The Gillmor Gang – Marc Canter, Dan Farber, Dana Gardner, Mike Arrington, Robert Scoble, and Robert W. Anderson – drills Microsoft’s David Treadwell on the Live Mesh platform announcements at Web 2.0 Expo. Recorded Friday, April 25, 2008.
Anderson: It’s Robert Anderson
Gillmor: Hey Robert, how you feeling?
Anderson: I’m feeling alright, how are you doing?
Anderson: Who else is on the call?
Gillmor: I don’t know.
Anderson: So what’s new? Are you ready to ditch your iPhone so you can have Mesh FX running on your Windows Mobile device?
Gillmor: No, is Microsoft ready to have Mesh FX running on my device? That’s the question.
Anderson: Well the answer is no.
Gillmor: Well the answer is yes Silverlight is…well it’s a question of whether or not it’s allowed to run.
Gillmor: I mean I don’t care, if it doesn’t come in through there it can be broadcast over html, so.
Anderson: Yeah. Well I do wonder about Silverlight vis-à-vis, well, there’s a lot of questions I’ve got. Well, Silverlight as a runtime is fantastic, but I’m wondering where in relation to all this new stuff Silverlight is. It looks like Silverlight is an alternate reality in some ways.
Gillmor: No I don’t think so.
Anderson: I hope not.
Gillmor: Well I mean it’s not about hope.
Anderson: You know, reading through Ray Ozzie’s April memo, there’s too much in there about Windowsmobile, in my opinion.
Gillmor: You know, he has to speak politically as well, and if you look at Jon Udall’s conversation with Ray, I think if you take the two then overlay them, then you understand what’s really going on.
Gillmor: Who’s that?
Farber: This is Dan.
Gillmor: Speaking of which…the overlay himself. So I loved your story about the recession times in the technology community.
Farber: You mean the video we just did?
Gillmor: No I didn’t see that. I don’t get video on my…did you do it on YouTube?
Farber: No, we did it on CNET, we just shot a little video…
Gillmor: Right, but do you have a YouTube version?
Farber: Yea, well a lot of it goes up on YouTube, I’m not sure if that one is up at this point.
Gillmor: Well if it isn’t then I can’t watch it cause I’ve got an iPhone…so you should, you know, consider that.
Farber: Well we do have a CNET TV YouTube Channel.
Gillmor: Excellent. So you should put all of your stuff on there. So that iPhone addicts like myself have an opportunity to look at it.
Farber: Good idea.
Gillmor: Ok great. Who else is there?
Canter: My name is Marc!
Gardner: Dana Gardner.
Gillmor: One at a time…so Marc Canter who else?
Gardner: Dana Gardner how are ya?
Gillmor: Hey Dana how’s it going?
Gillmor: So we’re gonna be having, what’s interesting is that the USTREAM feed seems to be..well the only thing “quote” interesting that’s on it is me, and boy, am I not interesting so…maybe Jerry you could put something on the rest of the screen or else put up a blank screen. Alright so Dana, what’s your story?
Gardner: My story is that I was not at Web 2.0, I had too much work to do at home and I’m gonna be travelling the next two weeks, but I felt like I was there given all the blogging and Twittering and micro blogging, and other things…and Dan was really cool. I liked how he did through Twitter the Marc Andreessen-Jon Battelle discussion, that was just good enough without getting the gory details, but it was a nice medium.
Farber: Well I think it presents some challenges because if you’re there doing the reporting and you have to file the story you can’t really twitter it at the same time. I think we need more people, you know, someone to write the story and someone to twitter it online, someone to take pictures, someone to take the video.
Gardner: Hey that sounds like a staff, real payroll.
Farber: I just do it all myself.
Gillmor: He’s just got to shame his employees into working harder.
Farber: Yeah, I don’t think that’s the case its just I think you know, one of the things you do when you go into a new organization is to, to figure out you know, “How fast can we run? And do it so we’re not burning ourselves out.”
Gillmor: Yeah but you know, you’re sitting there at parties taking notes.
Farber: Yeah well actually I did file a story from the NetVibes party on my Treo.
Gillmor: Really? What did it say?
Farber: How NetVibes is going to open-source all of its technologies.
Gillmor: Well what does that mean?
Farber: Well right now, they’re proprietary widget platform with a bunch of API’s. So it’s a way to…almost like taking a page right out of Sun [Microsystems] as a way to potentially grow their market in the face of Google Gadgets. They’re trying to get more developers through open source technologies to work with what they have. So that was Tarik Krim’s latest move. As you know PageFlakes got acquired by Live Universe, so that leaves NetVibes as the lone, independent company not affiliated not affiliated with Yahoo! or Google or Microsoft, AOL, etc.
Gillmor: And what do they do again?
Farber: They do personalized home pages, so that you can create a page and they have thousands of widgets you can put on it…very much like My Yahoo!, or iGoogle.
Gillmor: Well what about Facebook and MySpace, aren’t they personalizable?
Farber: They, well…similar, similar. See you have different usage…I mean there you go as kind of, hang out with your friends…and this is much more generalized. In fact you can’t have widgets from Facebook and other things too. I think what it is about the atomization of the web, and it’s these loosely coupled pieces joined.
Gillmor: Marc Canter what’s the story?
Canter: Well, um NetVibes and PageFlakes redefined the start page, it’s a form of an RSS aggregator in that what you’ve done is you’ve got all these little modules on your page which are these little RSS feeds. Tarik makes his money by charging for placement of commercial feeds in the feed directory.
Farber: Sponsored widgets is what, you know.
Canter: Right. So that’s the business model. They’ve got a lot of money from Accel and Benchmark…they are one of the…one went to Benchmark…PageFlakes is one, NetVibes is the other…they had a battle, it appears that NetVibes won…But NetVibes has always been focused on the internal on your own internal…what is your private experience. What they’ve never figured out really is what is the public shared experience? So NetVibes is really the public one-to-one relationship with you and the NetVibes service…and in that sense it’s very much like an RSS aggregator. Neither of them, PageFlakes attempted to, but neither of them have been able to transition to the public phase, which is what a MySpace and Facebook are all about.
Farber: Good explanation.
Canter: Now the Times of India just put out a social network that has a NetVibes type desktop to it. Basically NetVibes is like a dashboard. It’s your personal dashboard and so the Times of India decided to combine that dashboard with a social network that’s called ITimes. That’s just been launched last week.
Gillmor: Alright I’m gonna refresh my video Jerry because I don’t see the streaming chat. Maybe I can do it some other way.
Farber: So are we multimedia today Steve?
Gillmor: Jerry Schuman has been multicasting this Gang live for 3 or 4 weeks and he’s doing it again today. That’s what I’m saying you might lose me I might have to go back and reinitialize.
Canter: Dan congratulations on, I don’t know if it’s official but I saw some rumors about Yahoo! and CNET content…sounds like a pretty good match up.
Farber: Yes, we’re going to power the technology channel on Yahoo!
Canter That was a pretty dismal page…don’t mean to insult your partner…but I only went there twice and there was no reason to come back.
Farber: Yeah well…that’s not their expertise, I think they do a good job in sports and finance, but I think technology is not one of those areas where it’s easy to aggregate something and just throw it up.
Gillmor: So you mean in general, the only thing that Yahoo! is good at is sports?
Farber: No, I didn’t say that…
Gillmor: Yeah, Sports and finance in other words…and the rest of it like technology they’re not good at.
Canter: Snow reports!
Treadwell: If they were really good at technology they would staff it up and do it themselves, right?
Gillmor: You mean, avoid being bought, which they can’t do?
Gardner: That’s a different story. But no, it didn’t seem like they could either satisfy the consumer side of the tech or the business side of the tech, they really weren’t doing either. Sort of somewhere, mesh, mix in the middle. It didn’t really work for anybody. I’m sure that will change now that Cnet’s gonna be driving some seriously good content in there.
Gillmor: Alright we’re going to have David Treadwell of Microsoft and the Mesh team joining us in about five minutes or maybe less. So is there anything else we want to talk about before we get into the Mesh world?
Gardner: The failure of Web 2.0 Expo.
Gillmor: Why is that?
Gardner: I’ve heard from a lot of people that it seemed like a really boring show.
Gillmor: Well it was a really boring show last year but the show floor, the…where the actual business goes on was packed last year and seems to be pretty good this year as well.
Canter: So it’s really an unvirtual social network, not really a news making trade event.
Gillmor: well I don’t know what trade events are news making these days. Microsoft had their Mesh announcements pretty much off campus. They had a couple of announcements on the main stage thing but most of it was, I think Dan Farber had a great story this week about the kind of messaging that’s going on here in this recession…
Farber: There are a couple of themes. One is that if you went to the exhibit hall you saw a lot of companies: IBM, Oracle, F5, and obviously the little companies on the fringes. The company that owns the show which is O’Rielley and CNP are basically figuring out here’s how we make money on this and let’s talk about Enterprise 2.0 because there’s more money to be made there. Obviously there are a lot of good conference sessions for people on various topics ranging from AJAX or working with Facebook or developing Mesh applications, whatever, and that’s probably the most useful stuff that’s not news but more tutorial. And then in the keynotes there were a few highlights. One was a bad interview with Marc Andreessen in which John Battelle spent most of the time talking about past history as opposed to what’s going on now, what are you doing with that $100 million. And Andreessen talked again about the coming nuclear winter and so he’s squirreling away his $100 million to buy some rice, or something like that. But it’s kind of confusing because while if you look at the earnings for this quarter, pretty stellar earnings for Apple, IBM, almost everybody and yet the economy is kind of sucking wind. So I think this is the first time you can say that people are tech immune because people really can’t live without it and they’d rather buy an iPod than rice. I think I’ve seen surveys where consumers think that we’re in a recession or in a serious downturn but spending is mostly down, but it’s actually not down if you look at consumer and IT spending on companies that are in the IT space it’s actually up but down in healthcare and pharmaceuticals and other areas. So a bit of misreading but I think you have to believe that if things continue that people will start pinching. And the people who are gonna get hurt the most will be the people who depend heavily on advertising from lots of little people.
Gillmor: Well there was something I read this morning which suggests that the venture capital community is drying up.
Farber: It is but I looked at some numbers from DowJones and Ernst and Young. And they said that it was down about 7 percent. But it was up 29 percent in consumer, up 29 in IT but way down in healthcare, energy, pharma. So it’s kind of a good and a bad story.
Gardner: but you think the Internet is countercyclical this time.
Farber: I think to some degree. I think that if there’s a bubble it’s not that frothy.
Gardner: Right. There’s also a lot of build out in IT spending in the emerging Markets: Brazil Russia China India. And of course the valuation of the dollar is helping a lot of US companies…
Farber: Yeah, and I think those companies who did more than 50 percent of their business outside of the US are in much better shape.
Gillmor: So which is it? Is it that VC money is down as it relates to…it wouldn’t surprise me to see that VC money is down relating to the tech, particularly the Web 2.0 industries.
Farber: Well that’s not what the first quarter showed. But I’m sure that as this downturn goes on that people are starting to really tighten up. But you get these kind of flukes where somebody throws 60 million dollars at Ning makes it look like there’s more money going around than there is because it’s only going to a few companies.
Gillmor: Well speaking of the moment when Microsoft reenters the game when the venture money starts to dry up, David Treadwell of Microsoft, welcome.
Treadwell: Hi! Thanks very much for welcoming me.
Gillmor: Well that will be the first and last time right?
Treadwell: At least I got one.
Gillmor: As per our conversation the other day, I will now go into my grouch character.
Treadwell: That will keep it fun.
Gillmor: Oh so you’re ready for this?
Treadwell: Bring it on.
Gillmor: First of all, tell us what your title is, not that it’s important and how you’re basically driving the Mesh team for Ray Ozzie.
Treadwell: I’m the corporate VP of the Live Platform Services team of which the Live Mesh effort is a big part. We’ve been pretty pleased with the announce that we’ve been able to do the last few days and the reception to it. So that’s gone pretty well we think.
Gillmor: When you say ‘pretty well’ what were your expectations?
Treadwell: Here’s what I was worried about. What we gave out this week is a set of platform experiences. Some scenarios for end users that we think are really interesting and useful. But we were worried about the message of the underlying platform which is really the big thing that Mesh is all about, the infrastructure that enables scenarios like synchronization and collaboration. We were worried about whether we would successfully get out that message. And it seems like that message really has gotten out. In interactions I’ve had with a lot of influential’s like you guys, and the interactions that the press has encouraged, the fact that this is more than just the experiences, that it’s about the platform that has underlies it. We’re happy how this has come through.
Gillmor: Who’s joined the call?
Scoble: I’m here, Robert Scoble.
Gillmor: Where are you calling from?
Scoble: I’m at 7200ft overlooking Yosemite Valley at Glacier Point.
Gillmor: Excellent. So we…you may be intermittent right Robert?
Scoble: I’ll be here for a few minutes and then if we have to leave then I’m gonna leave cell signal. The cell signal’s great here right on the glacier point, we’ve even been doing live click videos. We’re here with Michael Adams who’s Ansell Adams’ son who’s been telling us interesting photo geekery stories about his Dad and his family.
Gillmor: We’ve got David Treadwell here from the Mesh team…
Treadwell: Hey Robert how you doing?
Scoble: Hey really well.
Treadwell: Doing real well.
Gillmor: So do you have any questions for David?
Scoble: The one thing I was still wondering was how does the mesh deal with multiple devices and one of the devices has a huge hard drive with a lot of files and I drag it into my mesh into my folder. How does it decide which of those files is going to get put on the other devices which might not have a terabyte hard drive for the storage?
Treadwell: Good question. The way it works is that the user ultimately has the decision to map which mesh folders to a given device. What we find people doing is segregating their data into different data classes, each class in a mesh folder. On my laptop for example, I don’t try and put some of my huge folders, because the thing’s just got a 60 gb hard drive. So it allows the user to decide which folders get mapped to which machines. Today the onus is on the user to decide…
Gillmor: But that doesn’t really answer the question which is what happens if the…don’t you have some way of detecting what the client’s resources are and how they’re growing and shrinking depending on usage?
Treadwell: Well we certainly technically can understand the client’s resources but if a user selected, for example, they had a folder that had half a terabyte of stuff in it and they tried to map it to a machine that only had a lot less storage than that. I think what would happen in that scenario is that we’d use as much as we could but then eventually be unable to move more down to the hard drive and it would start seeing failures in that instance.
Gillmor: So, you can’t use computers to determine if you’re about to have a failure message and stop?
Treadwell: Yea, we certainly could use computers to determine that users are trying to map more data than they should. Actually, I’m not sure what the current infrastructure does under that scenario, but it would certainly be easy to tell the user if they put more data on the machine than the machine can possibly contain.
Gillmor: Or the mesh basically understands the, you know, profile permissions that the user could input which would suggest that they want this kind of data but not that. You know, how the data conflicts works in Groove was an algorithmic function not a hard coded one.
Treadwell: That’s actually a good suggestion to do something along those lines, where you know we could have some sort of intelligence to tell where to map which file where. One thing I want to find out actually is, and this isn’t a current feature, but it’s in the pipeline: is the ability to do different versions of media files on per device basis. For example, we were talking a lot about storage… you’re going to want your 10 megabyte version of your photo on your cell phone. There’s no point in having a 10 megabyte photo since all you need in 10K. One of the things we’re talking about is having an intelligent system so that it will put an appropriate version of media file on the device. The capabilities aren’t exceeded, but you are getting the most appropriate experience on the device… if that makes sense.
Sanderson: Yea, that totally makes sense because the scenario I was thinking was photos: I have 10,000 high resolution photos and they are all sitting in one folder. I need granular control if I’m going to hard code it, first of all. Second of all, having transcoding so it works on cell phones or all 10,000 photos from my cell phone would be really, really cool.
Sanderson: Yea, I think as digital photography gets more and more common, having 10,000 or so photos on a hard drive is easy when you have a terabyte or half a terabyte. But cell phones, you only have a few gigs or something… so, you need to be much more precise.
Sanderson: It does beg the question though, where you’d store the transcoded versions and if you could transcode on the fly which could be expensive.
Gillmor: Just for David’s information since he doesn’t know, that was Rob Anderson asking the question.
Treadwell: Great thank you. Yea, probably what we’d do is have a cached version of the transcoded photo. Getting into the video scenario, transcoding would be very expensive as you know… very expensive on storage, etc. Certainly we’d need mechanisms to store transcoded versions so we don’t have to do it on the fly which would be way too expensive.
Gillmor: Sorry, we were just having a conversion (Dan Farber and I) about some videos he’d spent producing at CNET. I asked him whether or not they were available on YouTube. Now, what would prevent you from transcribing to the YouTube format? Politically, we don’t do that?
Treadwell: So, you mean sync automatically to YouTube?
Gillmor: Not sync to it, but be able to create a downloadable version that would work on the iPhone. The way the iPhone works now is there is a link in twitter that goes to a video file that is encoded in QuickTime and it [the iPhone] opens up a YouTube container and it plays on the screen.
Treadwell: Ah, I see what you’re saying. Yea, we definitely want to get there. I want to be direct, but we don’t have that ability yet. Having that platform infrastructure helps so adding it won’t take too much work. The kind of things you are talking about are person weeks of effort, not years of effort.
Sanderson: You could imagine you can use an account, Flickr or something else, and that’s one of your devices.
Treadwell: Oh yea, we’d love to have the iPhone capability on this thing I think it would be very compelling for a lot of people.
Scoble: The question I have is the opposite of that. I just put up several quick videos and I want those to be put into my mesh and then stored on various devices so when I get home, I have all my videos ready to be played on many devices. On my media center… or my Xbox or my Smartphone or iphone. We can certainly get into how far the device compatibility is going to get right now. So, there are two separate questions. Are you going to be able to subscribe to an RSS feed that has enclosures and suck things into the mesh?
Treadwell: Yes, it’s all going to connect with the mesh protocol. So, you can connect feeds each device to your mesh and synchronize across all your devices using its normal infrastructure.
Gillmor: Isn’t that the mesh vision? Isn’t Mesh about abstracting out the client… you know, you write something, and there’s some information that hits the network. And Mesh is intelligent about expressing that on different clients so that it’s uniform.
Treadwell: Yea, that’s a pretty good way to put it. So, basically the mesh allows you to connect each device into the folders and data services in your mesh. And then mesh handles automatically moving that information around to the devices that you want. So that the user doesn’t need to think nearly as much about: is this piece of information on this device.
Sanderson: What kind of device, a lot of my friends were asking it seems like a windows only thing. A lot of my friends were asking, I think Techcrunch had a headline that made it seem a Windows only thing. I know that’s not true because you showed me a Mac version. Are you going to be able to hook into my iphone or nokia phone and bring the mesh to other non Microsoft devices?
Treadwell: Let me say a couple things about that. One of the basic philosophies of the mesh work is that in addition to working super well with client devices it also works well with a web browser. So any device that has a web browser can access the mesh, upload and download etc. so we run tests of Firefox running on a Linux machine talking to the mesh. So anything that has a web browser is able to participate in the mesh. In order to get to the synchronization capabilities, there is a need for some infrastructure running on that device. Our vision is, all the devices that users would care about would be able to work well with the mesh. As many devices as we can. We’re not going to make it work on TRS80 model 1’s but of course we’ll need to be able to write and install code on the device so if the device has an SDK shall we say that will be a constraint on our ability to get it working on that device. But as long as people are willing to open up devices in a way that lets us put some code on there to do these synchronization functions we’ll pursue it very aggressively.
Gillmor: David isn’t that kind of ironic that Microsoft is looking for other people to open up their architecture.
Treadwell: I’m not sure I understand. Windows is an incredibly open architecture you can plug in all kinds of different devices.
Gillmor: I’m only being semi snarky about it. I think it’s sort of emblematic of a new kind of status and balancing on the network that Microsoft is actively looking for ways to be able to join the cloud as opposed to the cloud joining it.
Treadwell: That’s one of the philosophies of the home mesh thing is to connect the cloud and the client devices in a way that is very natural it obviously requires a fair amount of openness at the protocol level by virtue of basing mesh on tons of open protocols like TCPIP, http on top of that, XML on top of that, feedsync, atom, rss, app, etc. on top of that. So I really tried to embrace the standard protocol to make things super interoperable. And another thing, by virtue of embracing these interoperable protocols we’ll work hard to get it available on many devices ourselves. But there’s no reason why other entities couldn’t make use of these protocols and build the devices from the mesh and interact well with the mesh without any help from Microsoft.
Gillmor: OK now that’s two supers you’ve outranked your quota.
Treadwell: Shoot you warned me about that. I’m sorry.
Anderson: So David you talk about the synchronization capabilities required to have an SDK on a device to have the device themselves to be able to synchronize. Do you guys have a synchronization runtime if you will?
Treadwell: In essence that’s what the mesh is…it’s a runtime that…
Anderson: but I mean to run on an actual device. And I don’t mean in any way to denigrate.
Gillmor: Robert Scoble can you mute?
Gillmor: I can see that Mike Arrington has joined. But go ahead.
Anderson: But is there a runtime and what does that runtime look like?
Treadwell: Well Mesh itself, there is a that you install on the device. And so that, the answer is, yes we have a runtime and that’s the thing you install on a specific device is the runtime that does the synchronization and the communications and the like on your behalf.
Gillmor: And what’s that written in, .NET?
Treadwell: There’s a ton of .NET code in there, a lot of it’s based on C-sharp and managed code of course and then some of it’s native code because in order to interact with certain parts of the system, like the shell, there’s only native interfaces. So we’ve been pragmatic about using a combination of both.
Gillmor: And so that’s separate from the Silverlight runtime, it can’t depend just on the Silverlight runtime for example.
Treadwell: We do make use of the Silverlight runtime in the Mediaviewer in the live desktop on the web and we’re enthusiastic about further embracing Silverlight as aggressively as we can.
Gillmor: Cuz I’ve been promoting this vision, although obviously I’m not directing Microsoft strategy, the Silverlight runtime being available on more and more devices would enable things like mesh, like office mobile to run on non windowsmobile devices for example. But it sounds like this is a parallel strategy to that.
Treadwell: I really view mesh and Silverlight as orthogonal and complementary technologies. Essentially what the mesh client does, it’s the runtime for doing synchronization and collaboration those kinds of things. I’ve used Silverlight as a runtime that does the presentation engine. Mesh doesn’t really have anything for presentation, Silverlight doesn’t really have anything for synchronization and mobile communications. Working together I think you have a very good thought there about the combination of these and how they’ll come together. We’re working actively on that but we don’t have all the I’s dotted and t’s crossed.
Gillmor: So the real question is: whether or not there’s gonna be a transport to take Twitter messages in through the mesh, be able to assign them to multiple affinity groups, and then express those individually on Silverlight clients that can synchronize with each other.
Treadwell: That’s eminently possible. I don’t know if you had a chance to see Orameega’s (spelling) talks down at Web 2.0 but one of the things he demoed is integration of Twitter, with no changes required from Twitter, which makes use of standard protocols with the mesh.
Gillmor: Mike Arrington you got a question?
Farber: I have a question. When you look at the bigger picture of live mesh and what you’re doing as well as with Silverlight, do you conceptualize this as providing a pervasive layer of synchronization and low level communications functions for the web itself overall as well as a presentation layer with Silverlight so that effectively people can think of this as, ‘remember old Windows it kinds of sits at the bottom.’ Well this is what we’re emphasizing right now as the…what Microsoft’s platform really is about.
Treadwell: You’re right in the sense in that this does represent a pretty significant advance, we feel, on Microsoft’s platform strategy by virtue of bringing the web to Windows and Windows to the web and connecting those super well. We don’t feel that mesh should be a required piece of technology. There’s no intent to make it such that if you don’t have mesh then the web’s not going to work well. We do view it as something that will enhance a user’s experience and by virtue of depending so heavily on these protocols like RSS and feedsync and the like, we really want to make it easy for anybody to connect with it and for it to, for mesh to connect with other services like Twitter that may not have been designed specifically designed to work with mesh but that’s the beauty of the interoperable standardized protocols is that they enable scenarios that the original designers maybe hadn’t even considered. But when you have that tremendous interoperability in the protocols and everybody basically has similar ways of talking, all kinds of really compelling scenarios can fall out of that.
Gillmor: This sort of brings up the central question that I’ve been not asking so much as sort of projecting into this which is, when somebody says well when’s the Mac client? I don’t care when the Mac client comes because Silverlight’s already here and it has a Mac client. The notion that these open protocols are going to be good citizens on the web, that’s certainly obvious. But most of the successful web startups that have come along and achieved a lot of mind share if not actually commercial share like Facebook etc., have used these kinds of open protocols, including SSL and user password authentication in order to get on the system. So it’s not so much that this is…I understand how politically for Microsoft it’s a good idea to be completely transparent and open as a message to developers that they’re not going to get locked in but more important is, how is this going to effect the revenue models of the various products along the Microsoft channel such as Office, etc. It’s pretty clear what we’re interested in is what is Microsoft going to do to keep the revenue flowing while at the same time offering a competitive challenge to what Google’s doing in the on-demand office space.
Treadwell: One thing that we feel very good about the Microsoft business model, to answer your question directly, is that mesh makes it a lot easier for people to own multiple devices. Today people struggle with if you have 2-3 pc’s and a cell phone, and music player and other devices, it’s kind of a pain in the neck to keep the data across all the devices, to keep the applications across all the devices. So by virtue of making it more compelling, more easy, for users to have multiple devices that they use in different situations, it’s fairly clear to see how there’s some benefit to accrue to Microsoft because we think our devices are compelling, the software that runs on those devices are compelling, and so people will be more excited about owning multiple devices. That’s one, we think, clear way in which mesh will help Microsoft’s bottom line over time.
Scoble: Steve, this is Robert Scoble. I’m gonna have to take off, our van is moving.
Gillmor: It’s like he’s Matt Lauer.
Arrington: Hey Steve, sorry this is Mike. Sorry I was a little late to the call, I’d appreciate it, now that we’re working together if you could hold the start time until I can get on the call, in the future?
Gillmor: I’m gonna bend over now, so whatever you want to do.
Arrington: Could you give me, and you can do it on Skype if you don’t want to do it on air, give me a 2 minute recap of what we’ve talked about so far cuz it sounds really interesting?
Gillmor: I think it would be a good idea, Mark Cantor, can you give a recap?
Cantor: I’d be honored to give a recap because my job is to be the scribe of the conversation. Mr. Microsoft guy has, so eloquently, given us the company line as to this technology. We haven’t talked about business models, we haven’t talked about what this does to other sync technologies and we also haven’t talked about how this connects into live services. We’ve basically been talking about folder sharing etc. so that’s just the intro, there’s a lot to live mesh and so we’ve really just started, Michael.
Arrington: That wasn’t even close to any kind of summary of anything.
Gillmor: You still want me to say it.
Arrington: Who’s on the call?
Anderson: Robert Scoble, he’s gone. Jason Calacanis is at a board meeting. Saul Hansel can’t make it. I think that’s it.
Farber: Dan Farber’s here.
Gillmor: Right. So Dan Farber has been talking about all of the great writing he’s been doing. That about covers it. We were starting to get into some interesting stuff where David Treadwell has basically talked about his expectations of what he thinks is going to happen in terms of the media. There was a comment by Dana Gardner who mentioned the TechCrunch story focusing on the fact that this was a Windows-only platform. Is that the impression that TechCrunch gets about this?
Arrington: You mean the story that Erick wrote? Haven’t they announced some kind of Mac support haven’t they?
Treadwell: Yes, we will have Mac support. We understand that lots of users have Mac devices.
Arrington: Who are you? Who’s talking?
Gillmor: David Treadwell. He’s basically Ray’s point man on Mesh.
Farber: Here’s the deal. We have David Treadwell, who’s the Corporate VP of Windows Live Services and is responsible for the team that’s building Live Mesh and other services. We’ve basically talked about what is Live Mesh. Basically two things: one– it’s a way to create user scenarios and experiences using this underlying technology that deals with synchronization, collaboration, and other services. And then we talked about the Silverlight connection, that Silverlight and Live Mesh are more orthogonal– they don’t require each other, yet they work together better. And that’s kind of where we ended.
Arrington: Wow, that’s a great summary.
Gillmor: Any more questions Mike?
Arrington: No, I’m just checking out David’s hi-res photo on the Microsoft corporate site.
Gardner: I’ve got a question.
Gillmor: Go ahead.
Gardner: It sounds like this platform has a pretty good value of: create once, present anywhere. That’s particularly valuable on the mobile tier, which is a mess right now. It seems to me that you’re also creating maybe a big switching infrastructure through this platform. I’m wondering, is this just for asynchronous communication or is it possible, using this infrastructure, to move that into more synchronous communication like perhaps voice over IP?
Treadwell: One of the components of Mesh is network transparent communication infrastructure. Basically, it’s a fancy term for saying it allows a device to interoperate and connect to other devices regardless of the network topology the device is on. If you’re behind a proxy server or other network components, it might cause complexity. Mesh has infrastructure that ensures that any 2 devices that have internet connectivity can communicate with one another. Mesh does not have what we could call real-time communication capabilities. Some real-time communication capabilities could work on it. Essentially, if the networks are sufficiently over provisioned, in effect what we see with VOIP today, if they’re fast enough that they could handle moving audio just fine, even though there are not specific real-time guarantees in the infrastructures. As we open up the SDK on that, the network-transparent infrastructure could be used for some real-time scenarios, but it does not add any real guarantees in it.
Gardner: What’s the connection with SharePoint? Do you see bits of SharePoint being part of this Mesh so you can actually build a collaborative application much more easily?
Treadwell: One thing we’ve looked at is ways in which SharePoint can sit on top of Mesh infrastructure. There’s some pretty clear scenarios where a user would be interested in having SharePoint documents, for example, synchronized across multiple devices as those documents are stored in SharePoint. We’re investigating sort of a layering approach where essentially parts of SharePoint would sit on top of mesh. We don’t have any of that available today, but it’s not too hard to see some scenarios where it would be compelling with that.
Gillmor: Somebody from the chat group wants to know what the differences are between Mesh and FolderShare.
Treadwell: FolderShare is a beta-level product that has the ability to move folders just between devices. The biggest difference with Mesh is the platform infrastructure underneath it. One of the specific application experiences that the platform structure enables today is the ability to move files between devices as well. In terms of specific experience, one of the differences with FolderShare is that not only does Mesh allow files to be moved between devices, it also stores a version of the file in the cloud, so that you could access them from any device. So that you have what I like to call high SOA store for those files, so you could get at them even if you don’t have the Mesh infrastructure involved. You could be at a kiosk in India; all you need is a web browser, and you can get your data through that mechanism.
Anderson: That opens up the question again about storage. If you’ve got a terabyte at home, with all these files on it, certainly not all that data is getting up in the cloud for free?
Anderson: Does that just mean that you as the user choose how much of the data is synced to that cloud?
Treadwell: We don’t have the policy or mechanisms in place to allow the users to select what’s in the cloud and what’s not. It’s obviously a high priority feature to give the user that ability to determine what goes in the cloud given that today there’s a 5 GB limit on it. The way we like to think of cloud storage is it’s a special device in the cloud. Just like you can choose to map a folder to any client device, you’ll be able choose whether or not to map a given folder to the cloud storage as well.
Gillmor: Is this service intended to compete against or enhance, which is a good part of the question, Google and Amazon Web Services?
Treadwell: This service is a different level than Amazon Web Services. Amazon Web Services does a great job of providing what we sometimes call utility computing resources– the ability to do mass scale storage, computation, etc. This service doesn’t have anything that operates at that level
Gillmor: Well it sits on top of services that operate at that level.
Treadwell: That’s exactly correct.
Gillmor: So can you wire the two up? Can you sit on top of Amazon services and use Mesh services to synchronize and orchestrate them?
Treadwell: It would certainly be possible for somebody to create a virtual device of a sort that uses Mesh’s open protocols to use S3 as one type of device. We don’t have that support today, but again, by virtue of the open protocol, somebody could create such a device.
Gardner: If someone decides they didn’t want, for some reason, their content to go out through this platform, wouldn’t it be fairly easy for them to block it?
Treadwell: They could simply choose not to map the folder. Is that what you mean?
Gardner: Not the end user necessarily, but perhaps the large organization that’s got content.
Treadwell: I see your point. That’s one of the things we know we have to do a good job with Mesh; that is, to be IT-friendly, and to give IT control on what kinds of data can and can’t be stored on this thing. In addition, one of the things that we expect to do longer-term is to allow IT departments to operate their own cloud storage mechanisms, such that data owned by that enterprise would only be stored on that enterprise’s devices and that enterprise’s servers. They would never go to Microsoft and they would have complete control of that data.
Gillmor: Another question–will Mesh protocol specs be fully disclosed to enable mono-net LAMP integration?
Treadwell: In fact, today, the feedsync protocol, which is probably the single key protocol on which MESH depends, is publicly available. That’s one of the things we’re very committed to, is ensuring that the protocols that Mesh uses are very open, so that somebody could do a Linux-mono implementation of a Mesh client if they so desired.
Gillmor: Is banking everything on feed distribution, has Moore’s Law basically trumped the verbosity of XML?
Treadwell: Gilder’s Law is maybe the more interesting one. The biggest constraint I see in computing infrastructure today is not the number of transistors we have, not the number of bytes or disk capacity we have, but the ability to move all those bytes and instructions around. I think network capability is going to be one of the most interesting challenges the computing industry faces moving forward. You look at the growth of CPU capacity on the desktops you have, it’s still growing at Moore’s law rates. You look at the growth of hard disks and other classes of local storage, they’re growing at absolutely unbelievable rates. But the actual amount of network capacity that’s available to typical users has not grown by the same rate. So I really think that the communications infrastructure and the growth of capacity of communications infrastructure is going to be one the larger constraints that we see.
Gillmor: Well I see it as one of the larger opportunities, because as we saw with Groove, the whole relay services in the cloud, the scalability issue that allows you to virtualize what the device is.
Treadwell: Sorry, I didn’t track that.
Gillmor: We talked earlier about the idea that you essentially normalize devices according to the way the mesh infrastructure sees that device.
Treadwell: Yes, we will. We don’t have that infrastructure today, but that’s on the…
Gillmor: Right, but your relay services, I use the old grid term for what you call something different on your slide, but basically it’s the cloud services that intermediate between synchronous connections and offline clients, etc.
Treadwell: Yeah, in a sense that the amount of computing horsepower available on a local client is growing tremendously. You need to have them as interconnected and as interoperable as possible, but the growth of that recent networking communications resource, because it’s slower than other classes of computing resources, having data close to the users and having a complication on that data close to the users is what we think will build the best experiences for users.
Gillmor: Some people have been criticizing Microsoft for not showing any kind of scalability vis-à-vis what Google’s doing with Gmail. How do you respond to that?
Treadwell: Not showing any kind of scalability, uh…
Gillmor: There have been outages with Live, Mail, etc, etc. There’s been some concern on the part of people, not so much I don’t think you can’t do it, it’s that you haven’t done it yet and therefore people don’t realize or are concerned about your ability to scale up. I mean how many people do you have in this preview edition right now? 10,000?
Treadwell: Yeah, the Mesh preview right now is on the order of 10,000 and that’s just very intentional because we want to do a good job for those users and see how it goes and we’ll expand it over time.
Gillmor: That’s kind of the discussion that’s going on in the Twittersphere about Twitter’s scalability sort of in relationship to that. In other words you start slow and controlled. See what the market will bring and then start to expand it.
Treadwell: We want to have real people use it and I’m sure we’ll hit issues in the service as we scale it up. We’ll try to be intelligent about fixing the issues and continuing to expand the number of users in it as we are confident that we can provide good service to them.
Gillmor: But, in Johnny Dell’s interview of Ray and elsewhere, at the Web 2.0 conference, I don’t remember which session it was, there was a thought put forward that Microsoft has this tremendous experience in moving bits around the network. That it’s not necessarily transparent to the user right now, but you have a lot of heavy duty stuff going on right there and those of us who have been living on Gmail for the last 3 or 4 years, we know how scalable that has become. It’s pretty amazing how it stays up most of the time. So are you going to have to cross some sort of line where people are comfortable about what you’re doing?
Treadwell: Well we have almost half a billion regular users of Windows Live today, and that’s a pretty good scale number we feel. Certainly we’re trying to leverage some the that experience and the number of the components that support that nearly half billion users. I couldn’t quote what the numbers of the people that are on Gmail, but I don’t know if it’s quite at that level.
Canter: So what about synchronizing user accounts and social graph accounts? For instance the recent connection of the Windows Live contact APIs into Facebook. There is a web app you do control your own Live contacts and you are an equity shareholder in Facebook, can we expect that Mesh would enable us to keep our social graphs in synch?
Treadwell: Yeah, that’s one thing that we’re looking at and it’s a very compelling scenario I think. There’s an interesting challenge around that class of data portability, which is that you gotta be really careful about user privacy. You know, trust and security of what you do with users. So we are carefully exploring what’s possible there and we think there’s a very interesting scenario that Mesh could be a very enabling technology for doing things like synchronization, but we’re going to be cautious because we want to make sure that we don’t put user privacy at risk if we do it.
Gillmor: Well can people go in and grab those identity feeds without being part of the service and are those protected from phishing and other kinds of attacks?
Treadwell: The identity feeds? Do you mean their contact info?
Gillmor: Yeah, you’re architecture is broken down into feeds across FeedSynch or identity…you’ve got the slides in front of you. I don’t remember it, but identity was one of them.
Treadwell: Yes. So, once again that information is available via FeedSynch and of course and of course to access it you’d have to have to appropriate credentials to get at that information, but it’s because it’s available by those public protocols, as we solidify the SDKs and whatnot it would be possible to interact with it that way.
Gillmor: Can you explain delegated authentication?
Treadwell: The concept on delegated authentication is if you’re interacting with one application, say Facebook, and you want to let that website access some information that say Windows Live holds on your behalf, the idea is that you delegate authority to Facebook to access Windows Live information that is yours. Does that make sense?
Gardner: In some constrained way, right?
Treadwell: Yeah, there’s a timeout. You delegate that period of authentication for a period of time…minutes, hours.
Gardner: Even that delegation is not to give Facebook the ability to delete your account.
Treadwell: Correct, correct. You delegate specific actions.
Gardner: Right, I’m saying this is kind of obvious, but I just want to make it clear that Steve, when he talked about delegating the authority, it’s a subset of the different kinds of actions and privileges that you would have in the first place.
Gillmor: Yeah, I’m using a term that Microsoft people used in describing what’s going on here, sort of a sub-layer level. The reason I’m asking the question, it seems to me from what we’ve been shown is that the social graph types of capabilities are essentially abstracted by the Mesh architecture. It’s more like you can create a Facebook using these tools.
Treadwell: That’s a pretty good way to put it. Mesh operates at a slightly lower level. Mesh doesn’t have things like a contact store or friend relationships that it stores and supports. It doesn’t have a lot of higher level features that a Facebook or MySpace or another social network might have, but it can be used as a lower level enabling technology for some of those scenarios like dimension contacts incrementalization.
Canter: And you can do that with photos and media as well, right?
Gillmor: So Marc Canter, why are you excited about Mesh?
Canter: Well the desktop server scenario, the ability to do the work offline, I mean that’s been a big conundrum upon the cloud is what happens if your internet goes down? So, I just love this idea of this low thing and this desktop server technology, which has been around for 10-15 years and nobody has really been able to harness the notion successfully and it looks like they’ve really got their handle on that. The other notion of two APIs is intrinsic to the future of the Mesh and when I say Mesh I’m not talking about a Microsoft product, I’m talking about all of us working to interconnect together and to have two APIs is a foundation requirement. We can’t have just sucking. You’ve got to be able to spit as well.
Treadwell: That’s a good way to put it.
Canter: So those are the things I’m very excited about. Now what’s fascinating to me, Mary Jo Foley wrote it up about comparing the Live OS to the Live Mesh. I mean they both got announced at almost the same time and we can pretty much look forward to these two companies, Yahoo! and Microsoft, merge together. And so Yahoo’s finally opening up MyYahoo and mail and search and all of these great features. Now you’ve got to start thinking about the combination of Yahoo! and Microsoft together and it’s a pretty compelling, pervasive solution.
Gillmor: Michael Arrington would you agree?
Arrington: I think you broadly generalized the Yahoo stuff. I think what they are doing is interesting but it seems to me that even if they do all of things they are going to do just directly contradict with what in essence Microsoft is. And I am not slamming Microsoft by saying that. I have been really trying to put some thought on it. Would all the stuff just basically fade into nothing if the deal goes through? And I honestly think it would. If Yahoo is acquired by Microsoft, I think they have a very clear plan for that asset and about what it needs to do. I don’t think it has whole a lot to do with triggering one big social network, and…
Farber: I would disagree with that…respectfully. I think if Microsoft get Yahoo or not do it, then Yahoo absolutely need to do what they said they would do which was to create a single profile for every user. Once you have a single profile for every user, then it is much easier to build APIs that are going to be much faster and easier to build a social graph to work with other social graphs.
Arrington: Is that Dan Farber speaking?
Arrington: That is an interesting conversation and I don’t want to move too far away from Mesh. There is niche communities, Yahoo. You know, when I go to Flickr, I don’t think of Yahoo or when I go to del.icio.us, I don’t think of Yahoo. So they have to be careful in not generesizing everything to the point where those communities lose their own identity.
Farber: That is a user choice. They can support unified profile and you choice which part you want to expose, what data you want to share, or who you want to share with. In theory.
Arrington: Microsoft thinks in terms of big revenue streams. They have office, they have Windows and they want advertising. At the end of the day, all they care about is Yahoo pages and advertisement.
Farber: I think you are right about advertising piece, because you just look at Google’s first quarter and a lot of it is search advertising. There is nothing beyond search advertising which is this notion that if you have 500 million users and if each of them has 500 million users and a …… that has 750 million users or something like that
Arrington: You can’t have more than the hundred percent of the Internet…
Farber: What would you do with that and how do you create an infrastructure. That is where Mesh comes in. It allows you to say that these people are on the web, they are on the desktop and they are on mobile. What if an Infrastructure that support all these scenarios, and what if we can built applications on top of that, that work on whether on Yahoo property, or on Microsoft property or even on Google property, and what if you can unify those profiles and people actually got to manage their profiles.
Farber: I think overtime, people would say we will never move our applications to the cloud, why would ever want to do that. That is just not possible. Look at right now and that is a same kind of an idea.
Arrington: You may want to move your profile to the cloud because that gives you the opportunity to use many different things in many different ways.
Farber: But that is just moving your profile, its having a more unified profile
Arrington: What is the definition of a profile?
Gillmor: David could you define a profile please?
Treadwell: The way we Microsoft view profile is the information that user chooses to share with either some subset of their friends or the world. So, your name, age, maybe your birth date possibly other information like your email address, your IM contact information.
Arrington: So, it is meta-data.
Arrington: Where does that information sit in your vision? What is the core storage place for that data?
Treadwell: Well, profile information today is scattered across lots of different sites for users. One of the challenges that users have is they might invest in putting some information in one side, other information other side. Consolidating is difficult. Part of it is due privacy thing. You put your information on one side, you understand its privacy, and you put other information on another side…
Gillmor: Privacy is only a problem if the user isn’t in charge. Once the user says this is where I want my data to go, then privacy becomes a nonstarter.
Arrington: Or you have situations like happened with Twitter couple of days ago, where third party app started showing very private data.
Arrington: What about the idea of identity, what about the literal meta-data, my name . Who Mike Arrington is …twiter.com/techcrunch and flickr.com/Michael Arrington etc. If I just can give you 8 services that I use all the time then maybe my Facebook profile. That is who I am. It is not just data. It is not as easy as sucking the data from those services. It is more seeing what I am doing there.
Gillmor: it’s sucking the meta-data and the data, that’s what’s interesting about mesh, is that it abstracts those 2 different examples and allows them to illustrate them.
Arrington: meta-data, see all the date, when we’re talking about profile data, everything we’re talking about is either text or an image right but when you go to my Facebook profile, it’s not all about the data that’s on that profile. It’s also about seeing wall comments and how people respond, there is some core essence of “me-ness”, that is intrinsically tied to those services that can’t be pulled out.
Gillmor: I don’t agree. It’s with the assent of the user, the users behavior and the gestures that the user projects you know what he or she looks like on the network Can be aggregated and sent out along with those other data points that you were talking about. And that’s very valuable from a modification (modernization) perspective and as well as from a credibility perspective, you know if we’re looking at all your out puts, your pictures on flicker etc. etc. the ah… ah…,(confusion) oh I see what’s happening here Comcast, Mike can you sent the truck over, Comcast is down here.
[Talking on technical problems about podcast.]
Gillmor: My point is that as you live life on the edge, as you emit signals that people are interested in evaluating the creditability of what you are talking about. If you suddenly start on Twitter saying things that don’t make sense from the users’ perspective, they might start to look at your other channels, and they would come up with a balanced perception of what you are saying. What all I am saying is that Mesh does provide basic food groups to be able to orchestrate that information into something that can then be delivered as a service and…
Arrington: I think there is a subtle point that I haven’t even flushed out because I am just thinking about it on this call for the first time. Loic Le Meur first brought this up when he talked about the decentralized me as opposed to a few weeks ago and how all this data is all over the web and he sort of hates that because the blog is what he considers his core identity. But when I look at friend feed and I’m still, the jury is still out on that, at least my personal jury, it’s all this data and it’s like my twitters and some blog posts, it’s just so out of context it’s just not very interesting and I’m afraid that like in this, in this quest to one give users control over centralizing their data again at least making it appear centralized and still making it interesting from a contextual standpoint, that it’s just not going to quite be right and that someone might get it right and make a lot of money on that. If that makes sense?
Gillmor: Treadwell, you’ve been listening to what Mike Arrington said
Treadwell: Of course.
Gillmor: Ok, do you have any comments?
Treadwell: Well I think one of the biggest things behind all this aggregation is that people really want it in some ways, because you don’t want to have information in lots of different places but there’s this almost implicit segregation of privacy by where you put different pieces of information and there’s 2 risks, as we inevitably do more aggregation and have more data portability on the net—
Gillmor: Wait, you got to pause for a second, you said people don’t want their data in all these different places, did you mean that? You said it almost dismissively like it’s a given.
Treadwell: I think people DO want their data in all these different places.
Gillmor: Oh they do, ok. Go ahead, sorry. I misheard you.
Treadwell: The tricky thing is, sometimes you enter information in a site and there’s an implicit contract that states implicitly where that data will go, and sometimes data moves around in a way that users find creepy. I think Beacon is one example of that, where data moves around in a way that didn’t really fit with how users thought that their data should move around. So the implicit contracts that exist in some of these sites, I think it’s going to be important for the industry to be clearer about how those things work and educate their users about, as they’re entering their information into these places, where that information is going to go eventually.
Gillmor: So I disagree with some of that. I think I do, anyway. I’m less concerned with what you’re talking about, the implicit contract and the beacon program; I think we’re on a path towards people understanding that privacy is truly an illusion. As an aside, I think that that’s one of Facebook’s possibly fatal flaws, is that they’re too concerned with privacy. But that’s another issue. But the core point is that I think people in the long run will be much more concerned not with privacy but with control over data in the sense of literally the ability to manipulate it. For instance, when you delete something on Twitter today everyone still forgets that it’s still backed up on Friendfeed. It’s not just who has access to core data, it’s also the two way thinking. Imagine if I could, ten years from now, click a button, and a single photo that is on 100 different places on the web, make that photo just disappear.
Treadwell: Yea, but by Shia’s impossible DRM solution, we’ll never get to that.
Gillmor: But at least I might be able to make it disappear everywhere that I control. Right? So if I delete it off of Facebook, which would be the first place I upload it, it immediately goes away on Flickr, and Friendfeed as well. And that shouldn’t be that hard, and it should be a priority. It’s the management of that data going forward, particularly when, in a world that you guys talk about, data gets distributed and there’s many copies of it around the web, that a user continues to have control over it.
Treadwell: I think that’s a really interesting scenario because, say, I share a photo with you and I decide down the road that I want to delete that photo. Do I have the right to delete a photo out of your data store? That’s an interesting question.
Gillmor: I don’t think so, I think that the user determines the rights, let say being licensed by them to the cloud, at that point—
Farber: Uh Dan, the reason I buy things at Costco is because if I don’t want it anymore, I take it back and they take it right back. So what I’m saying is, people make easy privacy decisions on the web.
Gillmor: I don’t disagree with that, my point would be that different marketers are going to provide different offers to users and the Costco offer is better than the Fry offer but that doesn’t mean that the user doesn’t make the decision about what the contract is going to be. Now I just want to say, I want to tie this back to the remaining minutes that we have with David, which is to say that in the way that you are virtualizing or orchestrating devices, can’t you do exactly what we’re talking about here by orchestrating these different virtual objects?
Treadwell: Certainly you could have physical and virtual devices, like the AWS we talked about. I lost you when you said virtual objects; I didn’t understand your point on that.
Gillmor: Cool, what I’m talking about is that you’re talking about the release or rights associated with a visual, a graphic, whatever. That’s a contract that you have with the users. Is it without restriction? Or if you distribute it globally, that’s one thing, if you distribute it to a smaller entity group, where
those people have responsibilities to that relationship where they don’t distribute it outside that cloud. That’s a business model.
Gillmor: Ok, so you’ve got the internals, you’ve got the tools here with the matched services to be able to orchestrate this.
Treadwell: Well, let me tie it to specifics: in the photo example we were talking about earlier, say I emailed you guys a photo, I don’t have any ability to delete that photo from your inboxes. On the other hand, if I put the photo in a mesh folder and share that folder with you guys, and then delete that photo from the folder, it would be deleted from your guys’ version of that mesh folder. Of course, you could have previously copied it to some other location to prevent that action on my part, but that gives you a little bit more control over what happens with that information.
Arrington: I got to jump, thanks guys. (1:16:32)
Treadwell: Ok, thanks Mike.
Gillmor: In order to suggest what I’m talking about here, what if you have some information; all these different versions of a Tweet that you make that are going to be stored on various sites. What if you were to, at your endpoint, use mesh services in order to send it out to all these services?
Treadwell: That would certainly work.
Gillmor: Then you’d have control over it. And since you have access to all the API’s individually, you could also use mesh as a proxy to go out and delete them all, if you want to get rid of them.
Treadwell: Sure. So long as the users who received it didn’t take action to put it in some other parallel data store.
Gillmor: But my point would be, to the extent that you could control data, you could do so under a service layers controls.
Treadwell: It does give you a little more control; I just want to make sure I don’t represent it as some sort of absolute control.
Gillmor: Farber, probably you have another question for Treadwell.
Farber: I don’t remember where it was going…but I wanted to ask: maybe you could tell us a little bit of the process of coming up with Live Mesh and tell us how you stumbled upon this kind of solution.
Treadwell: Well I have to give Ray credit for the core of the idea; about 2 and a half maybe 3 years ago as he was joining Microsoft, it was sort of his idea to connect software and services in this way and to have a platform that would make it really easy to do that. Now, tons of people at Microsoft use Outlook. This is one of my favorite examples. So five or six years ago, Outlook didn’t do a very good job of caching information and sharing it in the cloud simultaneously. These days, we all really like the way that Outlook stores your data on your local hard drive. If you think Web 2.0 and your wireless network suck, you can still get it in your email, which is a little window of connectivity and you can synchronize it in your email. That kind of model is super compelling for users—once they experience it, once they see the advantages of having things automatically synchronized and available on all of their devices they really want it. So we thought hard about, hey, how do we make it easy for lots of applications to use a kind of pattern where data is always available on all your devices backed by the cloud and so that’s what caused us to walk down this path of live mesh and the platform mechanism that it has to make it easy to build lots of different experiences above it.
Gardner: So the heritage is with Groove? It certainly smacks of Groove.
Treadwell: Sure, there are certainly some similarities of Groove there, that’s certainly true.
Gillmor: You mentioned unintended consequences at the beginning of the show: don’t you think that there’s going to be a certain move in the direction of things that are more viral right now, that will help you—in other words, Microsoft get more traction in the internet space?
Treadwell: Well certainly virility is very helpful is a very helpful enabler of any new kind of technology. We’ve tried with Mesh to make it really easy to invite people to a live folder for example, if you tried to experience a live folder and you want to share that with somebody, try the invite experience, and that’s a dimension of virility that we’re trying to pursue with it.
Gillmor: Some people in the chat room and elsewhere are complaining that you have run out of invites. Any chance of Gillmor Gang getting invites?
Treadwell: How many are you talking about?
Treadwell: 100? Send me an e-mail and I’ll see what I can do.
Gillmor: I want to thank you very much for showing up and I want to thank Richard Ethel for making this happen. You will see it on your mesh objects a couple of hours from now.
Treadwell: I am super excited. … That was supposed to be a joke.
Gillmor: All right guys, anything else?
Time for lunch
Gillmor: This is Steve Gillmor. I want to thank those who have showed up and especially, those who did not. I am particularly excited about Mesh. I think it is going to be absolutely dominant in the best sense of the word in that we have choice between Google and Microsoft moving forward. Take care. Bye bye.