October 2007 Archives

The hot news of these days is Google's response to the Facebook API - OpenSocial.  I had breakfast with a Facebook employee today, and he was already asking me if I would see any value in developing using OpenSocial.  The competition is hot.  So many people are writing about it, and it's hard to find a useful summary because there's not much information available yet, but here is a rundown of what I consider significant about OpenSocial:

• It's really open: The API would allow writing applications for Orkut, Salesforce, LinkedIn, Ning, Hi5, Plaxo, and Friendster (oh, and Oracle - yes that made me wonder too).  TechCrunch has a gallery of some screen shots of certain applications.
  
• The application has access to shared data right on Google's servers - so you're riding on Google's scalability.
• There is no special markup language (FBML in the case of Facebook).  It's based on JS and HTML.
• It provides programmable access to the three key elements of the social networks:  Profile, Friends, and Activities.  Facebook provides sufficient APIs for profile and friends information, but access to activities is vague.
  
• Applications seem to have more flexibility in terms of using the canvas, or the their screen area.  Facebook enforces some serious limitations there.

But there are issues too.  The biggest source of the issues, like any other type of "aggregation" or "common interface" in the world of software, is the differences of these "hosts" (i.e., social networking sites).  It's interesting to watch how the APIs will handle different data elements and different layouts of information hierarchy among these sites.  Obviously Google wants to keep OpenSocial high level to make it work with all of these hosts - we have to see how much granularity will be lost in the process.  If I as a developer end up making host-specific API calls 50% of the time because the high level, common APIs do not give me the data I need, this project has failed.

Another related but outstanding issue is that the nature of these sites is not the same.  LinkedIn is highly business oriented, Friendster is friends and their birthdays.  Yes they both have a social graph, but that's where the similarities start to fade.  The usage pattern is different and hence the activity stream.

Last but not least, there's the issue of placing ads within the applications...  Google cannot be neutral there.  Ads are the "business" of Google, and at the same time most of social sites live on displaying ads.  There's going to be some serious conflict of interest between what the application shows, what the social network shows - after all, they already have contracts, and where Google plays a role in what's displayed where.

Hmmm...  And I cannot resist.  Does anybody bet over whether Facebook will adopt OpenSocial at some point or not?

Yes I'm behind - traveling back from Orange County and LA was exhausting with all the fire and burning going on.  I was not that close to the fire, but seeing the smoke circles going up both in Irvine and in the Valley was sad and horrifying.

So I'll have some blogging to do, but this one is hot off the press...  The official press release is out.  Microsoft invests $240M in Facebook for 2% of the company.  I guess Google did Facebook some good.  The bidding landed Facebook a great valuation, and Microsoft was not in the auction to lose.  But then hey, as somebody said, Microsoft has the money in the bank and $240M to them is just "rounding error".

So far the Web 2.0 Summit in SF has been pretty uneventful.  Even the interview with Facebook's Zuckerberg did not have much interesting or newsworthy information.  Steve ballmer, however, is clearly announcing that Microsoft is in acquisition mode, so that's good to know.

Also people are making a big deal out of Murdoch/DeWolfe talk about MySpace opening up its platform.  Well, what else did you expect?  Sure they'll do it.  And they did not announce that it's there now - it's coming in the next few months, whatever that means...

Being an old time Unix fan I have always admired Bill Joy, and It's great to see a person with his caliber focusing on Green Technology.  He has been with Kleiner Perkins for a couple of years now, which is a very respected VC firm, and therefore Joy has all the right resources to do things in this space.  I'm sure great things will come out of his new focus.

He is looking at all kinds of stuff, from more interlinking of transportation and electricity grids, new chemistries for better batteries, and better use of open parking spots for solar energy accumulation, to advances in materials for more efficiency and lower cost in photovoltaic products.

And he calls Internet investment wacky: "We're still finding really high-quality ventures at an early stage where we think the prices are fair. I wouldn't say that's true in the Internet space. Things in Internet space are wacky right so it's nice to be someplace where things aren't so crowded."

As a natural next step of purchasing Tellme, Microsoft is launching LiveSearch411 with 1-800-CALL-411 as the access number (the number is not active yet). This is in direct competition to Google's 411 service (1-800-GOOG-411).  I feel bad for the 1-800-FREE-411 guys - they could have been acquired in this war (although Yahoo! is still out there without such 411 service, and might approach them).  This is more validation for voice based services such as Frucall.

Amazon announced today that EC2 now offers three different server capacities.  For the uninitiated, the significance of EC2 (Elastic Computing Cloud) is that it provides virtual computers on the Internet which one can operate via web services APIs for utility computing, and the computer can run an OS image with everything you want on it.

So you can create a Linux/Apache/PHP machine as your web server and a Linux/MySQL as your DB server, all just by clicks and API calls - and you can create as many as you want, and shut them down or boot them up any time you want.  Last but not least, EC2 server images can use Amazon S3 (Simple Storage System) distributed storage.

EC2/S3 is something to watch, not only from a technology standpoint being a real, in-production, fully virtual and distributed computing environment, but also from a business standpoint.  EC2/S3 creates a totally new pay-as-you-need approach to data servers, which makes it a lot more easier for startups to build their infrastructure with little money and then scale it up as they move forward.

The new server (or per EC2 terminology, image) capacities makes EC2 more flexible in terms of architecting a solution (e.g., heavy hitter machine for DB, lighter for web servers), or special applications (e.g., video processing, genetic applications).

Not a technology post, but since I wanted to share this with a number of friends I'm posting this information here.

There will be a few great ballet/classical performances in the area:

• Michael Flatley's Lord of the Dance:  It's even better than River Dance.
• In San Jose: Nov 23 and 24. (As of this writing the 2pm show on Saturday has the best seats)
• In San Francisco: December 1 and 2 (Tickets will go on sale Sunday 10/14 10AM).
• Two great shows with one ticket, by Ballet San Jose Nov 15-18:
• Carl Orff's Carmina Burana
  
• Summerscape - Dance cjoreographed for  Dimitri Shostakovich's Concerto No 2 in F Major for Piano and Orchestra, Opus 102

Ballet San Jose will have more of the usuals (the Nutcracker, Swan Lake, and heck, even Snow White!), but the main thing that made me excited about it was the Orff and Shostakovich shows.  If interested check out their 2007-2008 season.

Also for the fans of Persian classical music, here is the information in case you have not heard about the Dastan Ensemble concert with Salar Aghili.  San Jose performance is on Oct 28 - tickets are $35 and there is no seat assignment.

Interesting statistics on consumer behavior - bottom line is that people start with search engines for shopping, and research products online only to go and buy offline, and last but not least they do not want ads on their cell phones.

Rubicon Project has launched their beta program.  Interesting startup to watch.  They are addressing the right problem, and given the profile of their team they must be addressing it right... And, being well funded does not hurt.

For those who are interested, Google has created a repository for course content, tools, videos, etc on their topics of interest.  Specificall do not miss the video lectures on distributed computing. More about how it came about can be found here.

I have to say I had not heard of Jaiku before - but apparently it's a twitter competitor and already there is all kinds of rumors and speculations as to why Jaiku and not twitter - some are saying Yahoo! is about to acuire twitter.  Anyhow, here's the official announcement on Google Blog.

Not much to say...  Just read on!

http://www.statesman.com/news/content/news/stories/local/10/04/1004whiz.html

The latest Internet Advertisement Bureau (IAB) study (PDF) performed by the New Media Group of PricewaterhouseCoopers is out.  Online ad sales has been $4.9B in 07Q1, and $5.1B in 07Q2 - a total of $10 Billion in the first half of this year.

For comparison, the numbers were $3.8B and $4.1B in 06Q1 and 06Q2.

Steve Tobak has put together some unsolicited advice for CEOs.  Very good article including a few good links, and in particular I like the Yogi Berra quote: "If you don't know where you're going, you may not get there."

I would argue that positioning is part of strategic planning and not a topic of its own.  Strategic planning, or as a good friends used to say "having the end game in mind", mandates doing a number of different key things including the right positioning.  I'll leave that to another - probably long - post.

What I see missing from Steve's list, a critical one, is sales.  A good CEO must know how to sell.  Sorry guys, but if you're not good at sales chances are you'll not make a successful CEO.  I have seen it many times.  It's beyond selling products and services - it's selling the image and selling the company, which starts will making investors buy it, making your customers buy it, making your competition buy it, and finally making an acquisition happen.  Every presentation you do is "selling" - somebody should buy what you're saying.

And of course there's literally sales for revenue, which needs a CEO with sales background.  Without that the company will most likely shoot in the darks when it comes to generating revenue.  Non-sales CEO do not even know the right questions to ask and usually fail - at the strategic planning step - to build the path to revenue.   One might argue that some of Web 2.0 companies go for building a large user base and revenue is not important to them - again a separate topic of its own - but try to get VC funding based on that premise and then we can talk.

Anyhow, thanks Steve for putting together the article, and I hope this adds some value.

Technology has its goods and bads, but this one is certainly among the good and most welcome ones.  UCB has started posting full course lectures on youtube.com, which means just about anybody curious can now tune in, watch, and learn without paying tuition.  To get going, visit:

http://youtube.com/ucberkeley

My love for UCB started with getting to know BSD Unix and my class with Kirk McKusic on Unix kernel is my greatest academic memory.  When I learned about the role of UCB and Berkeley in 60's and 70's history I loved it even more, and this tops them all off :-)


This is a serious move and I hope other important academic bodies follow. 

Amazon's CTO introduces Dynamo, a highly available and scalable distributed data store built for Amazon's platform in a paper that he has co-authored for SOSP.  Dynamo is not S3 in case you're wondering. It's an internal data store, not available to the public via web services.

The significance of the paper is the degree of high availability required for Dynamo, and how that requirement is addressed in Dynamo's architecture design.  Companies like Yahoo!, Google and Amazon have to deal with failure constantly.  The number of moving parts in their infrastructure is so vast that at any given time you can see disks, servers, or network devices failing.  Technologies like Dynamo are built to make such failures transparent to the end uses of the system.

A PDF version of the paper is also available.  It's a long paper, but a great read for anybody interested in HA and distributed systems - this is not academic stuff.  It's a production system, which is rare and therefore very valuable.

I got to know Brian Smith (www.comparisonengines.com) during the early days of Frucall, and I have loved his work ever since.  In addition to being a great person, he is a real journalist, and is excellent at finding bits and pieces of information, putting them together, and giving you a story that is worth reading.

A few weeks ago he wrote about Jellyfish's "Smack" feature, and suggested that Jellyfish is a target for acquisition.  Admittedly I did not take it seriously (my bad, Brian!) - and what do you know...  Less than a month later Microsoft acquires them.  The suggestion was not explicit enough to make me feel Brian had inside information, I think his article was his assessment.

Anyhow, I saw the news a few days ago but did not a chance to blog it until just now.  So I went back to Brian's site to get the link to his article, and I saw his post covering the acquisition.  You should definitely read it - a lot of good information, both on the Microsoft/Jellyfish deal, and on acquisition landscape in the comparison shopping market.

Well, most of us remember the jaws dropped at the multi-billion dollar valuation of Skype when eBay acquired the company.  It look likes, however, that eBay has started to regret what they did...

Adobe acquired Virtual Ubiquity mainly because of its online word processor, Buzzword.  Buzzword is a flash-based RIA which competes with Google docs.  As you have probably heard, Microsoft has also announced their plans for what they call "Office Online" (creative name, by the way!).  So the market for web-based documents is certainly getting hotter.  This acquisition is a surprise because Adobe was not in this market, although Buzzword makes sense to Adobe being a flash-based technology.  You can find a detailed account of Buzzword here:

http://blogs.zdnet.com/SAAS/?p=393