AWS Official Blog

Transcript of Jeff Barr’s Web 2.0 Interview

by Jeff Barr | on | in Thought Pieces | | Comments

I used the Podcast Transcription Store at castingWords.com to create the following transcript of the interview that I did earlier this year with Josh and Chris at the Web 2.0 Show. I spent about 10 minutes adding some links, and I fixed 2 typos in proper names.

I will post more about the transcription service (which is powered by the Amazon Mechanical Turk) sometime soon.

— Jeff;

 

Here’s the transcript:

Web 2.0 Show – Episode 6 – Jeff Barr
Josh: Welcome to the Web 2.0 with your hosts, Josh Owens,
Chris: And Chris Saylor. We’re a show about the new web.
Josh: The latest thoughts and technology behind internet development and content delivery.
Josh: Today we have Jeff Barr with us from Amazon.com. He is the Web Services Evangelist. Tell us a little bit about yourself, Jeff. Where you’re from, I guess, a little more about your job, that kind of thing.
Jeff: OK, well first of all, thanks for having me. This is really quite a new media and really fun to participate in. So, my name is Jeff Barr, I work at Amazon.com in Seattle. I’m on the web services team and I think I have, perhaps, the coolest job in the world: I’m the Web Services Evangelist and it’s my job to go out and talk to software developers all over the world about what we’re doing with web services.
Chris: That does sound like a great job.
Jeff: I’ve been doing that for about 3 years now.
Chris: What are some of your other projects that you’ve done in the past that led up to your evangelist job?
Jeff: Well, let’s see. I started my software career in actually 1976. I was a teenager and I was working at a little computer store in Seattle back in the days of the Altair and the IMSAI processor technology. And even way back then I was very happy to go out and talk to users at the time, tell them what we were up to with our little business. Since then I’ve worked for all kinds of established and startup companies. I spent three years working at Microsoft; I was there from 1997 to 2000. I’ve also have had this kind of perpetual habit of always having a little project on the side. For example, in the late 90’s I built the first desktop news aggregator product I called Headline Viewer. And then after that I built an RSS feed directory called syndic8.com.
Josh: Oh, wow. That’s awesome. One of the things we noted was that the open-source community often jests about Microsoft being the big bad wolf. What was it like working for them?
Jeff: Well, let’s see. The impression I got is they are certainly very big but they’re really not bad. There’s a lot of really sharp people there. They work on very very complex problems. When you have over a thousand people contributing to a single software product you get just a number of problems that the world really hasn’t faced before in terms of builds, dependencies, goals, and so forth.
Chris: I can imagine. I believe Scoble talks a little bit about, well, he’s Microsoft’s Evangelist, now; I guess he’s your counterpart over there.
Jeff: Exactly. He actually came over to Amazon a month or two ago and did a videotaped interview of me. That was a lot of fun.
Josh: Actually, speaking of Microsoft, we read on your blog that you and Microsoft are working together. Want to talk about some of the contests?
Jeff: Yeah, certainly. So, what we found is that the Amazon services are a great compliment to Microsoft tools because we can use our services to actually show these tools solving some real word problems. So we worked pretty closely with them or we probably have on the order of one or two conference calls and meetings a week with them to talk about various things we can do together, And so we wanted to inspire developers to be really creative and to show off their best work, so we came up with this idea of a developer contest and the concept is it’s going to run until the end of 2005 so folks still have over 2 1/2 months to get started. We’ve got some great prizes including: the top prize is a $5, 000 wish-list fulfillment from your Amazon wish-list, we’ve got copies of Visual Studio and we’ve even got some Xbox 360s in the mix.
Chris: Nice. Excellent.
Jeff: There’s full info on that contest at amazon.com/aws.
Chris: Oh, cool we’ll link it in the show notes for you.
Josh: Yeah, we’ll definitely have to check it out.
Chris: So what is–is the idea just to build a web service using Microsoft tools?
Jeff: Probably you’d be building more of a web service client. You’d be using one or more of the three Amazon web services to build some kind of a client application. It might be an online shopping experience, some kind of a rich visualization of our catalog, maybe a search engine optimization tool using our Alexa web services. There’s lots and lots of different things developers can do with our services.
Chris: What are those three services again?
Jeff: OK, the first service is called ECS, or the E-Commerce Service, and that service provides complete access to the Amazon product catalog. So developers can run queries against the catalog, they can say, “give me products that match a certain keyword, books by a certain author, movies with a certain actor. We return a very very rich very detailed data schema from that service. The service also provide s access to the Amazon shopping cart. The second service is called AWIS, the Alexa Web Information Service, and this service provides access to three hundred terabytes of data that is collected by our Alexa subsidiary. So you actually have a set of APIs to do things like searches against that body of data, you can explore connectivity of the web, you can get rankings and ratings of various pieces of content from the web. And the third one is called the SQS, or the Simple Queue Service. This one’s a bit harder to explain. the idea is that you can use web service calls to create queue data structure hosted on an amazon server. So it’s the classic first-in-first-out queue that you learned about in computer science 101. And then any number of applications could actually push data onto that queue, pull data from the queue, and each user account can have any number of queues associated with it. So it’s really an inter-program communication mechanism.
Chris: Is there any examples where you can see that in action or is that more of a back-end thing?
Jeff: This service is still in beta right now and there are people that are using it. They haven’t really shared with us just yet exactly what they’re building.
Josh: I noticed some great examples, I was looking around on your blog, and you actually do developer-books.com, and this is almost like a mini-amazon. This is amazing.
Jeff: That’s right and that’s a good example of one of the use cases for our service. So, Developer-books was actually built using a script called Associate-O-Matic from Associate-O-Matic.com, written by a very talented developer by the name of Justin Mecham. He built that script for his own use then found it useful enough he decided to offer it up for download, and he also has a professional version for sale. So from download through configuration though having that store up and running was a matter of 30 minutes. I did no design work whatsoever. i have zero design skill. You can look at any of my things I built you’ll see I can’ choose colors, I have no artistic talent. I just picked some colors and the rest has to be attributed to the quality of the script.
Josh: You’re in good company here. We’re definitely not designers.
Chris: That’s why we have Lisa.
Josh: That’s right.
Jeff: But this is a far cry from a just a simple link on your blog to a book you talked about with your billing ID in there, to, I mean, this actually has a shopping cart.
Jeff: Correct. He’s actually using the shopping cart API, so the neat thing about what he’s done is the actual shopping experience stays within his site so he has the user’s attention and focus throughout the shopping experience. Up until the point where the user says, “I’m ready to make my purchase” and only then do they make the transition from his site or my site over to Amazon. So this is a perfect example of a developer taking our service, having a great idea, building something, putting it online, and being able to profit from it.
Josh: Are there any other services like that, current or past that you want to talk about?
Jeff: There’s a number of different things folks have built, there’s another great store builder, there’s a developer by the name of Mr. Rat, believe it or not, at mrrat.com. He has a Perl script that’s been used to create several thousand associate sites, and he actually gives that script away for free as sort of a give-back to the open source community. He’s told me he really appreciates what he’s been able to get from the open source community over his professional career. And he said”I love this script, I’d love to be able to give it away for free to give back to the world.” We’ve got a lot of people who’ve built different kinds of graphical browsers against the product catalog. There’s a site called liveplasma, liveplasma.com. Uses a Flash interface to do a very very radical interpretation of what’s in the Amazon catalog.
Chris: Hm. We’ll have to check that out. Actually, I wanted to ask, how did you get a gig like Web Services Evangelist?
Jeff: OK, good question. So, I joined Amazon in the middle of 2002 with the idea that I would mostly be working on the associates team, but that I’d have a little bit of time at the edges to devote to web services. And then very quickly that little bit of time grew into almost a full time job, and then my boss encouraged me to interview for the evangelist position. I wasn’t comfortable taking the evangelist position from the outside, mostly because it seemed to come into a company and all of the sudden be thinking that you could represent the company. i felt a lot more comfortable learning about the company first, learning about the products, and only then feeling like I could legitimately go out and talk about it. So I spent some time before Amazon doing web service development work, some web service consulting, some web service work at Microsoft. I think as an evangelist a lot of what it takes to succeed is just having a really good attitude, not being afraid to go out in public and face complicated questions or smart audiences and have folks throw all kinds of interesting things at you, and just think, “well, I probably know at least as much as least as much as they do, so I’ll do my best to respond. and so I do a lot of travel, I love to talk at conferences and user groups and classrooms and I go out of my way to see if I can meet with developers wherever I go.
Chris: Cool. Kind of a combination of outgoing and a geek at the same time.
Jeff: Absolutely, yeah.
Josh: How many developers do you guys think you have doing work with the Amazon web services?
Jeff: We’ve signed up over 120,000 developers. They’ve signed up for our program, so they’ve given us their name, their email address, their contact information. In exchange we give them a little identifier called a subscription ID that they use to identify themselves when they make calls to us. We don’t separately track how many of those 120,000 are actually doing things, but it’s definitely a good proportion of them. once they’ve got that subscription Id they can start building and creating right away.
Josh: What about internal developers?
Jeff: In what sense?
Josh: Like, how large is the Amazon team? I guess it’s probably not a separate team, is it, that’s actually building the web services?
Jeff: We don’t really give out a lot of internal numbers inside the company, but the interesting thing is that the way I can describe it is that because of the architecture of our web services, our outside services are really resting on the same services that the rest of the company uses. So when you do a search through an ECS call, you’re using the same search technology that the web site uses. So in a sense everybody on the IT team at Amazon.com might in some way contribute to web services whether they quite realize it or not. They’re building things for a technology platform that’s then used in a number of different ways.
Chris: Wow, that’s really cool.
Josh: It’s definitely a good re-use of resources I guess. In our first Podcast we talked to Matt Mullenweg a little bit about managing an open source project, I guess in some ways your web services project is similar. How exactly do you manage that?
Jeff: OK, So, we don’t really have internal methodologies that we talk a lot about to the outside. What I do think we’ve done is we’ve picked up a lot of the good principles from open source development. To me some of those things are: release early and release often, we do incremental releases of our service very very frequently, and I think maybe the most important one is creating a community and listening to the community. When you’ve got all these users out there, you’ve got all these developers solving real problems, the best thing you can do is open yourself up and listen to them. And when they say, “I’d really love feature X. do Some more”, you need to take that. If they say, “Feature X could be good, but you need to get some things fixed”, you really really need to listen to that, and if they say, “Feature X really sucks badly”, you need to just take that input, and then understand what the problem is there. And do your best to respond very quickly. So I think when you look at open source and its relatively short cycle times, the fact that the community is very involved in development, I think we’ve successfully taken those attributes and incorporated them into what we do.
Josh: That seems to be one of the common themes among web 2.0 developers is especially people like 37 signals. Are you familiar with their products?
Jeff: I am.
Josh: You know, we’ll release a product and then it’s kind of like a starting point and it’ll gather feedback from the users and it’ll grow from there.
Jeff: That’s right. Because the developers at a certain point are closer to the problems that they need to solve than we are. And so they are in the best position ultimately to tell us: is it meeting their needs, or not?
Josh: Amazon was one of the few companies, I guess, of what we’ll call the web 1.0 world that actually survived past the bubble. It really even…
Jeff: And so they’re in the best position ultimately to tell us is it meeting their needs or not.
Chris: Amazon was one of the, you know, one of the few companies, I guess there’d be what we’ll call the Web 1.0 world, that actually survived past the bubble, and really even back then you guys had some features that were, you know, we’re now calling Web 2.0, like you guys have had the API, how long have you guys had the API anyway?
Jeff: The API has actually been out since the middle of 2002.
Chris: Oh, cool. Yeah, I mean you, even as a large company, a lot, you know, a lot of people associate Web 2.0 with all the small companies, but you guys have definitely have a lot of features of Web 2.0 kind of company.
Jeff: The Web 2.0 is a lot about listening to customers, it’s about being open, it’s about inviting folks in to help and to participate. And so the model I love to talk about is with web services, what we’ve effectively done is we kind of grabbed, found this kind of loose edge on the web site, and we pulled it off, and we let developers see what’s inside the web site. And you’re giving them essentially direct access to a number of our platform technologies.
Chris: Cool.
Josh: Earlier, you mentioned syndic8 and how, I guess, you created that. Can you go into a little more detail about what syndic8 is and how that came about, how you came up with that idea?
Jeff: Yeah, certainly. So, way back in the late ’90s I had this desktop news application called Headline Viewer, and at one point I think I actually had the list of every known RSS feed in the world, which was somewhere around 800, 900 feeds at that point. And a lot of people liked the application, but even more people liked the fact that I had this exhaustive list of all the feeds. And so a lot of people started asking me, hey, could I just have your feed list? And they said, we can help you, let’s kind of get this community together so we can find feeds, collect them, organize them, work on improving the quality. And I got the domain name syndic8.com because it sounded really cool, and in August of 2001 I started writing PHP, it was really the first big thing I built with PHP. And right now there’s about 80, 000 lines of PHP and there’s about 200 gigabytes worth of MySQL data behind the scenes at syndic8. So it’s been a really fun project.
Chris: Do you still do any work on it?
Jeff: I absolutely do. My general development model is I spend usually five a.m. to six a.m. and ten-thirty to eleven-thirty at night working on it every day that I possibly can.
Josh: What kind of infrastructure does that take to run something of that size?
Jeff: So right now there’s a dual-processor server with about 3 gigs of RAM and about half a terabyte of disk on there, hosted at a site called bocacom.net.
Josh: That’s pretty cool. We haven’t put this show up yet but we actually mention in the podcast for Kevin Rose that we also do hosting, so if people are looking for hosting, check us out.
Jeff: So hosting, it’s one of those great things that makes it so straightforward to put a company online, now. It used to be, in the early start-up days, as soon as you got funded as a startup, the first thing you’d do is you’d call your Sun Microsystems salesperson, and he’d come over, he’d write you an order for a million dollars worth of servers, and you’d wait a long time, and only then could you really get your startup really started. And now you go to any one of a number of really great places. You visit a web site, you put the right data in there, you put your credit card in, and an hour later a root password shows up in email and you’re online and ready to go.
Chris: It actually is, absolutely is amazing, you know, places like EV1 and rackspace and serverbeach, you know, a hundred fifty bucks a month and two developers, and an idea, and, you know, you’ve got a web service.
Jeff: Exactly. I really think that you can go from an idea, to figure out a domain name, to get it registered, to building something and putting it online, I think you could probably do that in twenty-four hours. Pretty much as close to instant as you could do.
Chris: I know, a lot of things being built now are, you know, they’re like from idea to conception is like less than a couple of months.
Jeff: That’s right. So that means when you see those needs out there in the community, if you can get your brain around this is what the need is and this is how I can solve it, you got a great shot at filling that gap before it evaporates and while there’s still time to make some money from it.
Chris: Absolutely. I know, well, Josh and I had started work on a calendar application, and having full-time jobs, kind of hard to dedicate much time to it, but, you know, in that amount of time, a lot of them actually popped up, but none of them really seemed to see the vision of, you know, what we have, and we hope to be able to finish our little project, too.
Jeff: You got to go for it. So what’s always worked for me is to never think, okay, if I only could do something full-time, then I could really make it work. I just figure out what are these incremental pieces of progress that I can keep on making. And even if I can only devote ten minutes to Syndicate in the morning, I make ten minutes of progress that day. And I’ve figured out a good development methodology that lets me actually get some value out of that ten minutes, and push things forward. If it’s only a tiny bit every day, it still adds up to something over time, is what I’ve learned.
Chris: Yeah. I think there are a lot of people out there, especially, a lot of our listeners I know, probably, obviously, developing their own web services, and that’s really what you have to do. Because there’s almost, there’s a lot of talk about, oh, will, should I take VC, you know, you don’t really need to take VC now, the 37 signals are out, and just do it yourself.
Jeff: Works for me. Turn off the TV, and wake up an hour earlier and make it happen.
Chris: That’s awesome.
Josh: We saw a post on your blog about textfo, I guess txtfo, can you tell us a bit about that?
Jeff: Yeah, certainly. So txtfo is a service based in the U.K., and the concept is, it allows mobile phone users to interact with online ads using SMS text messaging. So the concept is that advertisers will run ads, and I think they’re running some ads right now on the various subways in the U.K., they run these ads, and there’s a special txtfo number attached to the ads. So, you text your, you text to the actual number, and you include your email address, and they actually send you back product information. So there might be a movie or a book or a DVD advertised, you text over to it and it will then reply and say, here’s more information, give you the information needed to actually initiate your purchase. So a fairly innovative kind of application and a unique business model.
Chris: Cool. Can you actually get, like, the comments too, as well?
Jeff: I believe it does return the reviews. It’s built on SMS, so with SMS you get a somewhat limited amount of data you can pass through at a time. On the order of a couple of hundred characters.
Chris: That’s cool, because I actually, I don’t even buy a book without going to Amazon first, and reading the reviews and see whether or not it’s actually worth it.
Jeff: That’s a good practice, and to me the actual cost of a book isn’t the dollars, it’s the amount of time you’re going to invest in actually reading it. And so you want to know, am I going to get a good value out of the time I put into it.
Chris: Actually, sometimes what I’ll do is, I’ll be at the store, like, you know, actually be at a brick-and-mortar Barnes & Noble, sorry, and I’ll call Josh on the phone and be, like, hey, look this book up for me on Amazon. So that would great if I could, you know, if that service came to America. You said it was U.K. only, right?
Jeff: That’s right. You know, there are some services that will let you do that from, you could do it very discretely from within a physical store. I don’t have some URLs in my head offhand, but I can give you the, I definitely have seen some of those.
Chris: Oh yeah, ship those over to us later and we’ll put them in the show notes.
Jeff: Will do.
Josh: Yeah, that’s awesome. Yeah, I’d love if we had a service like that here so I could stop being Chris’s bitch. Earlier we had mentioned the Web 2.0 conference, but we weren’t recording. Give us your thoughts on that and what excited you there.
Jeff: So the industry’s definitely back in full swing. And what I’ve noticed among my friends is that, it used to be I had a lot of friends that were not fully employed, they were just out there looking for opportunities, and now everybody’s got one job or two jobs, and the opportunity’s just out there looking for people to fulfill them. So, the boom seems to be back. There was a lot of, a ton of positive energy at Web 2.0. Just a lot of cool things. Everything seemed very dynamic. It seemed very automatic. It seems people are now, out of the gate, making sites that look very, very sharp and very professional and artistically designed, unfortunately for folks like me who don’t have any design talent. I think web services is great, because web services is what’s empowering a lot of these things, in that, if you want to build an e-commerce site, you don’t need to build multi-billion dollars worth of distribution facilities like Amazon has. You simply use our web services and you do the fun part. You build your application or your web site, or your mobile application, and you let Amazon do all the dirty work of actually shipping things to people.
Chris: That is cool. One of the discussions going around right now is, there seems to be an awful lot of money, VC money, going into these web services, and people are afraid that there’s another bubble. Did you see any of that?
Jeff: Hard for me to judge. I was really there for less than a full day, but the, there certainly were a lot of well-dressed folks in the crowd, that seemed to be out there. Probably they had a hundred-dollar, packets of hundred-dollar bills in their jackets, I don’t know, but they seemed awfully, awfully eager to talk to everybody and say, hey, what are you doing, what are you up to, what’s your business model. But that business model question is very important, and I think that that perhaps wasn’t asked often enough in Web 1.0. Now it seems pretty fundamental that not only have the people learned to ask that question, but the folks have good answers for the question as well.
Chris: Yeah, I really hope to see people keep their heads on straight this time around, because there’s a lot of cool stuff coming out. It’s almost like everything has to work together for everybody to win, because you don’t want, you know, the things to slide and, I don’t know, it’s just going to be bad. But I’m really excited, there’s a lot of cool stuff coming out. And I, for one, am going to take a look at Amazon web services. I actually have an affiliate ID, I just, I’ve never had a web site with enough traffic to justify linking it in. I know my wife’s blog, she, like, reviews a book every once in a while, but I didn’t realize you could do almost a full-blown store. That’s really cool.
Jeff: Most definitely. And there are people who also will build stores, and then they’ll invest their money in search-engine optimization and traffic building, to drive folks to that store. And they can do pretty well through that.
Chris: Excellent.
Josh: Alright, well, is there anything else you wanted to discuss, Jeff?
Jeff: I think we covered just about everything. We haven’t really talked a lot about the future because as a public company I really can’t pre-announce anything, but we do have some neat things in the works, and I’d encourage people to go to amazon.com/aws from time to time, and check out what’s new there from time to time.
Josh: Oh, come on, you know you want to reveal that super-secret project.
Chris: We haven’t gotten anybody to do it yet.
Josh: Well, I guess Matt did, he talked a little bit about wordpress.com before they’d actually released it.
Chris: That’s true. And I think Kevin said he was going to give us the scoop on something, so…
Josh: It’s true. Alright, well, thanks for being on the show, Jeff.
Jeff: Alright, thanks for having me, this has been fun.
Josh: Yeah.
Josh: This has been a Steel Pixel production. For more information about Steel Pixel, you can check out steelpixel.com, or for more information about the show, feel free to check out web20show.com.

[song]