2008 Mobile Predictions: Location, location, location December 27, 2007
Posted by kevin in : LBS , add a commentOn Dec 4, Jeremy Liew of Lightspeed Venture wrote on behalf Jake Seid,
Location- based services can open powerful new business models for carriers and compelling new applications for consumers. Intent can be deferred from location and location can also significantly increase the relevance and utility of mobile services and ads. As an advertiser, imagine not only knowing that someone clicked on your mobile ad, but that they also requested directions on how to get to your store. Take this one step further and imagine that (without compromising privacy) an advertiser only pays when a user clicks, navigates and then arrives at the store. Location- based services are only possible in the mobile medium and have an opportunity to create significant utility for the consumer and value for advertisers and carriers.
This is the second LBS prediction I found for 2008. This is encouraging as Lightspeed Venture is a major Web 2.0 firm in the valley.
The Mobile Web Challenge December 27, 2007
Posted by kevin in : mobile Web , add a commentOpenLandmark and Fon11 are mobile Web systems. Mo’Blast is committed to this strategy. Like any new technology, it has challenges, which I would like to share with our users.
OpenLandmark is currently available on desktop Web browsers, and it will run on the latest Firebox, Safari and IE browsers. It will not run on IE 5; IE 6 is okay with minor layout problems; and it’s best in IE 7.
For the mobile Web, our reference is the iPhone Safari browser. We chose the iPhone for its cool user interface. As a result, OpenLandmark mobile web will only run on the iPhone for now.
Because we have OpenLandmark on the desktop Web and mobile Web, we have a filter that redirects you to the correct system. On desktop browsers, including Safari, you will be directed to the desktop Web system. On the iPhone, you will be directed to the mobile Web. Of course, if you know the path, you can get to the OpenLandmark mobile Web system with a Safari browser. OpenLandmark mobile Web only works on the Safari browser.
Today, Fon11 only runs on the iPhone Safari browser. We have a plan to have it run on every phone browser (see below). On desktop browsers, we only have a static site showing the iphone product. The Fon11 user experience is very iPhone.
We are still debating whether to keep the iPhone Fon11 user interface for other phones. The client code (JavaScript, HTML and CSS) should be portable to other browsers. We are very interested in having both Fon11 and OpenLandmark run on other mobile browsers. First, we need to get our hands on the phones to play with the user experience. Most of the improvement in our mobile Web products did not come about until we started playing with them on the iPhone. We do not believe you can build a good product just with the emulator.
Our plan is to release JavaScript and REST APIs required to support Fon11 and OpenLandmark mobile Web systems. This will come soon, as we now have a development sand box configuration in our server farm. We just need to do some developer’s documentation. Our goal is to invite independent developers to port the products to other phone browsers. Our model is to push high-value location-based ads to the Fon11 and OpenLandmark pages, and we will share the ad revenue with the developer who ported the respective web client.
You can now post to any metro December 25, 2007
Posted by kevin in : OpenLandmark , add a commentYou can now create landmarks in any metro around the world. So you know we added some landmarks on the Las Vegas strips. I should note that we still needed to do some other minor adjustments. For example, the default radar in certain metros, including Las Vegas, is not on the action.
Fiona, you can proceed with Paris. Tim and Weng Kett are going to do Kuala Lumpur.
2008: The Year of the Location-Based Services? December 25, 2007
Posted by kevin in : LBS , add a commentI’m tracking blogs saying 2008 is the year of the location-based services. Om Mailk predicted upswing in M&A activities. He is still looking for a cool LBS service that let him find what’s cool within few block radius. I need to show him OpenLandmark and Fon11 on iPhone.
Fon11 Alpha December 25, 2007
Posted by kevin in : Fon11 , add a commentIf you are a Fon11 alpha user, you should be aware of the following with the latest drop:
(1) Don’t use single quotation mark (apostrophe) in name and message fields. It will mass up the JavaScript
(2) You need to click on Country twice to get the selection dial.
(3) We are still debating how to sort Contacts. Most likely, we will add the favorite feature.
(4) We are still having problem with double byte languages.
The system is quite functional now. Thank you for alpha testing the product. Please continue to send me feedback.
PS: If you are an iPhone user and would like to check out Fon11, send me mail at kevin@moblast.com.
Why we enable so many metros without data? December 17, 2007
Posted by kevin in : LBS , add a commentNormally, you would focus on one geographical area when you ship a product. We did and that was Berkeley and San Francisco. There are reasons why we add so many metros without landmark data. One reason is the way we partitioned our data for scalability. If we only have the three metros (San Francisco, Silicon Valley, and Paris), all posting outside of these areas (denote by latitude and longitude) will be clustered in the servers serving these three areas. Then, the Most Recent page would not be usable. Secondly, we also need a way for our Fon11 users to fix current posting in other states and countries.
Therefore, we elected to enable all metros. So you know, currently we are only focused on building out San Francisco and Silicon Valley. I will keep you posted when other metros begin to take shape.
Fall 2007 Capstone Lecture on iPhone and Carrier Grade Architecture December 8, 2007
Posted by kevin in : random , add a commentVijay Eranti and I teach software engineering at SJSU. This semester we have about 150 graduate and professional students from Silicon Valley including a group of students from IBM SV Lab. Vijay’s course covers Web UI development. I cover backend architecture, mainly how to build scalable Web 2.0 app using Enterprise Java?
At the end of each semester, we bring our students together for a capstone lecture where we present the next wave of technologies. The hot topics for next semester include:
• How to write iPhone Mobile Web? Stay with standards
• Divide your system into tiny apps to conquer scalability and high availability
• Carrier Grade. What are the issues? From 10 millions to 100 millions users?
• GlassFish:SailFin: SIP Servlet container for presence and VoIP applications. (Jagadesh Munta (GlassFish’s QA Lead) and Binod (SaliFin architect) will present this topic.)
Where? SJSU Engineering Building Auditorium Room 189
When? Sat Dec 8. 9:30 AM to 4:30 PM
$: Free
Come and see a sneak preview of OpenLandmark and Fon11 on iPhone. Lecture is open to public. Anyone is welcome.
We shipped! December 3, 2007
Posted by kevin in : random , add a commentFinally, we shipped Open Landmark. It has been a fun, but exhausted six months in this Sunnyvale apartment. Salute to the team: Fang, Florent, Patrick, Tim, Weng Kett, William, Francois, and Marvin the Cat. Thank you Yih-wen for the support.
The job is not done yet. Open Landmark is just the beginning. It is the base. From it, we have a lot of other ideas. It is easier now to build out the rest. Fon11 should be out by the end of this week.
The URL for Open Landmark is www.openlandmark.com (same for iPhone). Please give us a few days to load our photos.
Why scalability is ever more important? September 19, 2007
Posted by kevin in : random , 3commentsI have been following Adonomic lately which make me think about scalability. On the Leaderboard, the widgets are from a handful of companies. When a widget becomes a mover or shaker, it moves fast. It must be able to scalable to face this fast ramp-up. If not, it will never reach the Leaderboard with 7+ million users.
There are so many Facebook apps today that I wonder the developers thought about scalability. Yesterday, I asked several students to add 5 new apps on their page. They say the apps were interesting but too slow and they have to take them out. Interesting! I would say scalability is ever more important with tiny-app (widgets).
Facebook canvas pages work on a proxy model. The widget servers must be able to handle the load Facebook throws at it. On Facebook, when you get traction, it means growing from zero to millions in a week. Therefore, you are planning to create a popular widget then you’ve to plan and benchmark for high concurrency situation. Otherwise, you are testing the water for Slide and RockYou boys to quickly duplicate your widgets.
Before I go any further let’s visit the recent blogs on scalability. The first is Tim O’Reilly on Hadoop, where Tim elaborated on Yahoo’s position.
While nearly everyone agrees that the “divide-and-conquer using lots of cheap hardware” approach to breaking down large problems is the only way to scale, doing so is not easy.
No further explanation is needed on this statement. The second is Todd Hoff’s High Scalability site where you can read about how people scale large architectures.
At Mo’Blast, high scalability is one of my hot button. Scalability is more important for us because we are building a platform, not a single Web 2.0 app. Before Facebook and before we have Open Landmark working, I pounce on engineers to divide and conquer. I wasn’t easy. Now, with the development infrastructure in place especially our frameworks, built and performance test processes. Now, everyone is on the same wagon. Then, there is the issue of seeing other people going to market earlier with un-tiered Ruby-on-Rail (RoR) or PHP apps using virtual hosting. On the other hand, we build our own servers.
At the university, whenever a young entrepreneur comes to me for advice. It was very difficult to convince them to think about scalability. Before, the strategy of getting something out there fast and if it gets a lot of eyeballs scalability is a good problem to have. Investors will jump on this problem. This strategy no longer works with Facebook widget because they are tiny apps. If your widget gets traction and if it does not scale someone, someone already with an infrastructure, such as Slide and RockYou, can drop the same widget in a day or two. Have you ever wonder why with the technology they have at Slide and RockYou, i.e. video sharing, that they are doing simple widgets, like Top Friends and Questions. These are very basic text with a photo widgets.
For me, the idea of tiny apps (widgets) running on Facebook, iGoogle, Mac Widget and shortly on mobile devices is a very powerful concept that makes us itches. Actually, to architect a high scalability you have to thing tiny apps. Basically, you want an architecture that divides into tiny apps that you can conquer by scaling on many servers. Not only this enables you plug in more servers to scale, but also allows you to reuse the tiny apps and frameworks across multiple apps.
Red Bull Hanger 7 in Salzbury, Austria September 7, 2007
Posted by kevin in : random , add a comment3 billion bottles Red Bull sold last year. Dietrich Mateschitz created this incredible structure in Salzbury, Austria to house his 15 show planes. The building is also a chic eatery with visiting chefs from around the world.
Check out the conference room at the roof of the glass building. This is brand marketing. Read Deitrich’s story here.