User Interface - You know what pisses me off!!

Just consider three email internet interfaces. Gmail, Yahoo, and MSN. The latter two have really bad interfaces. The latter two interfaces are images & advertisement ridden and they never start off with emails. After I log in, they start off by showing me "Today's top news". What crap!! It might be more understandable if they put up a huge ad. With that ad I might even be sympathetic to their revenue cause. But they are showing me "top news". This is crap. No wonder, people are moving to Gmail. And to think of it, Microsoft and Yahoo are planning to get in bed. Man, that would be a really ugly looking progeny!!

Sun’s blackbox – That’s neat innovation!!

Yesterday over dinner & beer, one of my close friends who works for Sun Microsystem told me about Niagara processors and then he digressed to Sun’s BlackBox System (or now nicknamed Sun MD). Since he had got my attention, I ended up spending sometime reading on the Blackbox. Wow, now that’s a neat innovation. Note, it is not a technical innovation where people brag about 10% more CPU cycles. This is end-user innovation. People need datacenters, which can be deployed in a very short period, can be placed anywhere, can be easily scaled on demand, and saves money (primarily by saving power). All the components for BlackBox datacenter were available for so long but it required vision and risking a lot of money on intuition. Sun ran through the risk and the results are spectacular!! Lately, as computing is moving towards internet and enterprise computing, the need for such datacenter would grow substantially especially the cheap ones. Companies like Google, Yahoo, Ebay, Facebook, etc. are just some of big names. But the real masses will also be involved because I can just take one of the Blackbox and place it anywhere in the world without the need to create a building around it. Just imagine, Cell phone, Wimax, Wireless, DishTV, etc. all use cell towers. These cell towers might not be too intelligent at this point and all the intelligence resides at one focal point. With such innovation I can embed a group of cell towers with one Blackbox and my infrastructure becomes intelligent, more robust against attacks, and with a lot better routing algorithms. The possibilities are numerous. I believe, this is a thought leadership innovation and it will change if not shake the datacenter market. The datacenters will become more distributed, more independent, easy to scale up with lower barrier to entry.
Having said that Sun launched the blackbox in Mid 2007. Since, it is still datacenter and still expensive (around $1 million), it will see a typical datacenter adoption cycle. It will take some time to realize the breakeven for this product. Secondly, the major ROI will be from AMC (Annual Maintanence Contract) and probability of Sun’s more muscle power to push its Sun Server machines such as T1000 & T2000.
All said, with this innovation, I am ready to invest in “Java” (Sun’s Stock ticker is Java). The stock will see growing confidence from the investors. The only question left is how does Sun lock out its competitors (like Dell, HP, & IBM).

Good job all the nerds at Sun. We want such innovation!!

Data and future of Data miners

A lot of internet data has been mined already. What would be next sources of data for DataMiner companies. Okay, so data that is being already mined are:
(a) Internet webpages – Most of the internet search companies have tackled webpage mining. The new focus will move towards to vertical domain searches. Successful business model and easy to great ROI from digital advertizing.
(b) Blogs, wikis, usenets, forums, news – Again, in last couple of years, there has been a rush by every search company to index any sort of text data. Again great ROI from digital advertizing.
(c) Map Information – About 3-4 years back google started its maps.google.com. At that time, the map data wasn’t that good. It just contained static information like cities, mountains, and roads. As the time passes, all the local businesses, personal houses, regions would start showing up on the maps. Google.com. It will be a huge value add especially when it could be tied to a users intentions and good local business information. Local businesses with good content delivery system to end users would be a great money earner and would generate multiplicative ROI.
(d) Images – Even after decades of research, this field has not seen success. Currently, most of successful image search engines are based on Tagging. It is highly likely that in near future this trend will continue. True image searching is far from mature to be adopted for searching applications. Business model is hard as ROI is hard to justify as digital advertizing would be successful in a limited manner.
(e) Library/books – Again a lot to do with Image processing and Optical Character Recognition (OCRs). This is still nascent and plagued with copyrights etc. Easy to justify ROI as digital advertizing would be highly successful in this segment.
(f) Videos – Video searches had seen stunted growth in previous decades because of computation power and network bandwidth. In future, the video searches would see a growing market share primarily as a way of sharing extremely high bandwidth communication channel between two people. Clarifying the above point, reading a book takes time, listening to audio podcasts removes the visual context, whereas a video gives both the aspects. Video is still a passive communication channel and if it can be made to interactive communication channel then IMHO it could be a killer application. Again just user video segment has a difficult business model and has a hard to justify ROI as it is unlikely that digital advertizing would be successful in this segment.
(g) Social Network – Currently the most hyped end-user innovation in web is social networking. Fortunately, this segment is such that it also allows new small ventures to also make their mark and does not require a huge investment in terms of research, development, or infrastructure. Business model is based on number of users rather than ROI. The question is how successful would digital advertizing as a money earner be in this segment. Or is the question not the success of direct digital advertizing but more personalized advertizing. Social network mining can substantially increase the quality of digital advertizing for more personalized advertisements.
So, where is the next data source for the Data Miners to attack?
(a) Vertical domains: Such as Government sector, Architectural CAD/CAM drawings, health checkup data etc. It would be difficult to get this data because of legal & privacy issues. Otherwise it could generate great ROI from digital advertizing.
(b) My day to day life data: Such as Grocery shopping, to my traffic pattern, to my fitness program data. Unfortunately, such data collection equipment, techniques, & technology is still under development. Otherwise, it could again generate great ROI from digital advertizing.
Yeah, this is sort of a stupid post!! But anyways, nobody is reading my posts, so I will continue to be stupid J.