Q&A: Google's Alan Noble on the future Web

Alan Noble, engineering and site director, Google Australia.

Alan Noble is the engineering and site director for Google Australia. ZDNet.com.au sat down with him to find out about the future of the Web, and what Google really thinks about Microsoft's move into online applications.

Where is the Web going?
Noble: There are a couple of pretty amazing trends happening. One such trend is the move towards gadgets, mini-applications or widgets. Gadgets are doing for applications what RSS did for content, they are democratising it and making it possible for developers to disaggregate applications in a way that have never been done before. It's exciting.

I think we are going to see a lot more gadgets, and we are going to see these gadgets syndicated across the Web on hundreds of thousands of different Web sites. It's not just the big guys here, and that is really important.

Another major trend is the almost inexorable move towards cloud computing. That trend has started, it will continue, and I think it will have profound implications. Computing as we know it becomes a utility.

In same way we had the advent of the electricity utility 100 years ago, at the time it was all about street lighting. Little did people predict it would be about every [electronic] device on the planet. I think we are in for some pretty profound changes we just don't really anticipate right now.

For example, our Google Apps -- we have taken Gmail, our Google Calendar, Google Docs, Google Spreadsheets, and taken a whole range of applications that users traditionally thought of as client side applications and moved them online. There are a bunch of applications that reside in the cloud.

It basically means you have access to your applications anytime anywhere. Today I am lugging around a laptop, but if I had access to a public terminal, I could access all my documents, my spreadsheets, everything.

It's early days, and we need a lot more applications to move to the cloud, but the trend is clear.

Microsoft recently launched Office Live Workspace beta. How is Google dealing with competition in this area?
We have a view that collaboration has to be at the heart of applications, I think Microsoft have a different view. My view is that they haven't completely decoupled themselves from the desktop. Our view is that the Web is the platform, and the desktop is nothing more than how you access the Web platform. [Microsoft] are still wed to client side applications and that complimented by server side applications. We might say that ours is a purer model, a pure server side model.

The threat of a US recession is forcing down online advertising budgets. Will this push Google into different business models?
I don't think a recession alone would force us to embrace new business models, but I think that [fear of a recession] is certainly a reality in the US. Already we get a lot of our revenue from outside of the US. In some respects we have a bit of immunity by virtue of our global footprint. We operate in over 100 countries, we do research and development in over 40 countries. We have a true global footprint as a company, which will help us, even if the US goes into recession.

Clearly advertising has been at the heart of what we have done in our search business. We are looking at other ways to grow the business. Apps are certainly becoming important, and we will be looking to grow the apps business.

Can you give us an insight into Android?
The Open Handset Alliance, or Android -- is a completely open source operating system and application platform for mobile phones. Essentially we are trying to open up mobile computing. We think for too long there have been propriety operating systems from vendors who haven't really wanted to make their mobile platforms available to third party applications.

Android is about opening things up. We think it is about increasing the size of the pie. There will always be a demand for high-end propriety systems -- the Apple iPhone is an example of that. We think, especially in the developing world, like China and India, there is a huge potential for low cost phones that can run useful applications. Right now the cost of the operating system is prohibitively expensive. Essentially that is making it very difficult to deploy low cost phones.

If you take that out, and the OS and application stack becomes free, [we think] it will put a lot more interesting and capable handsets out there.

Recently, a group of Australian Football League (AFL) sites were labelled as malicious. How does Google remain equitable when so many businesses rely on it for referrals?
I am not intimately familiar with [The AFL site incident] -- but part of our technology is to try and defend against sites that have malware.

We are equitable by keeping humans out of the loop. We are firm believers in fair, algorithmical approaches. You may recall that when Gmail first launched, there were a lot of people upset with the notion of ads in Gmail. 'Oh my goodness, there is someone or something reading my e-mail and targeting ads at me.' Of course, it wasn't someone at all, it was an algorithm that was trying to match the content [with ads]. Now that has completely died down now, and people essentially accept it.

We are a reflection of the Internet, and we try and do this in investing heavily in the best possible processes and algorithms to do that. There are imperfections obviously, but I can step back and say that we can never stop innovating, because things like this happen. We have to continue to improve.

We realise that when anyone uses a search service, whether it's Google or Yahoo or Microsoft, you are really only one click away from another search engine. There is no switching cost. That's why we put so much effort and R&D into making sure our offering is as user friendly and accurate as possible.

As Web content becomes richer and richer, how is Google going to overcome the software engineering challenges associated with searching Web content such as videos -- which can't be searched -- and images?
I don't agree that you can't search a video, or can't search images. Admittedly some of the techniques are fairly rudimentary. Essentially there are image processing techniques that can glean the gist of a video. I have seen research here in Australia based out of the University of Queensland, where they have been able to classify videos and detect similarities with other videos.

It's early days for image search. Clearly we can do better, and we will do better. Certainly it's a valid point, the Internet is becoming richer and being able to search that is crucial. Trust me -- that's why we have lots of engineers working on these problems.

Is there a huge increase in computing power associated with these complex image processing techniques?
Yeah, there is. It's computationally expensive.

How do you plan to meet the challenge of rising search costs?
[Laughs] We meet those challenges by building huge datacentres, which are very expensive. However we do them in a way which scales. Anyone can build a datacentre, but not anyone can build a datacentre as efficient as a Google datacentre.

Users often associate Google with the end result -- a search engine, or perhaps you use Google Reader or Google Docs -- but underneath all of that there is a fantastic scalable infrastructure. That's the real secret source.

Author's note: Alan Noble wrote to ZDNet.com.au after the publication of this story regarding the final sentence. His preferred quotation is "secret sauce".