Buy Viagra - HQ Pharmacy

The second day at IdeaFestival kicks off with a talk by Dr Craig Nevill-Manning who is Google New York Engineering Director. A native of New Zealand, Nevill-Manning joined Google from the Computer Science Program of Rutgers University and  has been with the company for the last 7 years.

Nevill-Manning tells the audience that there are five big points he wants to make in today's presentation.

His first point is the need to think broadly. He says that computer science is not just programming (the not just part is important) He wants to teach the audience computer science, so he asks for five volunteers from the adenine, who get on stage. That being America, there is no shortage of hands in the audience. Each of them gets a little piece of paper with dots on them. By now, it's clear there is going to be some fun on stage.

He asks them to sort themselves (by the number of points). Then he asks if there is a pattern to the numbers (they double each number). Then he asks the people to show exactly 23 dots, which is a bit of a puzzle (that takes them by surprise). Then he asks for nine, they do it quicker (and one man folds a a 16 dot paper) – apart from folding a piece of paper. No, there is only one way, he says. Then he asks what's the next number of dots they can show.

Now he asks to count from zero to 31, which is a slightly challenging task. After many laughs in the audience, they succeeded. He asks if there was a pattern – they were each showing their paper with some interval. Manning says the same can be done for buckets of water or anything else. Then he explains the binary number – and there is a unique combination for each number, which many in the audience find interesting. It's probably the first time I see somebody explain binary numbers in such a highly interactive manner-- I am sure everybody in the audience got the point.

He talks about teaching computer science without the use of computers. It doesn't have to be necessarily computers, he says. He is showing up csunplugged.org which is a program to teach kids computer science (now bough by Google and free).

The second point that Manning makes is the need to enable others, after which he proceeds to show how to write a little program. He asks the audience how many people have used Google Maps—pretty much everybody as turns out. Then he asks how many wrote a program using Google Maps. He talks about users who decided to reverse engineer Google Maps and how Google decided to make it much easier for people to build their own products by using Google's API and build their own stuff on top of it. Manning starts coding on this computer – the audience is starring at a white screen. He asks how many people have used “I'm Feeling Lucky” Button on Google–some people did, but very few actually know what it does.

He opens two windows on his screen, one displaying Google Maps API Concepts and the other one displaying Google search results (a little crash course on search). He proceeds to register on Google Maps and get his own GM API key. Manning doesn't pay much attention to the license, which gets him some laughs at the audience (well, nobody really reads those terms, anyway). He zooms in on the convention center and copies the geo coordinates of the place. Talks about Java Script – not sure how many people understand it. The basic point is to show how easy it is to take Google data and play with it in your own ways. He brings up the famous example of Chicago Crime Map, not many people in the audience seem to know about it. Then he shows the mash-up between Google Maps and Craiglist for house prices. He talks about people who had reverse engineered the Google Maps code before the API was released (we hire such people, sayss Manning). Another example is about buses in Seattle – something that people can easily understand. So Google doesn't build directly, but enables others to build.

The third point that Manning mentions is the need to use deep technology. He talks about spelling correction software that Google has looked at. He talks about Google's spelling correction, which corrects words based on context. He mentions the word coffee spelled as kofee pointing to either a cup of coffee or Kofi Annan. When somebody says “kofi cup”, it automatically goes to coffee, but when they type “kofee annan”, it points to Kofi Annan. So Manning talks about the algorithm that Google has developed using artificial intelligence. Manning pitches this as a handy way to spell sophisticated foreign names – should probably work in the US. He shows a graph with hundreds of possible misspellings of Britney Spear's name, which gets many laughs from the audience.

The fourth point that Manning makes is the need to build for scale. He talks about doing projects on the massive scale. The question, says Manning, is how to build such projects that deal with massive amounts of data. He shows a sign that he saw ona a lab door in Stanford “ forget quality, go for numbers”. The point is that it's sometimes important to sacrifice quality, but go for the data. He shows a quote form Rodney Brooks from MIT aobut robots that would be “fast, cheap, and out of control”. Manning says you can do a lot of things with “dumb, unreliable, massively parallel, on lots of data”. Google doesn't try to have a perfect answer every time. Instead they focus on doing things incredible broadly, sometimes the algorithms are complicated, but most often they are rather simple.

Next he points to the intersection between academia and insutry. In the industry, they build things that people will use and in academia they build things that are interesting. Google's approach is what Manning dubs “computing at scale, cheaply”. The strategy is to exploit processing poer of off-the-shelf PC hardware, and make it reliable in software. PCs are cheap and fast, says Manning, which he prefers to large framework machines.

He says that Google archives reliability through replication. THE basic principle: is to replicate everything. There are always other computers with the same data, which can always back-up if the first machine crashes. So the result, as Manning points out is that single (or even multiple)_ failures don't hurt, they only reduce capacity. Redundancy isn't waste, since they need that capacity to respond to millions of queries they get. He talks about Google's way to cure machines—it gets a bit to technical for the audience. He does a “backup” slide in big red font, which audience finds rather amusing.

Cooling failures happen at Google from time to time. Manning says they are exciting, and shows a photo with firefighters rushing to the scene. That does raise a lot of questions about what happens if Google goes offline for a day or two – how much of the world's infrastructure is now dependent on it? Manning says it's actually worse when it happens at night, since most of Google's queries come from abroad.

Manning shows a really old picture from google.stanford.edu – a bunch of really old computers (what Manning dubs heterogeneous enclosure). He shows hard disk enclosure that Larry and Sergei built out of Lego – quite a good example of user-driven innovation at low budget.

“To be effective in the Silicon Valley, a start-up needs at some point operate from a garage”, says Manning showing a picture of Goole's founders posing in one.

The fifth-- and final-- point that Manning makes is that you need to detect scales. You can do many things that are very cheap nowadays—computers are cheap, software writing is easy—so doing things is easy. But detecting trends is important. He talks about the analysis Google does on seach queries, pointing to the ways in which Google detects all new trends, while also preserving users' privacy. Points peole to Google Trends before showing a slide on Google Zeitgeist. He gives an example from “who wants to be a millionaire”, where the contestant who has a 1 one million question to answer makes a phoen call to his friend, and asks the question to google the name of the winner. Later they checked how many queries came for tha name from Google users. You see the difference between the West coast and the East coast – people were searching for the name after seeing the question on TV.

Then Manning shows daily traffic patterns, comparing USA, Spain, and France – there are some time differences.

At this point, the discussion goes into Q&A. Somebody asks about the number of PCs that Google has. Manning points to the dropping costs of computer technology, saying that now Google has many computers all working at different speeds.

Another question about explosion of information in live sciences and the question is whether Google will build a search engine for it. Manning says Google has invested ina number of companies that do search in live sciences and that's definitively on Google's agenda This live science data is incredible important, so one may expect that Google will move into that space rather quickly.

The question is what's the next big thing for Google. Manning says, there are a few. Manning talks about verbally-queried search, and says that in 10 or even 5 years, we'll look back and understand that 's very primitive how we search now. Voice-recognition and unified answer with some sort of visualization are the areas where Google is looking.

Question about local /niche search and about Google's attitude to this. Manning says that Google already helps to find local businesses, but he does agree that there is still a lot of untapped information about there (I wonder why he doesn't mention Google Business Referral program). Google wants to build something much more ambitious than another Yellow Pages. Working closely with gadget makers – Apple/iPhone – is also a priority, so that people can get access to data they need on their own devices.

Another interesting question is about Google culture and fostering creativity. Free food, free massages are there, but one of the best perks is that every employee can spend 20% of the time doing whatever they want. So, that's Google's way to promote entrepreneurial thinking – that's where many of Google products come from. Flexibility and the ability to organize their time the way they want is crucial to innovation success. He says they have too few managers, which makes software engineers collaborate with each other. It's also great because as there is a shortage of managers, nobody tells software engineers what to do. So they hire people who they think are going to have good judgements. That is the end of Manning's presentation.