Alan Lewis

Mozilla Weave -- Towards a User-Centric Internet

I recently got an invite to the Mozilla Weave alpha, and after playing around with it for a while I can say that this product has a chance to be a very big disruptive force on the Internet.

The features provided by Weave are modest right now. It offers bookmark, history, and tab syncing, and of these features only bookmark syncing seemed to be working correctly. Other features set to be released in future versions, such as cookie, password, and, most exciting of all, extension syncing, have the potential to make Weave much more useful.

So why am I so excited about Weave, especially considering that production-quality extensions like Foxmarks already offer the same features as the current version of Weave? I'm excited because Weave has the potential to shift the balance of power away from the large Internet services and back to the user. Like many users, I gave up using browser bookmarks many years ago because I was using more than one computer on a regular basis. I started using del.icio.us, and then switched to Google Bookmarks. While I gained some useful features by using these services, I was at the same time handing my data over to a third party, relying on their good will and the strength of their security systems to keep my data private. Weave is different -- because user data is encrypted, when the data is uploaded to a Weave server, the data can still only be accessed by the user who first encrypted it. So even if Mozilla wanted to access the data (or if there were a security breach and someone stole it), user data remains safe. More importantly, even if the data were to be subpoenaed, it would remain unreadable by a third party. Of course this puts the onus of security on the client, but since Weave is open source, you don't simply have to trust the creator's word that the client-side encryption is secure.

You don't just have to be a security spook to be excited about Weave, though. I'm very interested in the possibilities for third-party applications that could be built using Weave. The hard part of building a web application is scaling it. What if you could build a web app that used Weave to store and sync all the user data instead of building out your own backend servers to do the same thing? Right now the best alternative out there to building your own backend infrastructure is to use services like Amazon S3. But again, this centralizes too much data with very few service providers. Building apps that leverage Weave would keep user data close to the user, where it belongs.

There are significant challenges that face Weave if it is to be a success. The project creators have spent much of the last year building a scalable infrastructure, and the alpha users are eagerly awaiting completion of the basic feature set that was promised when the project was first announced. The emphasis on security necessarily impacts the user experience. Not only will users have to get used to "signing in" to their browser, but they will also have to understand what is the difference between a "password" (authenticate the user with Mozilla) and a "pass phrase" (encrypt/decrypt user data). With enough testing and design iterations, Weave could be made acceptable to mainstream users, but it will take time and effort to do so.

February 24, 2009 in Software, Tech/Web | Permalink | Comments (1) | TrackBack (0)

Adobe - Go All In on Open Source!

Earlier this week I was on a panel at the Adobe Sales Conference, an internal event attended by 1,000+ Adobe sales and marketing people. The panel was to discuss "marketing to developers", and the discussion was, I heard after its conclusion, helpful to those who attended.

During the panel I was asked by the moderator, "what advice do you have for Adobe?" My reply was that Adobe should stop sitting on the fence regarding open source. I followed up my answer with an acknowledgement that I didn't think that those in the audience were necessarily the ones who had the ability to make that decision, but that they could perhaps influence those within Adobe who do have that responsibility.

I was glad to see the announcement that Adobe is open sourcing part of LifeCycle as BlazeDS and even more importantly is publishing the specification for the Flash binary data format, AMF. These are important steps, and along with the recent moves to open source Flex and standardize the PDF format, they show an increased acceptance of open source within Adobe.

But it's not enough. To gain the trust of the open source community, Adobe needs to open source the Flash Player and publish specifications for the SWF format. Only then will it have a chance of winning over this key constituency as the battle against Microsoft heats up. Sun waited much too long to open source Java (with a real open source license, that is). If it had done so in 2001 instead of 2006 it may blunted the rise of PHP as the web environment of choice.

So, Adobe, buddy, good work. You're almost there.

Note: These opinions are mine and not necessarily those of my employer.

December 13, 2007 in Software, Tech/Web | Permalink | Comments (0) | TrackBack (0)

OpenSocial - It's Just a Widget Container

I had a big, grand post planned out - I was going to dig into both the strategy and technical issues behind OpenSocial. Unlike those who proclaimed "Checkmate" based off of nothing but the PR hype, I was going to astound the blogging world with my insights and biting criticism. I even had a clever Saturday Night Live inspired title picked out: "OpenSocial is Neither Open Nor Social: Discuss"

Then I found that Dare Obasanjo and Tim O'Reilly had already done it for me.

Basically, there are two problems that I'll summarize here:

  • OpenSocial doesn't give most developers what they want, which is the ability to mash up data from multiple social networks in order to create something greater than the sum of its parts.
  • The APIs exposed are woefully inadequate. So you get access to a friend list, and the friend object contains a name, updated date, link to profile, and link to picture. Umm... Thanks? Anything beyond that needs to be specified as a custom extension to the schema. This is going to make it really hard on developers to create applications that work in more than one OpenSocial container. Translation: a MySpace app isn't going to work out of the box in LinkedIn, unless it only loosely integrated into the social network.

With these limitations, most OpenSocial apps are destined to be nothing more than glorified widgets - widgets that have the ability to spam your friends. Is that what we all want?

Note: these opinions are my own, and not necessarily those of my employer.

November 08, 2007 in Tech/Web | Permalink | Comments (1) | TrackBack (0)

eBay Desktop Public Beta Now Available

The project formerly known as "San Dimas" is now available in public beta: eBay Desktop launches in public beta today. Anyone can download eBay Desktop for free.

eBay Desktop is, well, eBay on your desktop. You can use eBay Desktop instead of the website to search, browse, bid, watch items, read and write messages -- most everything that you use the eBay website for to buy stuff. eBay Desktop also goes beyond the web experience through features like desktop alerts (for when items are ending or when you get outbid) and favorite search feeds, which let you spot when items are listed that match one of your searches.

Ebaydesktophomepage

We've made a number of improvements since the private beta (besides changing the name, of course). The number one piece of feedback from users was "make it more about me." We took this to heart and have completely revamped the home screen to put more information about your eBay activity up front. We've made significant performance and memory-usage improvements, and more will be coming soon. The alerts feature now works even when the application window is closed -- eBay Desktop will optionally keep running in the background so that you can get notified when you are outbid.

Remember when it used to be possible to browse on eBay? Well we wanted to leverage the power of the desktop to make browsing on eBay workable again. On the home screen you can drill-down into the categories, all without refreshing.

My favorite new feature is search feeds. eBay has offered favorite searches for a long time, but unless you enjoy getting emails containing your favorite searches, they have been of limited use. Now in eBay Desktop you can access your existing favorite searches or create new ones, and view those lists of items as "feeds" just like you would in an RSS reader. In this view you can see if an item is new, and you can delete items that you aren't interested in.

Download eBay Desktop, and let me know what you think. Feel free to contact me directly (see link in the sidebar) and the application has a built in feature request and bug report widget.

October 01, 2007 in AIR, Apollo, Desktop, eBay, eBay Desktop, San Dimas, Software, Tech/Web | Permalink | Comments (5) | TrackBack (0)

Want Open Source Flash Player? Contribute to Open Source Flex!

It is awesome news that Adobe is Open Sourcing Flex. To me, this is a significant announcement, and I am encouraged that Adobe is moving in the right direction. I hope that they choose to eventually open source all of their core technology and standardize all significant data formats. With the standardization of PDF and the open sourcing of the Actionscript 3 virtual machine (Tamarin) not that long ago, they are much of the way there.

As others have noted, this is not a half-assed open sourcing effort as other companies have tried. That they are using a well-understood open source license (Mozilla) is the most important thing to note. When we (eBay) open sourced our SDKs last year we made sure to only well-understood licenses (CDDL and MIT) so that other companies or developers didn't have to figure out the ins and outs of a brand new license. This was a bit of a pain to educate our legal department on why this was important, although they eventually got it, and now are totally on board.

So what is left to open source? The big thing is clearly the Flash Player. I would guess that their intention is to eventually open source Flash Player. But I'd also guess that they main hang-up may be a litany of patent and intellectual property issues (especially with licensed code). Flex itself was (as I understand it) developed completely by Adobe/Macromedia, and does not incorporate any third party code (that I'm aware of) so it was "easier" to open source. However, I'd also guess (third guess this paragraph) that the open sourcing of Flex is a bit of a trial run for the eventual open sourcing of Flash Player. What issues will come up? How will the community respond? How will this move financially impact us? Will it be a net positive, or a net negative? These are all questions that I can imagine are on the minds of the folks at Adobe.

So, if you want Adobe to open source Flash, contribute to and support the open source Flex project!

Disclaimer: I am an eBay employee and these opinions are mine alone. I should note very clearly that although I am under multiple layers of NDA with Adobe, I have never discussed open sourcing their technology with them, not do I have any inside information on this topic -- I'll tell them the same thing in private that I am on this blog! 

April 26, 2007 in Apollo, Programming, Tech/Web | Permalink | Comments (0) | TrackBack (0)

Slides from my Desktop 2.0 talk at Web 2.0

Here is the presentation from my talk at Web 2.0: "The Future of the Desktop". I hope to blog about the material I pulled together for the talk, but it can be summed up as: web services enable rich desktop applications, users like desktop apps, and there is life beyond the browser. The Q & A session was really great, and unfortunately the awesome questions aren't captured in the slides. Thanks to everyone who attended!

I don't know what to think of the whole conference. Moscone sucks - way too big to have a community feel. The show floor was interesting, and some of the vendors like Effective UI and Bungee Labs seemed to really get good value from it, which is great for them. Kevin Lynch's demo was awesome - I love the rapid fire demo style that he has, and it was awesome to see San Dimas demoed in front of such a large audience.

I won't say who it was, but lots of people were peeved at one of the keynote presentations from the CEO of a large Internet company who basically gave a 15 minute vendor pitch. Of course many of the sessions were nothing but vendor pitches. I tried to keep that to a minimum in my talk, and I hope that folks got some useful information out of it that wasn't eBay-specific. There was great buzz about Rolf's "Web 2.0++" talk (my co-worker) (Birdz, OMG!). 500 slides in 45 minutes, and none of it was a vendor pitch. You have to see this talk... and you can at the eBay Developers Conference in Boston - June 11th - 13th! Hey, I know its a pitch, but everyone is doing it...

April 18, 2007 in Desktop, Tech/Web | Permalink | Comments (0) | TrackBack (0)

360Flex Conference at eBay - Day 1

I'm at the 360Flex conference at eBay's North Campus. Its a packed house in the presentation, and the conference so far has been a great success. There seems to be a good balance of people here -- some are experts, some are beginners. As I expected, many of the people I've talked to thusfar are independent consultants. They recognize the economic opportunity that is out there for developing rich internet applications.

On Monday night we had a recruiting event and I talked non-stop for 2 hours with folks about San Dimas. Everyone was very excited about the application, and I think I convinced a few people to come work for eBay based on the strength of the demo. I was too busy to take pictures for that part, but I did get some from earlier in the day:

Christian Cantrell and Ted Patrick sign in.

The keynote audience

Mike Downey talking about Apollo

Tom Ortega talking about Cairngorm

March 06, 2007 in eBay, Tech/Web | Permalink | Comments (0) | TrackBack (0)

Chris Anderson Talk at eBay on The Long Tail

Chris Anderson, author of "The Long Tail" spoke at the eBay South Campus yesterday as part of the eBay Speaker Series. I've gotten a bit of a repuation for writing good speaker notes for these events (I think its because I can type fast) so I tried my best again. I got a signed copy of his book, and started reading it this morning. It was cool to see an acknowlegment in the book of Terapeak (an eBay Certified Provider), who helped out Chris with his data analysis. These are notes, and while I tried for accuracy please don't take these as direct quotes from Chris.

 

"The Long Tail – Who needs blockbusters?"

You guys inspired the book, in many ways. You have shown me the way. This was the result of a 2 year project involving Stanford, Harvard, and MIT.

Let me talk a little bit about the 20th century. In the 19th century, we were fragmented by geography. Then came the cultures of mass media, especially the radio in the 30s. The power of broadcast enabled one-to-many communication. It became incredibaly efficient to take one thing and distribute it to everyone. This caused us to become a blockbuster culture – this was what the physics allowed. This peaked on March 21st, 2000. NSync’s No Strings Attached, which was the fastest selling record ever.

Hit records have fallen dramatically since 2000. The music industry is flat, but the hits have fallen dramatically.

In the 1950s, we were synchronized culturally. We are fragmented now. Our lens has been on the “head.”

Music – Walmart is America’s largest CD retailer, but they have 25,000 items for sale at most in the store. On Rhapsody, there are 900,000. Music – 40% of Rhapsody’s business is with tracks that are not available in Walmart. This is the market that we thought was effectively 0% -- it is heading to 50% over time.

Chart comparing Blockbuster’s sales rank vs. Netflix. Example of trying to find “Akira” in a Blockbuster, and how finding is important too.

“Power laws” – they are incredibly ubiquitous. They are in nature and ecomonics both: wars, earthquakes, biology, etc… A proper power law should be a straight line, but with movies, its not, because there is a limited supply of screens. There is a bottleneck in distribution in the American market.

Example of Amazon sales data 2001 vs. now. They are seeing more sales in the long tail. If a listing doesn’t have all the metadata, reviews, etc… it will not get sales. This is an information problem.

Example of Netflix – their tail drops off too, but Chris thinks this is due to a switch to foreign languages.

Normal distribution under scarcity (non-zero marginal costs) = bell curve. If things cost money, businesses focus on the middle. Normal distribution under abundance = curve going up to the right. Blogs – large number of low quality blogs, small number of high quality blogs. Network effects (word of mouth) identify the good things and diminish the bad things. Google Pagerank is a perfect example of this. (Graph of frequency vs. rank)

eBay – Great long tail example. We did a study on eBay, and it kind of blew up on us. Typically what we do is rank sales of products. But eBay doesn’t have the concept of the SKU (in most categories), because everybody lists products differently. So – we weren’t able to come up with a proper long tail analysis.

 

Q&A

Q: Do you see the culture going back to the “Heavy Head”?

A: I don’t. Let me answer that 2 ways. I see more and cheaper distribution. Digital technology will continue to lower the costs of distribution. The shift toward niches is natural. However, the internet is a fantastic amplifier, but what is interesting is that you see those amplifiers being applied to things that come up from the bottom, instead of just those things that come from the top down. You will still have hits, but the story is different.

Q: The Power Law – if you plot it on a hyperbolic scale, does it plot better, or is there still a drop-off?

A: First time I’ve got that question (laughs). No matter how you plot it, you’d see a drop-off. It would look different, but there would still be a drop-off.

Q: Have you seen a difference in the profitability in the tail?

A: It’s actually more profitable in the tail. The acquisition cost of a DVD when it is released is 21 dollars. Blockbuster has to charge the price set by Walmart: 16 bucks. Netflix delays getting copies until the acquisition costs go down (a few weeks later). Also, the longer a movie has been out, the more satisfaction goes up. Blockbuster is in a market where the acquisition costs are higher, and the satisfaction is lower.

For Amazon, they lower the discount as you go down the tail.

Q: For products where it costs a lot to produce them, does it apply?

A: Cars are an example where the long tail doesn’t apply. The most important thing in a car is quality. Any variation on the manufacturing side will lead to lower quality, so customization is not common. The dealers hate customization because it cuts them out.

Q: Because we see the power law in nature, do you think that humans have brought it on ourselves to find quality things?

A: That’s a great question. What you are talking about is some sort of evolutionary advantage – and I think it’s true. What’s interesting is why do you see it happen in earthquakes?

Q: Does the exponent change as you go down the tail?

A: 2 important things: they are fractal. Music follows a power law, but so do the top sellers in jazz. You always have hits and non-hits. What you are talking about is the predictive value. You should be able to predict the demand by adding items, but it turns out that you don’t have that. Adding things to the tail change the dynamics, but we don’t know how the exponents change.

Q: What do you think the impact on the creation of content?

A: Forces that create the long tail: democratization – this populates the tail, distribution – anyone can put something out there. This probably will have an impact on the blockbusters. It’s not dramatic yet, but I expect that model is reaching the end of its life.

Q: Is there a long tail for revenues/profits for media items?

A: Media that doesn’t pass word of mouth muster falls off much quicker – these are the “bad hits.” But those media that are organic hits, the good hits, have more opportunities for audience.

Q: How much of the YouTube phenomenon is a result of culture change/the “ADD effect”?

A: That’s a great question. On news sites, half the traffic goes not to the front page, but to individual stories. The album has been micro-chunked. What you are seeing is not necessarily ADD, but one size doesn’t fit all. We thought that we wanted TV in 30 minute increments, but in fact it was the advertisers. Online, the peak time is less than 3 minutes. Sometimes you want a snack, not a meal. This is the natural shape of demand that had been suppressed by the distribution.

Q: Will the tools themselves (The filters) succumb to the long tail?

A: Absolutely. Take music filters – you might start with taxonomy. The filters at the head are not necessarily the same at the tail.

Q: It seems like the long tail effect is stronger in things that we take a personal interest in, like books and media.

A: I took things I thought wouldn’t have a long tail effect, like flowers. Some people care passionately about them. (mentions “The Omnivore’s Dilemma”) Most of us are happy with Walmart and Safeway, but there was demand for the kinds of food that wasn’t sold there. As you know, Walmart beats up suppliers – there is a shortage of organic milk, and Walmart has to beg for it. Longtail Libations – long tail of beer is micro-brews. The supply chain wanted 4 beers, but now the supply chain is more efficient.

Q: One of the effects of everyone getting what they want is a lost culture. Do you think that will continue?

A: That is a great question to end on. What is the loss to society? My argument is that there is certainly a loss, but the upside outweighs the downside. Us all watching I Love Lucy is a superficial connection. In your niche interests you can find people who share your interests passionately. The rise of tribal culture -- you think of tribes being in conflict, but think of being part of many tribes simultaneously, and expose you to a rich variety of culture, even if you lose something in the mass culture.

July 20, 2006 in eBay, Tech/Web | Permalink | Comments (1) | TrackBack (2)

I Love Digg

Digg. Digg. Digg. I love Digg. Why do I love it so?

The site redesign is really, really nice. Ajax done well. But that's not why I love Digg.

I scored an invite the Digg V3 party (actually, Delyn scored the invite -- I just tagged along), and had a great time. Met Jay and Kevin and lots of other Digg folks. But that's not why I love Digg.

They are opening up an API (horray!) very soon. But that's not why I love Digg.

Even though my initial experience was bad, it wasn't Digg's fault. Its the stupid spammers fault who make it so that Digg has to be extra careful with suspicious behavior.

eBay and Digg are corporate cousins (same, umm... grandfather?). But that's not why I love Digg.

No, those are all reasons why I like Digg. The reason I love Digg is because it has personality. What put me in the "love" column was seeing a page like this when Digg was down on Sunday (I forgot to capture the original, but I found a cached copy that looked a bit older than the one I saw):

Diggdown

That, to me, shows real personality. How could you not like a link to Ask a Ninja? And three of the other links on that page were really useful, Top Left Pixel, A List Apart, and digg the blog were new sites for me. Just a simple page with what looks like everyone at Digg's favorite site. Anyone can build a website. Many can build a useful website. But its really hard to express genuine personality, and thats what won me over.

P.S. I added Digg links to the bottom of each post (via FeedFlare) to celebrate my new love. Feel free to give them a try.

June 26, 2006 in Likes, Tech/Web | Permalink | Comments (2) | TrackBack (0)

Salesforce AppForce Conference Wrapup

I was at the Salesforce AppForce Conference in San Francisco today to see Infopia's eBay/Salesforce Mashup demoed. I played matchmaker for Infopia and Salesforce -- I hooked them up about a month ago, and in a very short time Infopia integrated their product into Salesforce. It seemed to get a good reception at the show. Here's Eric and Ralf from Infopia demoing the product to a Salesforce employee:

Infopiaproductdemo

Here are more pics of the conference.

For three consecutive weeks I've been at conferences up in SF. First was the Ajax Experience, where I learned that Dojo and Prototype/Script.acu.lous (Scrip.tacul.ous? who knows...) are the hot Javascript frameworks. Then was JavaOne, which hasn't seemed to change much since I was last there five years ago. The exhibition hall is still by far the most exciting thing there. Then this conference, which was a totally different crowd from the conferences I normally attend. Lots of dudes in suits. One guy did have a Firefox t-shirt, though.

I hadn't heard Marc Benioff speak before -- that guy is good! It shows how passionate he is about technology, and he way more techno-hip than me. (he has a Writely account, and he demoed it on stage)

Other stuff of interest from the show:

  • Lots of mentions of various online spreadsheets. JotSpot Tracker, iRows, and NumSum. I went and tried out iRows, and it is neat-o. Check out my spreadsheet of bands and their lead guitarists. I've set it to public read/write, so feel free to play around with it. I also tried out JotSpot Tracker -- it worked pretty well too (no public sharing in Tracker. sorry.) Don't like the name, though. "Tracker" is the name of eBay's internal project management tool, and I'd have a hard time shedding the word "spreadsheet."
  • Ken Norton gave a great talk on application design. Ken doesn't blog often, but he has some must read posts, including this one on his Enron experience.
  • Marc Benioff thinks Google hosted email is going to be very big. He wants to get rid of his Exchange servers. Is it just me, or does Marc seem to really dislike software?

If you are going to the eBay Developers Conference you can see Adam Gross from Salesforce show off the Infopia/eBay app and the Skype application for AppExchange.

May 25, 2006 in eBay, Tech/Web | Permalink | Comments (0) | TrackBack (0)

»
Subscribe to this blog's feed
View Alan Lewis's profile on LinkedIn
See how we're connected

Recent Posts

  • Alan Lewis Has Joined Auctiva
  • Mozilla Weave -- Towards a User-Centric Internet
  • The eBay Article You Must Read
  • Mmmm... Chocolate Jobs
  • Alan Lewis Has Left eBay and Joined Ning
  • Blogging from my phone
  • eBay iPhone App Shown at WWDC
  • eBay Desktop 1.0 is out. Whew.
  • I'm Supporting Barack Obama
  • Adobe - Go All In on Open Source!

Categories

  • AIR
  • Apollo
  • Apple
  • Auctiva
  • Business Rules
  • Desktop
  • eBay
  • eBay Desktop
  • Food
  • Google
  • Hardware
  • Life
  • Likes
  • mashupcamp
  • Microsoft
  • MIX06
  • Mobile
  • MySpace
  • Ning
  • Programming
  • San Dimas
  • Skype
  • Software
  • Tech/Web
  • Weblogs
  • Yahoo

People

  • Rolf Skyberg
  • Bjorn Behrendt
  • Seán Crotty
  • Rogelio Choy
  • Delyn Simons
  • Julia Wilkinson
  • Greg Isaacs
  • Adam Trachtenberg

---

Blog powered by TypePad
Member since 01/2005