Code Warriors Debate Whitehouse.gov Robot Commands

As the tech community poured over the new whitehouse.gov site, one of the first subterranean changes noted was that of a file most people would never notice called robots.txt. This file serves as a notice to search robots informing them of what files they should or shouldn't survey. Upon seeing the new version of the file, some noticed that it only had two lines of code excluding robot searches vs. the former whitehouse.gov robots.txt that had nearly 2400 lines of exclude lines by the end of the Bush administration, sparking excitement and controversy over what the change means in terms of government transparency.

The text from the new robots.txt file:

User-agent: *

Disallow: /includes/

A sampling from near the end of the previous file:

Disallow: /president/text

Disallow: /president/waronterror/iraq200404/text
Disallow: /president/waronterror/photoessay/text
Disallow: /president/winterwonderland/iraq
Disallow: /president/winterwonderland/text
Disallow: /president/world-leaders/iraq
Disallow: /president/world-leaders/text
Disallow: /president/worldunites/iraq
Disallow: /president/worldunites/text

Cory Doctorow, Editor of Boing Boing and Former Outreach Director for the Electronic Frontier Foundation was one of the first to report this finding, with just the facts followed by a bunch of commenters asking for explanations.

Proponents of the belief that the move to the vastly smaller file was a statement about transparency claimed were ecstatic. According to Patrick Thibodeau of ComputerWorld, New York blogger James Kottke "thinks that by eliminating the Bush disallow list on its first day in office, the Obama administration was sending out a symbolic message." Kottke, in his post on Tuesday, alluded to the "huge change in the executive branch of the US government." In e-mail to Thibodeau, Kottke wrote: "One of Obama's big talking points during the campaign and transition was a desire for a more transparent government, and the spare robots.txt file is a symbol of that desire."

Presenting an alternate view, Declan McCullagh of CNET News pointed out that the Bush whitehouse.gov robots.txt file followed the letter of coder law for the most part in terms of what to disallow with the exception of a few incidents that were corrected. McCullagh brought to attention the idea that perhaps the new robots.txt file is actually too short. "It doesn't currently block search pages, meaning they'll show up on search engines--something that most site operators don't want and which runs afoul of Google's Webmaster guidelines."

While most of the technical experts weighing in suggest and expect that the robots.txt file should grow, most of them explain it as just a normal process a website undergoes over time. Andy John, a search developer for DeepDyve, puts it like this: "robots.txt is just a request. Robots can do whatever they like anyway." He then went further to describe what that means. "For example, there is a program "wget" (web get). You give it a URL, it downloads it and saves the file... You can tell it to download an entire site. It honors robots.txt by default. But by just adding these parameters you can tell it to ignore robots.txt and get everything: wget -- erobots=off whitehouse.gov.

As to why those who developed the new whitehouse.gov site would want to code it this way, Jaelithe Judy, a Search Engine Optimization specialist and political blogger says "Google does generally encourage webmasters to use disallows to keep from having their search pages spidered; this is to help keep a Google search from returning a whole page of search results from other sites' internal search engines, instead of relevant original content. However, in some cases a search result from a site is a meaningful result. For instance, when you are searching for 'DVD recorders' and the Amazon search page for 'DVD recorders on Amazon' pops up, that might actually be useful to most users."

She added that "Google is still trying to work out how to sort annoying search-generated page results from the useful ones. The Whitehouse.gov ones may lean toward being useful. For instance, if you are a middle school student doing a report on the First Ladies, and you get a Whitehouse.gov search page for First Ladies, that has all sorts of different links to different sorts of information, that might actually be useful."

The bottom line about robots.txt? John says, "It's really more of a serving suggestion."

Cross posted from techPresident and the Personal Democracy Forum.

Comments

[lapoutlander]

Just as likely they cleared it out to hide things. The robots.txt method is generally considered obsolete for securing webpages. Pages should be first protected by stealth. Use unique directory names and set permissions. Use server side scripting to generate the pages rather than expose the complete source. Interactive elements can help identify if a human is on the other end. Secondly use the META robots tag inside any generated page to protect especially any search page (form is META NAME="ROBOTS" CONTENT="NOINDEX, NOFOLLOW").

The robots.txt file actually makes it more likely that pages you don't want exposed will be cataloged by some bots. It is analogous to having a do-not -call list that anybody can access online. Result being that those very phone numbers get targeted more by the fraudsters and bad telemarketers. However if your webmaster is stupid enough to leave things out in default directories then you might as well create the fool-proven robots.txt­.

[RadCenter]

I hope the tech community "pored" over the site instead of "poured" over it. That could be messy.

[Sarah Granger - Huffpost Blogger]

Yes, thanks. Editors missed that one too.