Tuesday, September 17, 2013

Google's latest Chrome OS moves

Google has announced a renewed focus on the Chrome OS operating system.  Many of my readers know that I invented the Chrome OS webtop back in 2006 while working at Google, so I thought I'd answer some questions that people have asked on the web about the first version we built at Google.


Did you work with the Chrome team or anyone else on Chromebook?

Of course. The Chrome Team played a key role later in 2007.  Mike Jazayeri, Director of Chrome and Apps, assisted more than any other single person.  He was heavily involved in organizing and pushing through the project with the Executive Team at Google.  He ran the meetings with Larry Page and the other executives, as well as the meetings with external partners like HP and Asus.

It was SVP Jeff Huber who supervised the Chrome team and also gave the initial executive level approval for Chromebook, including Google's most valuable resource, full-time engineers.

Other members of the Chrome Team did the core work of porting Chrome to Linux and helped me get it running, including Ben Goodger and Brian Rakowski.

It's impossible for me to overstate the importance of the Google Apps team, particularly the Google Docs project.  If it had not been that Google had already completed a huge number of webapps to replace the functionality of a consumer operation system by July 2006, a consumer operating system based on web apps could not have been written.

Hundreds of other Google engineers provided testing and feedback for the operating system.


What are your thoughts on Google latest announcement regarding improving Chrome OS' offline capabilities and positioning it as a full fledge operating system?

I always thought the Chrome OS operating system could be marketed as a competitor to Microsoft.  Much of the original reception, back in 2009, was that Chromebook was just a toy for occasionally browsing the web.  I never thought of it that way.

When I came up with the invention, I used it myself every day as my main computer while I was working at Google.  I believe the business buzz word for that is "dogfooding". In my opinion, you should always, always dogfood your product, otherwise its difficult to understand your customers and their needs.

How do you square that with consumers who are used to Windows?

Consumers are used to this OS architecture where you have files and applications, and you carry all your files around with you on a hard disk, and a PC is a CPU, a hard disk, a Wifi, etc.  But there's really no reason it has to remain that way or even that those expectations are the best way to interact with a computer.  That's just what people are used to.

Microsoft, I certainly feel, has stagnated to a huge degree.  Their operating system hasn't significantly changed in the last 20 or so years.  That's unheard of in the PC industry.  You don't have software business that have products that remain identical and unchanging for decades.  If you don't innovate in software, you tend to go out of business, very quickly.

Some people have questioned the RAM-based architecture of the original Chrome OS product you came up with in 2006?

The original version was based on a Linux kernel called PuppyLinux, which I loved, and still love for that matter.  Puppy is an operating system designed from the ground up to boot entirely into RAM, so it was very, very fast.  It was the perfect starting point for a lightweight web-based operating system, when you have the expectations that you aren't going to have applications and data sitting on the same computer. Again, once you get over the hurdle of teaching consumers that their files aren't sitting on the same computer, it becomes kind of natural to expect the OS itself can just boot to RAM.

The main reason is simply speed.  When you look at the BUS architutecture of something like the Samsung Chromebook, its running on an MMC bus plugged into an SSD.  Well, the MMC bus is only 30 Meg/sec, that's crazy slow.  Why would you ever plug an SSD into such a slow BUS?  I opted for a RAM bus architecture instead, because the RAM BUS is generally going to be in the 1000 Gigabit range, thousands of times faster than MMC.  That was a very easy decision.

Your blog mentions HP and ASUS were in talks with Google prior to their webtop product launches in 2008?


Google had some informal talks with HP and Asus about partnering on Chrome OS back in 2007. We demoed the operating system for the companies and showed them a bunch of screenshots and product roadmaps.

I don't have access to any sources inside HP or Asus to get the full details, but it appears from the outside that when the partnership talks with Google broke down in 2008, HP and Asus decided to go their own direction and launched their own webtop offerings without Google.  As of 2013, though, Asus and HP are back as Chrome OS partners.


Why do you think the originally HP and Asus offerings were not more successful as compared to Chrome OS?

Good question. You'd have to do a market study to find out the exact answer to that, regarding how individual customers made their buying decisions.

From a technical and marketing perspective, Google was smart to launch it as an open source operating system, positioned as a cheap, secure, easily administrated alternative to Windows.  I think that message resonated with many businesses and consumers.

So you invented the webtop, not just Chrome OS?

I would offer a qualified yes.  I described the invention process and the work I was doing back in 2006 and 2007 at Google in my first blog post here.  All the products that were launched in 2008 and 2009, by Google, HP and Asus, can all be traced back to my work and the meetings between Google and potential hardware partners that took place in 2007.

However, that is a qualified yes, because the term 'webtop' has become so overloaded - it has many different definitions and has been used to describe many entirely different products.  Motorola even has a smartphone docking station that they call the 'Webtop'.  SCO (a Linux company) originally coined the term 'webtop' to describe a browser plugin.

I invented the idea that you could write an entire fully functional operating system entirely with browser components and web services.  To the extent that is definition of 'webtop' you are referring to, then yes, absolutely.

Did Google pay you a bonus for inventing Chrome OS and filing the patent? 

Google paid me a small bonus, a few weeks before they publicly announced Chrome OS in 2009.  I had not been an employee in well over a year.

Why did you leave Google?

I tend to make decisions like that for many reasons - not just any one single reason.

I suppose one of the issues was that I personally failed to get as much traction with Google management on the Chrome OS project as I felt the project deserved.  We had full-time head count slotted for the project.  We even had some backing from several executives and had demo-ed in front of Larry Page, but I personally still had other responsibilities. (I was programming web services for the Google Apps team, at that time.)  I repeatedly asked to transfer to the project that I had founded, instead.  At some point, it all came to a head, and I found myself walking out the door.

There were other reasons, but I think that if I had received more direct support from Google management at that time, I probably would not have left.
I see on your web site that you also worked at eBay.  What was your favorite company to work for, Google or eBay?

Well, I've worked for many companies, not just Google and eBay.  I was just recalling this the other day, I got my first IT job doing data entry as a temp working for a bank, Arizona Federal Credit Union.  That was 27 years ago.  Yikes.

It's a difficult question to answer, you know.  Each company is different, each culture is different. 

I remember this one tiny company called Shockmarket.  When I was hired, everyone else in the company had PhDs, except for me.  (I dropped out of my PhD program to pursue my first Internet startup back in 1995.) That was a fun company, because we were all super-motivated to make a big hit during the Internet gold rush.  We had people pulling all-nighters on a regular basis, which was cool, because you'd come up with a product idea in the afternoon - and by the next morning, you had a product ready for market.  Someone stayed up all night to write it and make it happen. Good times.

Google was also good in many respects of course.  Google gave me a long leash to experiment and do my own innovation, so in that sense I'd have to say Google.  Google has been able to hire some really outstanding people, also.  It's always nice as an engineer to be working with people who are great at their job and always exceeding expectations. 

We had some great people at eBay, also.  Many people don't know this, but eBay ramped up their hiring just as many Internet startups were failing back in 2001, so we really could pick from the cream of the crop.  We were able to get some real rockstar engineers, who have all gone on to do great things either still at eBay or at their own companies.


Any more questions?  Post them in comments and I'll respond as time permits.


Chrome OS: A Timeline

I just wanted to note down some dates relevant to my invention of Chrome OS.  You can refer to the original Chrome OS patent on the USPTO site.


July 2006.  Initial prototype of Chrome OS created based on PuppyLinux kernel.  Demo for coworkers and immediate management.

May 2007.  Published first version of Chrome OS to a internal Google company wide mailing list.

June 2007.  Executive management approves full-time head count for project. 'Chromebook' is coined.

August 2007.  Informal talks begin with hardware vendors Asus and HP to partner with Google on a webtop operating system.

January 2008.  Jeff Nelson departs Google for other opportunities.  Talks stall between Google and hardware partners, Asus and HP.

September 2008.  HP and Asus announce plans to move forward (without Google) on their own webtop products.

March 2009.  Google files patent on webtops listing Jeff Nelson as inventor.

September 2009. Google announces plan to launch Chrome OS as open source project.


Google+ | Twitter | Facebook | Quora | LinkedIn | Tumblr | jeff-nelson.com