Google and Obfuscated JavaScript

February 10, 2008

Sunday mornings are generally calm in rural Kentucky. There’s the normal pop of gun fire as my neighbors hunt squirrels for burgoo, and there is the routine salvo of news stories about Google.

I zipped through the “Google to Invest in CNet” but paused on the Google’s “obfuscated JavaScript” story here. A number of Web log and Web sites are running the news item. Google Blogoscoped’s story ran on Friday, February 8, 2008, “Why Does Google Obfuscate Their [sic] Code?” Philipp Lenssen does a good job, and this post contains a number of intriguing comments from his readers. These folks speculate on Google’s compressing JavaScript to save bandwidth; others hint that Google is intentionally creating hard to read code. A possible code siting is here.

Speculating about Google whys and wherefores is fun but semi-helpful. My hit-and-miss dealings with the company reveal “controlled chaos.” The way to get a look at what Google does is to dig through their technical papers (do it daily because the list can change) and read some of the company’s patent applications, patents, and if you are in Washington, DC, the “wrappers” available to some savvy researchers.

Some hints of the JavaScript mystery appear in this document: “Method and System for Dyanamically Composing Distributed Interactive Applications from High-Level Programming Languages”, US20080022267. The invention was filled in April 2005 anad was published on January 24, 2008. When an application is published, Google often has examples of the document’s “invention” running or even visible for the intrepid investigator. Three years is a long time for gestation at the Google. My hunch is the JavaScript is produced by the Googleplex’s auto-programming techniques, possibly the one disclosed in US20080022267.

I’m no attorney, and patents are difficult to analyze even for the experts. Read the document. You may find that the odd ball JavaScript is a way to eliminate manual drudgery for Googlers. US20080022267 may shed some light on what Google may do to spit out JavaScript for browsers, for instance. What do you think? I am toying with the idea that Google does automatic JavaScript to improve efficiency and eliminate some grunt work for its wizards.

You can obtain US20080022267 here. If you haven’t used the USPTO’s search and retrieval system, check out the sample queries. The system is sluggish, so you can try Google’s own patent service here. I’ve found that Google’s service is okay, but it’s better to go to the USPTO site, particularly for recently issued documents.

I want to conclude by saying, “I don’t think conspiracy theories are the way to think about Google.” Google’s big. It is innovative. It is — to use Google’s own term — chaotic. I think Google operates like a math club on steroids, based on my limited experience with the company.

I’m inclined to stick with the simplest explanation which seems clearly set forth in US20080022267. “Controlled chaos” is a way to disrupt monoliths, but it doesn’t lend itself to highly-targeted, human-motivated fiddling with JavaScript. Not even Google has that many spare synapse cycles.

Stephen Arnold, February 10, 2008

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta