Track is Back The Movie

buchheit1I’ve been filming segments with various folks in preparation for TechCrunch’s Realtime Stream CrunchUp this coming Friday. One of these conversations took place last Thursday in the wake of FriendFeed’s announcement of what they call Realtime Search and what I call the return of Track. Paul Buchheit and his co-founder Bret Taylor have been on numerous editions of the Gillmor Gang talking about FriendFeed’s adventures in realtime, and since Bret will represent the startup at Friday’s event, I filmed Paul.

There’s always been a lot of pushback about the significance of Track, just as there used to be similar downplaying or pigeonholing about other transformative technologies such as RSS. In particular, Track started as an afterthought by a Twitter engineer that he coded in just a few hours. It was used initially by SMS fans to keep track of incoming messages of interest on cell phones, a reasonable proposition in the early days when Twitter’s flow was emergent and the dynamics of major news events still outside the scope of the service.

But for some of us who never really “got” Twitter until its social cloud started becoming more useful than other services, Track became the only way to interact in realtime when messages came in from points unknown. This quality of serendipity or discovery, combined with a realtime IM feed that allowed upstream messages in the same window, significantly expanded Twitter from a one to many broadcast service to a communications service rivaling not only the host IM service I used (GTalk) but also email and the remnants of collaboration tools built out around the Y2K period.

This capability was both shortlived and politically charged, highlighting as it did the tenuous nature of the relationship between Twitter and its third party developers. As stability collapsed, Twitter kept stripping out all but the core functionality, and both Track and IM fell by the wayside. Over time, it became clear that Twitter’s crown jewels were in fact Track and access to the full firehose of messages. Attempts to revive Track by independent developers were rate limited to the point where the roundtrip communications aspect was effectively locked out.

It could be (and is) argued that Twitter has now brought Track back via various API services, including at least three levels of flow. Some third party services have even been given access to the full stream, but that access is based on private relationships and could be altered or terminated based on Twitter business objectives. Certainly there is no current open process whereby developers can get unfettered access. Also, Twitter’s focus on mainstream brands in some of the API levels tilts discovery away from less well-known identities and toward those promoted by the much-maligned Suggested User Lists.

As a communications medium, this is akin to some phone numbers not working when called or more precisely, 911 calls not being heard on a variety of local phone systems in a national grid. While I see no problem with SULs or any other ranking systems, the lack of an open playing field as an alternative encourages the talk of a private company “owning” the micromessage bus. I don’t buy that one either — no one forces us to post messages to Twitter and they have every right to the fruits of the service they’ve created — but talk of Track being back as either an API to be leveraged by third parties or opaque business relationships not available to all comers is the talk of those willing to accept take it or leave it terms.

It’s not that companies haven’t the right to get what they can get for their assets, it’s that I have little confidence in a third party ecosystem where the keeper of the crown jewels can summarily gate or shut out those services without notice or rationale. As an example, in editing video for the CrunchUp, I went looking for a great utility I’d last used several years ago for converting between various formats and resolutions. Clicking the icon produced an upgrade requestor that announced the company had ceased operation and recommended a procedure to back up the last version.

With Apple’s upgraded iMovie that accepts most of the camcorder formats and simple tools for uploading to the dominant YouTube, the interchange tools of the legacy app are really not required. I imagine that may well have provided the impetus for the company discontinuing development. In the same way, with no confidence in Twitter providing equal access to the firehose without the chance it would be shut down or incorporated into Twitter or some favored partner if successful, the motivation for playing along is diminished.

That’s one of the reasons FriendFeed, along with Facebook to some extent and even Twitter itself, represent better bets than others. Even before Thursday’s announcements, FriendFeed’s realtime comments architecture has allowed exactly the kind of realtime communications possible when Track was initially available on Twitter. FriendFeed has demonstrated a rigorous commitment to openness in its API strategies, a financial strategy that does not require short term monetization (largely self-funded by Buchheit), and an aggressive pursuit of realtime services such as Track and stream splicing. As you’ll hear on our conversation, I tease Paul about the timing on fleshing things out, but behind the humor is the fact that the company has come thorugh repeatedly on their roadmap and promises.

It’s not clear whether Facebook has further to go down the realtime road in the near future, but already the move toward opening the newsstream to the Everyone designation is fostering a larger footprint in open micromessaging traffic. And of course there’s nothing to prevent Twitter from restoring Track themselves should either or both competitors prove successful with their versions of Track. FriendFeed’s forthcoming stream splicing and realtime APIs will make open Track available to both FriendFeed users and third parties.

FriendFeed’s open widgets and APIs may encourage Twitter to punish its third parties for working with the smaller service. But doing so risks drawing Facebook in with its enormous scale and its search tools, not to mention Microsoft and perhaps Google when Wave becomes more integrated into or around the Google Apps. Alternatively, Twitter can consolidate its growing scale and open more or all of its cloud to Track, minimizing the rationale for average users to move to what some call a too-complex interface.

Whatever the outcome, FriendFeed’s realtime services will within weeks provide a discoverable communications platform that should concentrate the cream of high value microcommunities on the service and in cooperating third parties. As Paul and I discuss, the tools for maintaining realtime value while allowing the community to police itself are top priorities for the service. Private groups allow conversations to be curated by the group, leaving dissenters a healthy role either within community structures or in those of their own design.

The intersection of realtime video and realtime Track will prove an immensely powerful platform. The experiments Leo Laporte and the Gillmor Gang participated in may be over, but if anything the pace will quicken in the coming months as more of these services become available openly. In the meantime, Friday’s CrunchUp will underline the speed with which Realtime is transforming the Net. Enjoy our conversation as an appetizer.

Here’s the full transcript of the interview via the Simulscribe API:

Steve Gillmor: Okay. So Track is back, right?

Paul Buchheit: Well, we’ve launched (Real-Time Search?), so I don’t know if that qualifies as Track or not but it’s very exciting for us because it lets you see the results stream in such instantly as soon as someone puts the new entry or comment that matches your query, it’ll appear instantly in all of the browsers of all of the people watching that search around the world.

Steve Gillmor: So, what is your understanding of what Track is?

Paul Buchheit: So, I’m… I think maybe that people have different ideas of what it means. So we think of it in terms of a couple of different features. So the real types are those that we have launched today. And, you can look at the searches all so you can embed it somewhere and updates instantly as it happens. It isn’t… If I post a new comment and it matches someone’s query, like they’re searching for Gmail or something like that, it will appear in their browser with less than a second. So it becomes almost like chat because it’s so instantaneous, so you can see those conversations occur at Real-Time. What we’re planning to do next will be just… at least a few more weeks still, is actually enable you to subscribe to those searches in the same way that it’ll allow you to subscribe to any other feed and that includes notification options such as Email, IM notifier. So for example, you know, I could do a search for people mentioning my name like an (Excelsior Polo Bagg?), things that have, let’s say more than one comment, that I could actually have that, let’s say, go to IM when that happens. And then, you know, if my phone is set up with IM, I could actually get that off my phone…

Paul Buchheit: In an instant someone mentions my name, which could be kind of fun. Obviously that’s somewhat (frivolous?) in this case. But, if you’re… If you have a product that you’re monitoring closely or a brand that you’re… Your job is to stay on top of, that can be very useful.

Steve Gillmor: Now, are you going to start decorating these searches with advertisement?

Paul Buchheit: Ah, not currently. We have no advertising plans right now.

Steve Gillmor: Well, I mean that would seem to be a logical extension of this technology…

Paul Buchheit: Yeah, this search has been proven to be a fairly good space but it’s just isn’t our priority right now, we’re focused on building out the product and adding this great features like the ability to subscribe to searches.

Steve Gillmor: What about the… I mean, obviously IM also flows that same stream into, let’s say, G-chat where it is archived…

Paul Buchheit: Right…

Steve Gillmor: Inside of G-Mail. So that’s the advantage that we’ve been waiting for, in quite a while for.
Paul Buchheit: Right.

Steve Gillmor: What about the ability to, what I would call feed splicing, to be able to splice this stream of data along with the home feed…

Paul Buchheit: Right exactly. So that will be what is enabled by subscribing it. So for example, I could do a search for (Tech and Friendfeed?), you know, for whenever those two terms appear together. And then actually subscribe to that search and say, I want that search, anything that matches that search to appear in my home feed or maybe I want to put it into one of my friend list and in that way, you know, I don’t have to go do that search every day, it’ll actually just appear along with all of the updates from my friends.

Steve Gillmor: Right. And then you can subscribe to… How can you get that… That composite feed out of the system?

Paul Buchheit: Well, when we launched it, it has the embedding feature so you can put that on a blog or wherever else…

Paul Buchheit: The ability to export this composite feeds, as you call them, is possible by the API or it will be very soon. But, another feature that we’d like to add… I don’t know exactly where this is available. Is that group can also likewise subscribe to Feed. So for example if you have a group that’s called, you know just (Friendfeed?) fan or something that they wanna do a particular search to pull in more content for their group, they can automatically subscribe to the group to that feed so that everything matching that search automatically appears in the group and people have a discussion there.

Steve Gillmor: So the group is the new term for what used to be called rooms.

Paul Buchheit: Exactly.

Steve Gillmor: All right. And then, so you could, you can also approach to turn those rooms private, so you can… And how does direct messaging, intersects with this?

Paul Buchheit: Well, direct messages are also searchable so, you know, search works with direct messaging and I don’t think that there’s, I don’t think that there’s too many other intersections between the two features

Steve Gillmor: So, in other words, the people that have visibility into private rooms would be able to essentially have a custom servers that became available to them as part of their overall usage.

Paul Buchheit: Right. Right. Right. So addition to be able to pull it into, in RSS Feeds or someone’s blog or something like that, they can also just pull in the searchers all in real time, so that, you know everything that matches that is going to come into the room.

Steve Gillmor: And, so you say that the API will be available shortly for this?

Paul Buchheit: Yeah. We’re working on an entirely new revision of the API that we hope will, significantly simplify development of (Friendfeed apps?) and we have internally kind of a (data?) version of it and we’re hoping to get out at least to maybe to a few people next week and so it should be released to everyone in within a few weeks. That will make available in every feature that we have today. So our current version 1API is missing a few other features that were introduced for the (Friendfeed?) design like, direct messages for example, and also obviously real time search, but those features will all be available in the (APIV2?).

Steve Gillmor: All right. How about the ability to, create rooms to create… Well, I’m sorry, groups, is that going to be available as part of the API or you had another…

Paul Buchheit: You know that’s the question I, I have to double check the docs. I think that may not be in there yet, so we can certainly add it.

Steve Gillmor: Right. I mean from a technical perspective, I mean there was some issue regarding, not being able to, I mean it’s sort of recursive issue with having RSS feeds coming from…

Paul Buchheit: Right…

Steve Gillmor: Coming from (Friendfeed?)…

Paul Buchheit: We don’t import RSS feeds back into (Friendfeed?) because that creates loops in the system which is a problem. And so, what we will allow you to do is actually just subscribe, explicitly support that feature of subscribing to another feed instead of going by RSS, in that way our system can automatically detect and manage those loops and prevent them from being a problem.

Steve Gillmor: So, when is that gonna happen?

Paul Buchheit: I don’t really have a date yet for that. But it’s a feature that I want myself very much so, hopefully…

Steve Gillmor: What would you want if this is featured too and…

Paul Buchheit: Yeah, I know. You know these things are always, it’s a little more, a little more subtle that you might think

Steve Gillmor: More than a week?

Paul Buchhheit: More than a week.

Steve Gillmor: Less than a year?

Paul Buchheit: Less than a year.

Steve Gillmor: Less than a month?

Paul Buchheit: I’m not sure if it’s more or less than a month.

Steve Gillmor: Okay. So now we know it’s at least a month and maybe 2 months?

Paul Buchheit: It’s probably right.

Steve Gillmor: Okay.

Steve Gillmor: Let’s shift gears a little bit to the sort of overall, you know, we’re doing this conference in 10 days and on real time, the real time stream. I guess you certainly now qualify on pretty much all grounds for, you know, a significant place in that build (down?) of the architecture. So, how do you think things are going in real time, and do you think that it’s, you know, flash in the pan, it’s just this year’s peer-to-peer or what?

Paul Buchheit: You know what, I think this is a fairly fundamental trend. You know, a lot of people have asked this question, like is it real time? You know, why does anyone want real time? But, the way I view it, it’s the original form of communication, you know. Real time is how people worked for hundreds of thousands of years, since language was invented. When we have a discussion, it’s in real time. It’s not, like, I say something and then tomorrow I hear your response. And so I think it’s just a more efficient way to communicate with the people. It’s a more natural way, and I expect that it will become more popular.

Steve Gillmor: But it also has the tendency, you know, as you might guarantee this, discuss the famously with our struggle. It has a tendency to create or accelerate potentially, the so-called mob aspects of communities. You know, shoot first and think later or not at all, that kind of thing. I’m not sure I agree with the overall premise, but I certainly agree that we can see this kind of behavior materializing.

Paul Buchheit: Certainly, certainly. Yeah. Anything that enables people to connect and coordinate more effectively is inevitably going enable them to do, to use that power for good or evil. But that’s just, you know, the same thing as true as with, let’s say a cellphone is now. Groups can coordinate much better and all the criminals have their cellphones. The freedom of assembly isn’t always going to be a, I guess a positive thing in everyone’s view.

Steve Gillmor: All right. So, what do you do about that?

Paul Buchheit: You know, a lot of it just comes down to giving people more tools to control their experience. So, one of the features we’ve talking about is, for example, you know, just letting people control whether they have comments on their entries, for example, if I just don’t want them, or limit who can comment. So, you know, if I don’t want the whole world commenting or maybe if a conversation is just going on for too long, and I just want to shut it down, you know, those kinds of features that I think maybe helpful to add, certainly even on my own I’ve had some… Yesterday, I posted some video, and then now it’s kind of drifted off into the world of 9/11 conspiracies, and so if I had that feature, I’ll probably just shut it off.

Steve Gillmor: And so, that sort of gives rise to blacklists and that kind of stuff.

Paul Buchheit: Sure. Well, we already have the ability to block users. So, you know, if there is someone who’s harassing you or is causing a lot of trouble, you can block them and they’re then unable to view or comment on your entries.

Steve Gillmor: Right, you know, that’s a little bit draconian for what we’re talking about here, which is more, you know, managing a conversation, as opposed to just making a, you know, binary on/off. So, you haven’t thought much about that?

Paul Buchheit: Well, I mean, I think that’s maybe not quite as big of an issue as… It’s been talked about quite a bit. But just in my… My users are friend. It’s pretty rare that the conversation gets to the point where I’d want to kill it because, I think it’s… there’s occasionally something that’s very inflammatory or very, just a problematic topic comes up. But, you know, that’s a very small fraction of the conversations. Most conversations never venture into that…

Steve Gillmor: Yeah. But again, I’m not really talking about killing a conversation, but rather, providing tools so that the community can manage it. For example, there are some people who, I just won’t comment in their streams because they have demonstrated, you know, propensity for editing things that they’ve said previously which orphans the comments that are in reaction to it, things like that. So, there’s a certain kind of self-management that users go through. But there’s… The tools are not fine-grained in terms of being able to suggest to somebody, you know, I’m gonna, you know, mute you for a little while here because, you know, you made your point. And I’d like you to move on. And if you don’t move on, then I’ll block you.

Paul Buchheit: Right.

Steve Gillmor: Or something like that, I’m not suggesting that this is your problem right now. But the passion that real-time engenders is definitely something that has a quality to itself that we haven’t seen elsewhere.

Paul Buchheit: Right, right. Yeah. I mean, I think there is still some unsolved issues there. I don’t know. I don’t know what all the answers are, but there’s also… I think we’ve learned that it’s, you know, important to keep a lot of these things relatively simple. So you have to be careful about adding too many of the different features. Because there is a, you know, a tendency to…for every problem that comes up I can think of, there are kind of 3 different things that might address it. But actually making it, you know, creating those features, maintaining them and then effectively communicating them to users, what their purpose is, what the right time to use them is.

Paul Buchheit: What the interface is and what it really means is can be a little bit of a challenge at times.

Steve Gillmor: What about… I lost it. Something about… It’s a feature I’ve asked you about on a direct message several times with no response, which is…what about commenting, being able to comment directly in-line. In other words, more, you know, threaded comments.

Paul Buchheit: Like a nested, nested comments.

Steve Gillmor: Yeah I, I’m not necessarily if, you know, if there was some sort of nomenclature for doing this, or some sort of signal where you could, you know, basically go in and say…you know for example, I liked this comment rather than I like…

Paul Buchheit: Commentary.

Steve Gillmor: That’s the example that we’re saying

Paul Buchheit: That’s actually like a great example of something that I think we all like to feature maybe, maybe me more than others even. But it’s just a matter of… we haven’t come up with the (UI?) yet that we’re all happy with. But, on the back-end, we’ve actually made a change, probably a month ago which is invisible to everyone but enables that feature among other things, the way that we store and manage comments was moved over to a more general purpose infrastructure so we can…

Paul Buchheit: …We can more easily implement that feature.

Steve Gillmor: And, you know, so is the answer that the (UI?), you’re not happy with the (UI?) yet?

Paul Buchheit: Yeah, the answer is that it’s largely held up just on figuring out the (U.I.?) that we’re all happy with. We now have, technically, you know, no reason that we can’t very easily add it. But it’s just a matter of figuring out the (U.I.?) that we’re happy with, and that, you know, won’t make the page too complex. You know, we don’t want to have smiley faces on every single line, for example. And so, we just have to work out what the right interface is for that. But, that’s actually true overall with comments that we’re going to need to spend, I think, a little bit of time looking at them and figuring out how to add maybe a little more richness to the comments. Because, you know, for example last week, we launched the feature where you can attach files to an entry. So, if you want, like a Word document, or just a text file, or whatever it is, you can attach it to an entry. Well, you know, logically, you should be able to do the same thing with the comment as well, right. You know, if someone has some other revision of it that they want to attach, that should be easy to do. And so, we don’t currently support that, but it’s kind of a logical feature to have, and something I think people would like to have. But again, we would have to make sure we can come up with a (U.I.?) that’s still lightweight, so it doesn’t become… You know, part of the reason the comments work is because it’s so easy to leave a comment. It’s just a little box like I.M. So, working out the right (U.I.?) so that it doesn’t become too heavyweight to just too much to manage is really the challenge there.

Steve Gillmor: So, we’re talking less than a day, more than a day and less than a decade?

Paul Buchheit: Yes. Yes, unfortunately, (U.I.?) problems turn out to be the hardest thing in many cases, because it’s a much more subjective thing than a lot of the technology. You know technology, if I read a piece of code, either works or it doesn’t work. But, the user interface… I might come up with something that I really like, but, you know, someone else on the team thinks it’s awful, or likewise, they might come up with something they love but I think is horrendous. And so, generally, as a team, we don’t typically release a feature unless we’re all, like, at least moderately happy with it. Because, you know, once you put something out there, it becomes a lot of times difficult to change it, because even if only 20% of your users like it and a lot of them kind of… It takes away from the experience. Those 20% will be very upset if you change the feature in a dramatic way.

Paul Buchheit: So.

Steve Gillmor: So, more than a month and less than a month and a half?

Paul Buchheit: I don’t know. I would like to have it very soon. But, I…

Steve Gillmor: The reason I…

Paul Buchheit: Yeah. We don’t… We work out a fairly short schedule so it’s very difficult of for me to give anything other than like these extremely crude guesses that it’s something that we’re interested in doing, but, at this time, we have just like a couple of features that we launch real time search, and our next thing in that front is that we want to do maybe to just subscribe to the searches, so that’s definitely the next step there. We’d want to get out the next version of the API. We just did themes, so we wanted to do a couple more things on themes. We wanted to enable people to theme their groups, for example, so that they can have like a more, kind of a brand experience there, and also just to launch a few more themes that are kind of on the. But beyond that, the next step is essentially we do this process probably once every two months where we just all get together and go through essentially all of the ideas that we have or that people have suggested and figure out what we think is the next step for the next couple of months. And so, I think we’re doing that, maybe next week. So, I don’t actually have, none of us actually knows the answer to this questions, because haven’t gone through that process.

Steve Gillmor: Well, I mean you know, now that you brought Track back, pretty much, it sort of begs these other questions. So, you, you know, this is the bed that you’re making and you have to lie on it. There’s going to be more pressure for you to innovate here.

Paul Buchheit: Yeah. I mean, I certainly agree though, with the idea of, like, the liking comments of pointing out the best comments. Because there is a lot, you know, really the best content (online?) feed is in the comments a lot of times, and so to be able to highlight the best ones is…

Steve Gillmor: Right. And I use and others use likes as a method of broadcasting over Twitter and other, you know, micro-message networks. So it, you know, it’s a way of not only telling people about something that’s going on, but also you know, through the (FFTILI Link?) pointed them back in to what is going on, but, you know, so much of the time this goes to this whole crowd, you know, behavior issue. You know, when you say I really liked when you like the post on the Holocaust. It’s like, you know, I don’t like the Holocaust.

Paul Buchheit: Yeah. Yeah. I did a little post on this once on FriendFeed where I did a screenshot of like, this is how you interpret it. We just kept the word short. And it said… I…the link…changed the link text from like to, I like that you have chosen to share this entry with us, which is, you know, more, in the spirit of it, is I liked that you shared it with us, not necessarily, I like the Holocaust, or I like, you know, whatever the tragedy of the day is.
Steve Gillmor: Have you seen any kind of reaction on the part of other, you know, developers in the marketplace to your (adventures?) in real time?

Paul Buchheit: I think it’s definitely raised people’s awareness of the feature, and you know, it’s hard to (tease?) a part to exactly, you know, what people’s motivations are, but there’s certainly a lot of interest in real time right now. There’s… We’ve had a number of people who contact us about getting access to real time data. They’re interested in the forthcoming real-time APIs. So, there’s definitely…definitely something that people are working on and there have been, you know, a few of these real-time search engines and (laws?) just in the last couple of week, though I think their definition of real time search is different from ours.

Steve Gillmor: Yeah, it’s real time according to however, what level of service that Twitter provides.

Paul Buchheit: Right. I mean, there is this real time meaning that is searching recent data where there is (.55) is actually literally instant, or, like, within, you know, within a second, as soon as something happens, it actually just pops out to your browser right there, so I think that’s fairly unique at this point. But I’m sure others will follow.

Steve Gillmor: You know, we’ve talked about this a little bit in the past, but the impact wave, you know, how do you see this being responded to? I mean, Facebook seems to have sort of moved in that direction and sort of backed away from it a little bit at this point

Paul Buchheit: With their (UI?)?

Steve Gillmor: Yeah.

Paul Buchheit: Yeah. I mean, but Facebook obviously has different constraints with, you know, a couple of hundred million users, and it’s… You have to move a little more carefully because people are all at different places in terms of technology and expectations and usage. And of course, they have a product that’s just tremendously successful, so you always have to be careful when you have something that successful. You don’t somehow break it because this sort of horrible truth about all successful products is that no one is quite sure why they are successful as they are, so you have to be careful you don’t accidentally, like, step on the goose that laid the golden egg. And so, as always, you know, start up soon in smaller operations or?? you’re able to to move a little faster and just take more risks.

Steve Gillmor: All right, just a couple of, sort of, check-offs here… We’re at the conference where they’re having a demonstration from the (PubSubHubbub?) folks, Brad Fitzpatrick and his partner-in-crime, who’s name I have forgotten at the moment, Brett…

Paul Buchheit: His last name is escaping me.

Steve Gillmor: Yes. Me, too. And, you know, basically the thrust of that is to, you know, bring RSS up to real time, essentially. And, as we’ve talked about, this is more complimentary than competitive with your SUP functionality.

Paul Buchheit: Right. They both have, I think, fairly similar goals, which is to be able to make RSS more real time by adding in the notification element. So, in traditional RSS, you just have to check back every hour or however often you want to track it. And so, that means most of the time there’s nothing there, and when there is, there is some latency of (potentially?) an hour or how often you’re checking. And so, both PubSub Hubbub and…

Steve Gillmor: Let’s call it RSS Hub.

Paul Buchheit: Yeah…and (3.39) are trying to solve this problem, and… But it is absolutely, I think, complimentary. At least it has some potential to be complimentary. The Hub product does this with a call back mechanism where you register interest and then they’ll actually ping you with the feed updates, which is actually, I think, a much easier way to consume the data, because as… Let’s see, writing a little app, it’s much easier to make an (HEP?) call and get that update. The purpose of SUP was to make it very easy for publisher sites that have large number of feeds to just expose that update information. So, you know, it’s literally something that (4.20), (SUP Support?), and, you know, just a small amount of time. And it works for private feeds, for example. So, let’s say, you don’t want to expose your URLs, you don’t want to broadcast them to someone else, SUP can still work for that, because it uses opaque identifiers for each feed, which means, you know, you can just pick, like, a random number and so, someone on the outside can’t tell anything. It doesn’t expose any information. So, a number of services use these secret URLs, like, Google Reader is one example where the URLs are actually secret. And also, sometimes, there are some services, even if the URLs aren’t secret, they still may not want to just emit the activity information for the sake of private users. So, you know, Twitter, for example, has private users, has this FriendFeed. And so, SUP is a great way for those publishers to very easily expose the update information. So, for example, YouTube has actually added (SUP Support?) a couple of weeks back. And, so, I think in my view of kind of, I think, the ideal combination is for the Hub product to actually also support (SUP?), because the two work well together. So, for example, if they added (SUP Support?), any publisher could subscribe to a feed that supported Sub using the (PubSub?) feature and so they would then get a call back from them and then (PubSub?) would monitor the SUP.

Steve Gillmor: Yeah. Is there something you can do to support them as well?

Paul Buchheit: Yeah. I think, unfortunately, I haven’t had a chance to read all of their spec and everything yet, but it’s something that we want to look at more, just seeing if, for example, any feed that supports their system of call backs, we could just automatically hook in to get those call backs as well so we can update it quickly. So, we are, you know, not tied to anyone, (protocol?). We’re happy to get updates by whatever means possible. You know, SUP, we launched last September, I think, or at least we announced last September, but we’ve actually been using, for example, the ping service, the blog pings,, also Google Blog Search has one. We’ve been doing that since February of 2008. So, we’ve been actually been already using those preexisting ping servers long before we came up with (SUP?). And we’re happy to find any other source of update information we can get. So, it isn’t one versus the other. Any place that we can get information

Steve Gillmor: No. It’s just the (7.01) moving to real time that’s going to expose, you know, the longer latencies of things like RSS going into your systems so, it’d be good to, you know, use this opportunity of the real time, the emergence of real time to clean up the, you know, the varying latencies on the network..All right, well, any final words about the dawn of real time for our conference?

Paul Buchheit: It’s exciting that everyone give our search a try, and let’s just know what they think.

Steve Gillmor: Thanks, Paul.

Paul Buchheit: All right. Thank you.