, at play here and one is that Twitter came about I think, at the right time with the right circumstances with an API that was extremely easy to use and that gave them a real I think, edge. You look at folks who even came before them like Jaikoo. And they had a number of additional features. They had groups and a lot of things that people kind of clamored for.

Twitter itself actually has not changed substantially since it was first created. The two features that it has really added are Twitter Replies and Track, everything else has kind of remained the same.

Gillmor: Track was there from the beginning according to Blaine Cook.

Messina: Well, then maybe it wasn’t exposed, but I mean regardless, in terms of major features I feel like – and search I think, might have been added at some point, but still summarize, I think, is a very capable competitor there, using of course XMPP. So, what we see though…

Gillmor: Be careful how you use those competitors because that is part of the issue here, but go ahead.

Messina: You are right. I guess, what I am sort of looking at is a general trend where again with the marketization of these tools, in a sense that I mean that more and more people have the ability to publish. And Twitter made it dead simple for people to get a lot of value out of publishing these 140-character updates. For example, when I introduced the notion of the Hashtag a while ago…

Gillmor: Oh, that is your fault, huh?

Messina: Yeah, it is my fault.

Gillmor: OK, we will take that up later, go ahead.

Messina: Yeah, I simply wanted a way of adding some context to individual posts and what I have been amazed at… I think, it is on Tweens or maybe it is, whatever, Hashtag or something. What I have been amazed at is that the Red Sox community is actually one of them who is dominant users of Hashtags as a way of coalescing and getting together.

And I think that there are a couple of things at play here. One is that Twitter provided such a low barrier to entry and through the design of it, the very effective, very low friction design gave a lot of people a way of saying “Oh what is this crazy weird thing that you got in your cellphone” or “What is this Twitter thing I keep hearing about.”

They are able to go there to sign up to get an account, add all their devices and all of a sudden they are able to stay in touch with those five to ten groups of people that they cared about. And now is essentially, when this publishing stuff actually really took off and so now people are taking this for granted across multiple services.

So, Twitter has the fortunate position of being the early lead in sort of capturing the [inaudible] surplus market, getting people really interested and working on it. And now we are seeing a lot of other services like Plaintiff, like BreakKite also offering this 140-character net publishing paradigm. So, that is common. And now, we are also seeing that again with little bit of marketization of these tools, more and more people are publishing.

And so, now it is no longer just about, gee what are my friends blogging about, it is like: oh I have got 300 friends that I need to track across all these different services and they are publishing more and more stuff, how do I actually make sense of and get value out of the things that my friends are publishing; how do I stay up to date with them in a very convenient way.

And Twitter is starting to fall off of that because you see the rise of FriendFeed coming up and saying well actually people are publishing a lot more stuff than just these 140-character status updates. They are publishing photos; they are Digg’ing things, they are adding videos on YouTube, how do I make sense of this. So, FriendFeed came along with a very Google’esque type interface and interaction model – obviously I mean that is where those guys came from – made it very easy for people to stay up to date with everything, didn’t even offer their publishing own tools really, besides the ability to comment on stuff.

And that is kind of where we are at, where we are dealing with this deluge of information by essentially subscribing to smaller and smaller chunks of information. So, in terms of GNIP’s opportunity, it seems to me that a large number of folks are going to be like: you know what, we really can’t handle thousands and hundreds of thousands of people polling us all the time, asking for updates, we’d rather just offload that with you guys where you become kind of essential provider and you guys manage that infrastructure and we will pay you to deal with that.

And the real tension here I think, is between GNIP having a business for providing the service for a lot of different providers who are not making their money let’s say out of certain types of reliability versus Twitter whose business model, I think, long-term absolutely relies on them being a central provider and providing a great deal of robustness through their APIs as you said.

So, I think that is kind of the way in which I am understanding this current trend towards not just micro blogging, but the notion of micro objects or the representation of social objects in a much smaller and more consumable snack size form.

Gillmor: Right and I would agree with most of what you said. I would push back a little bit on the notion that FriendFeed is somehow a broader environment because I think that using the TinyURL, using the ability to be able to launch different capabilities from 140 characters. Robert, is that you with your headset that you are breathing into your headset?

[crosstalk]

Doc Searls: No, this is Doc. I thought I had the mute on.

Gillmor: Hey, Doc. How’s it going?

Searls: Hey, guys. Hi.

Searls: … lurking because I’m dealing an interesting situation here. I don’t know if what you are talking about, Chris, is relevant or not. There is a fire here in Santa Barbara right now, and I want to set up in much the same way Ritter did with the Twitter kind of consolidation saying a really good way of covering the San Diego fires last fall.

It’s a small fire now, , but it could be quite big if it gets going. So, I kind of wanted to get ahead of the curve so I’ve been working on that this morning.

[crosstalk]

Gillmor: Just grab some oxygen and then go back.

Searls: Yes, exactly. [laughs] Yeah, probably some nitrogen and probably too much carbon as well. Who knows.

Gillmor: Well, let me finish I was saying with Chris Messina and then bring you into the conversation. Hang on.

So, Chris, what I’m basically saying is that using a TinyURL there is no reason why you can’t emulate and then encapsulate much of the FriendFeed conversation. I’m not suggesting that that’s what people would always want to do. But, my biggest criticism of FriendFeed is that by creating this larger more expansive space that they’re isolating so much of the interesting information there, and that the granularity of these micro objects, as you call them, is difficult to be able to extract.

And therefore, flow across the different activity streams, which is what you are talking about doing. So, I think, we’re close to the vision that you’re proponentizing, if that’s a word, which it isn’t. But, I think that there is some significant roadblocks, aside from Twitter’s reliability.

Messina: Yes.

Searls: I have a question, Chris. You said something about how the producers will pay money to GNIP to basically build and host that infrastructure. Is that happening now? Has Digg paid GNIP money for this?

Messina: I have no idea.

Gillmor: What Eric said was that…

Messina: That sounds like an actual business model.

Gillmor: Eric said that Digg is very interested in moving, in getting as broad an acceptance of their activity stream eventually as they can and that this represents an opportunity for them to be able to maximize or make much more efficient their output. So, they’re willing participants.

I think that, to the degree that they can reduce the traffic to their site, in terms of redundant pings, this is economically very valuable to them. So, I don’t think [inaudible] paying them money. I think, it’s more like the… And I think, what Chris said before and certainly what Eric and Judd said was they’re goal right now is to get critical mass and get users and relationships with vendors.

So, yeah. So, Doc, have you been following any of this? Are you on FriendFeed? I would imagine that you are.

Searls: I am on FriendFeed. I’m not very often on it. I actually fired it up while you guys were talking to see what I’ve got going on there. And I just posted a small thing. But, I’m not that active on it and I probably should be. I’m just not. I just got a lot to do, so recommendations welcome.

I want to find out more about what you guys are talking about, so I don’t sound too stupid talking to you.

Gillmor: All right, well sit in the background again. Go ahead.

Messina: The thing that I would offer in terms of FriendFeed and I’ve been hesitant about jumping on board with it as well, just because again it is yet another channel. There are a couple things. One is the super, super low friction, in terms of the interaction. You know, I am able to go on now with the iPhone interface and put things like and that’s all you have to do to express something.

That’s actually really critical. That’s the same thing that Twitter has with the ability to star things. And there’s actually a gesture involved there where you can actually get some value out of it.

Right now, it is interesting that, again, Twitter has all these super great opportunities and of course they have let a lot of developers, external developers, build on top of their stuff, but you don’t see a lot of good statistical analysis being delivered on top of Twitter from Twitter itself.

By that I mean you look at a site like Flickr and you go to the tags page and it gives you the tag cloud of recent activity going on on the site. They have a calendar view of stuff that has gone on. They have interestingness. And Twitter doesn’t actually offer any of these things. So, in terms of actually finding out what’s good or what’s happening, you have to go to external sites for that.

They don’t even have a way of saying, what is the top most favorite in tweets over the last 24 hours from your group of friends or from the entire Twitter ecosystem.

Gillmor: You can say that they don’t have this, but you can mine the data that’s available assuming that it can escape from their servers.

Messina: The point that I’m making….

Gillmor: You can derive all of that same information an…

Messina: Yes, that’s absolutely true. But, the point that I’m making is that they’re not actually imposing their own feedback loops on the service to make it better and more useful for other people. And so the comparison that I’m making is that FriendFeed is already starting to do that from the beginning.

And I think, you see from, the heritage of the FriendFeed founders coming from the Gold Newsreader or Google reader. They started very early using the Share functionality, using the favoriting functionality to drive recommendations, saying what’s cool, what’s happening across all my different friend groups.

And that’s driving, I think, more and more folks to find actually good, useful and relevant things within in FriendFeed, compared to Twitter where everything is just kind of a big river of stuff.

Gillmor: Yeah, but the problem with that logic, from my perspective, is that when you have an algorithm and particularly an algorithm that is supported by voting. In other words, an explicit gesture, that’s one thing. The problem with that inevitably is that it gets gamed by the participants.

You start to find people who are moving into an area in order to make sure that they can… I mean, this happened in the old trade space when IBM, who had hundreds of thousands of people working, they would enter into these online polls and just completely tells the results over to IBM products, because they had such clout and an army of people whose job via memo was to go out and do this.

Messina: The major thing that I would push back on that and I completely agree with what you are saying generally, is that this is where the social graph, as a personal filter, becomes very, very handy. And Twitter already has this. So, you have to explicitly follow somebody on Twitter in order…

Gillmor: Absolutely right. And the same thing is somewhat carried across with FriendFeed, so this is not an attack or a criticism of either FriendFeed or Twitter. What I am saying, though, is that harvesting implicit gestures is much more valuable in the long run, because it allow users to be able to influence and to create affinity patterns without having to be subject to gaming and the sort of tyranny of the commons.

Messina: Yeah. Yeah.

Gillmor: OK. So, that’s where we are right now. There’s a lot of pressure, exclusive of Twitter staying up or down. There’s a lot of pressure on the part of new entrants like FriendFeeds. The problem I have with FriendFeed’s attempts at building infrastructure or features that accelerate their user friendliness is that they are building that on top of a comments system which is, at the moment, at least, very difficult to traverse.

So, it is kind of accentuating the inability to be able to move through the information base, whereas in Twitter, you can characterize it as being this shit fest of – I know you are not being negative about it, but this torrent of unmoderated information.

There are signals, in fact you are one of the proponents of this Hashtag scenario, which is a way of providing context. There are ways of embedding, particularly I suggest in the TinyURL, I am using that as a Kleenex, not as a brand name, embedding information which would allow vertical groups to be able to form and talk to each other, so you can derive that information.

So, to suggest that Twitter is not functioning in the same way that FriendFeed is, I would agree and I am not so sure that it is a bad thing. They are leaving the canvas open and available for innovation as opposed to innovating at a single point, which is what FriendFeed is doing. If the two were able to talk to each other in a ecumenical way, it would be much more powerful than what it is today.

Messina: I mean, it strikes me that there is a another opportunity here, and this is something that I have kind of struggled with and actually goes back to some of the stuff I was trying to do with Flock. And the very simple idea that I had back then was that you would read all of your friends’ updates or whatever, essentially in what Flock has today as like your world or whatever, the home page thing with all your friends’ updates.

But, my vision there in building a blog writer into the browser was essentially to be able to see things that were interesting to you, to respond and react to them and then to post them to your own blog, your own individual blog somewhere as a blog quote citing back to the original source. And that through doing this very low-cost semi-semantically marked-up information, essentially the blog quote tag as a citation link back to the original source that you are quoting, you would be able to create a distributed mesh of conversations, essentially commenting from one blog to another and creating this wide web of discourse.

The issue right now in some ways in what you are talking about, with FriendFeed in particular, is the lack of portability of these comments; that the only convenient way of actually creating comments on an object that shows up in FriendFeed is to go to FriendFeed itself and to leave that comment on FriendFeed servers.

Now of course the real challenge here is that not everybody has a place to store comments. So, what else are people who show up at FriendFeed sign up for an account for the first time, is the first web account they have ever had that say – and they want to create a comment some place. Well, it has got to be stored with FriendFeed.

So if I were able to – and this is actually a very interesting thing that I have noticed now on Blogger where you can leave a comment – I believe this is the case – leave a comment on Blogger and then also what is called CC the FriendFeed object as well. I mean, I don’t know how this scales when you have 1000 FriendFeeds out there if we ever get to that point, but the notion is still there that these comment streams can exist in multiple domains and places.

I would definitely like to see eventually at least the opportunity or the choice to comment on my own blog service [inaudible].com let’s say that is where I store all my comments, but then also federate those updates out to whatever objects I am talking about, so whether the object is stored on Flickr or on Breakcape or on YouTube or on Magnolia or del.icio.us or whatever, I can essentially talk about any object in the wild that has a permalink. And that those social objects know how to go out and receive those updates and say, ‘oh by the way, here is the full comment text’ and we are going to link back to the original source of that comments. They can go back and find out more about the author and…

Gillmor: Oh, but not only that, but you could also by examining that link back, you could use that as an identity tag for the object.

Messina: Yeah, that is right, that is right, exactly.

Anderson: You know a large part of what Marc Canter has been talking about in his open mesh architecture, is the ability to go from some object somewhere and comment on it in any system that you want, write about it in any system that you want and have that output go to where ever you want it to go. And that is a pretty good thing…

Messina: And how we make that user friendly is the real problem.

Gillmor: I think, that is the…

Anderson: Right, and he does have ideas about how to do that as well about registries where you tell the world what your tools are that you want to use, but there is definitely issues about how you get, let’s say, FriendFeed to give you some kind of a URL or some kind of an ID that any old blog tool can use.

Gillmor: That is exactly the problem that I mentioned of granularity is that you can go down to the thread level by clicking on more and then selecting a link to that particular thread, but you can’t go to the item level. And to be fair, you can’t do it on Twitter either. When you reply, it is not to the item, it is to the stream of the Twitter.

Anderson: Yeah. I mean, you can, but it is harder.

Gillmor: Well, what do you mean you can, how do you…?

Anderson: Well I mean, you can find the URL for the specific tweat and you can use that as part of your tool.

Gillmor: But, that is hardcoded to the individual tweat, I see what you are saying…

Anderson: But, it gives you the time…

Gillmor: What Chris is saying I think, is that perhaps if you could relate the two so that there was a citation, maybe a tag or a field in the XML object that points back to the originating source. That could then be used by intelligent software – god help us – to identify the original micro objects and therefore allow people to be able to see how frequently that particular object is being cited, which would open up I think, the credibility of FriendFeed in a very large way.

Because among other things there is a lot of information that is being cited now, but it has only being cited on FriendFeed through a conversation where much of the material – I don’t want to overstate this – but a lot of the material, the signal to noise doesn’t justify trying to find these little nuggets. And right now, there is no way to be able to point into the nuggets.

Messina: You know, I am curious, I don’t know if Doc is still with us, but FriendFeed has this feature called Rooms and one of the thing that I think, is compelling about the discussion around FriendFeed and you know, I don’t think that FriendFeed first of all is the end all be all, neither is Netbur or any of these services. There is certainly room for competition.

But, in terms of collaboration and in terms of bringing people together, one of the things that is very useful about Twitter is how generic it is and how it serves kind of this general watering hole. And it is kind of a like a modern pub or coffee shop or whatever. FriendFeed has another version of that. It does of course aggregate a bunch of conversations happening within Twitter, but then it provides kind of secondary table, let’s say within the Internet bar or whatever – I know I am terrible with these metaphors.

But, the notion here that I am I guess asking about is, when people want to get together and talk about these social objects, what is the most efficient or effective way for them to come together presuming, let’s say, there was a design that we thought actually was the right design for distributed web, where we don’t have to worry about lock-in, where FriendFeed has permalinks to every comment and so forth and they are doing the thing. I mean, when does that happen.

Because I think, one of the things that Twitter is lacking is kind of like the breakout space, where you can kind of like go off and say: well, let’s have a conversation about the Red Sox or about FriendFeed or about for example the Santa Barbara fires. I mean, that is what the Hashtag idea was kind of about, was like saying well this is what I am physically talking about, but I only have one channel.

And that channel is my broadcast channel, that is ham radio channel, but I am going to make sure that this broadcast has this certain tag, so that other people understand that when I say, “Oh my god there is a fire,” they don’t immediately think that it is in New York or some other place, but that it is bounded.

Gillmor: Doc, are you there?

Searls: Yeah, yeah. I just turned the mic back on. It seems like what you are talking about is de-siloing some of the functionality one gets from FriendFeed and Twitter and siloing it in a positive way that kind of thing originating from ourselves. Is that right? Is that a short version of the same thing?

Messina: There are two parts. One is letting people use the tools that are most effective for them so they have a choice. So, if they want to use Twitter, FriendFeed, Gychu, Pounce or whatever, they can use those tools that are comfortable for them and then, as well, providing kind of a common room space, in which they publish related information. So, for example, in your case, and this is something that I’m seriously thinking about: how do people come together around the Santa Barbara fires?

They can get good information but where everybody doesn’t have to sign up for Twitter, or everybody doesn’t have to sign up for FriendFeed. Does that make sense?

Searls: Yes, it does. And I like that. And I think that some way that’s external to every commercial silo is a good thing. Even if it makes use of some of those things.

I’m not sure how… it seems to me like so many times the answer to every problem is yet another site. And I think that that’s the static web. That’s the old way and we need that. It’s necessary and sufficient, but there’s something that’s on the live web in a generalized sense, that is not locked into any one of those things and allows anybody to start adding useful value, external to any particular site.

If that’s a doable thing, and I think, it is, but I’m not technical enough to say that with certainty, so I’m just relying on guys like you to do that. So, I think, it’s very cool. At least that’s what I think, it is.

Messina: I think, there’s a balance between coming up the right conventions or standards, in which you codify behaviors or activities that we know have value, and furthermore have value across contexts. And that’s what I’m to push forward with the activity stream formats.

Then there’s also, and this is really critical, my background as a designer. We need to weigh convenience and the ability for people to come to a system, to know nothing about it and yet to drive value in a very limited amount time with limited effort. Because, of course, there are more and more sites out there that are demanding people’s attention and you just can’t make sense of them.

What we need is much more, kind of like disposable websites. For example, in your case, you want to be able to throw up a Hashtag and have people come together, let’s say, for the next 24 to 48 hours. Hopefully until the fire subsides, and they get good quality information from people who are actually experiencing this thing, where they don’t have to learn this technology, but it goes to whatever device they might have with them: whether it’s a laptop, whether it’s a cell phone, whether it’s an iPhone or GPS device, what have you. And I think that there’s a real need for that.

Searls: That’s exactly it. So, the Hashtag is SB gap fire, because gap fire is the name of the fire and it’s good for as long as that lasts. And there’s not going to be another thing called the gap fire or in Santa Barbara probably. So, therefore, that particular ad. hoc. license plate for that event will be self expiring in a kind of folksanomic way. And I think that’s a good thing.

And I think, the Hashtag is great way to do that if we can generalize it in a way that doesn’t isolate it’s a Twitter or FriendFeed or something else like that.

Messina: Yeah. I can agree with that. And I think that’s one of the things that’s interesting about the Hashtag thing is like, when I was at Reboot over the last week, weekend, we used the Reboot pen tag. That’s useful. There will never be another need to make use of that tag, so there will be this coalescing around this mime, maybe it’s a hash mime or however we want to call it.

And then, as you said, it dissipates again. I think that’s really important. That’s the way in which language, through these systems and these channels, needs to change. So, it’s great that you are already using a Hashtag. It will be interesting to see and this is what we saw actually with the SB Fire stuff. This is actually where we came up with the Hashtag. I looked across Flickr and I looked at the most popular tags in the last 24 hours and, of course, at that time, I had to go SD fire or something like that. And that’s how we came up with the tag. We looked at existing behavior, tags that people were already using for their photos and we brought that into the context of Twitter.

So, you can do that in terms of Pounce and do it in terms of Bright Kite and whatever else, that gives people a really useful way of just coming together, using, as you said, the folksanomic approach, letting people just kind of develop terms that will work with whatever context or limitations they are in.

Gillmor: Well, I want to summarize this because we’re past my self-imposed 11:30 limit. I’m liking the way these things are short and sweet. You can read the transcripts in about five minutes or less. So, it’s really bite sized and consumable.

So, I would suggest we were talking about here is moving data from inside silos to some sort of free-range space.

Messina: Letting people work with silos, sure.

Gillmor: Then we should probably rename this hash-pipe. I think that would…

Searls: That’s a good line. I like that.

Gillmor: Thank you. You’re welcome.

So, these hash pipes, we also need to toke up on the idea, pardon the expression, but seriously I’ll be here all week.

Messina: Yeah, I thought so.

Gillmor: With the notion that we also talked about a little bit earlier, which is that this payload, if it contains a URL reference to the originating micro-object, that that URL can then be mined for obscenity and recommendation information that would jump start this kind of alternate highway here.

So, I think, those two things need to be developed further and hopefully that, what I just said makes sense to you, Chris as you explore this activity stream that you are doing.

Messina: Yeah, I think, so. And I think, the thing that is most interesting to me right now is the notion of having activities, which have permalinks and are comprised of an actor, a verb and a social object. And that those pieces combined, create this micro-object or activity object, however we want to describe them. Such that, you can point to it and, as you said, it can be mined into its component parts. This becomes very, very interesting, in terms of creating almost like a molecular type of approach that these otherwise atomic pieces of data.

Gillmor: And one thing, one additional object. I was talking about Loren Feldman earlier today. He was talking about something he started to do, which I have been experimenting with along with Robert Anderson and others, which is the note functionality inside of Google Reader. That’s a little bookmarklet that you put on the top of your browser window and you click on it on any web page.

Then, you can add a note or not add a note and it will flow into your Google reader shared stream, which then becomes perhaps the citation back to the original object. So, we may have already some tools that we can use while we are building this idea. It also would have the advantage of shifting some of the infrastructure over to Google, which means it might stay up.

Messina: [laughs] Away from the nurture, yeah.

Gillmor: All right. Well, this is Steve Gillmor. This has been the Gillmor Gang, a fascinating one and I appreciate everybody who showed up and especially those that didn’t.

We’ll see you again tomorrow.

[music]

Bye-bye.

This entry was posted on Wednesday, July 2nd, 2008 at 3:01 pm and is filed under Gillmor Gang. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.