The Hindu Business Line : When the Govt comes calling...

The US Government seeks access to portions of Google's database of people's Internet searches. And Google is unwilling to open the data gates. In the spotlight is this case of searching through searches.

YOU don't have to search for the hottest case, because it's about search: Google versus Government. Thus, "One of the most active reader polls on mercurynews.com," is the query, "Should the Government have access to portions of Google's database of people's Internet searches?" Online poll results, at the time of writing this, are "Yes 136 votes (8 per cent), No 1,668 votes (92 per cent)."

Derek Sooman writes on www.techspot.com, "Google has been ordered by the US Government to hand over data revealing what its users are searching for online - and the company has refused!" One of the postings reads, "Google is certainly gutsy and should be applauded for this. Why does the Government really want these records? I suspect more sinister reasons."

COPA challenged

If you trace back, you'd land in `Civil Action No. 98-5591', American Civil Liberties Union (ACLU) vs Janet Reno. She was the Attorney General of the US, and the case came up before the District Court for the Eastern District of Pennsylvania. "The plaintiffs, representing individuals and entities who are speakers and content providers on the World Wide Web (the `Web'), many of whom are seeking to make a profit, and users of the Web who use such sites, filed a complaint in this Court challenging the constitutionality of the recently enacted Child Online Protection Act (`COPA') under the First and Fifth Amendments," reads the judgment of Judge Reed, dated November 20, 1998.

COPA was part of the struggle of US lawmakers against the proliferation of Internet pornography, educates Wikipedia. "Parts of the earlier and much broader Communications Decency Act had been struck down as unconstitutional by the Supreme Court; COPA was a direct response to that decision, narrowing the range of material covered."

Judge Reed's decision applied brakes to the law, by saying, "While the public certainly has an interest in protecting its minors, the public interest is not served by the enforcement of an unconstitutional law." In 1999, the 3rd Circuit Court of Appeals upheld the injunction and struck down the law, ruling that it was too broad in using `community standards' as part of the definition of harmful materials, chronicles http://en.wikipedia.org.

"In May 2002, the Supreme Court reviewed this ruling, found the given reason insufficient and returned the case to the Circuit Court; the law remained blocked." In March 2003, the law was again struck down by the 3rd Circuit Court. The Government appealed, and it was the Supreme Court that upheld the block when deciding Ashcroft vs ACLU, on June 30, 2004. "The court referred the case back to the district court for further investigation and action," and mention was made of the use of filters as an alternative form of protection.

John Ashcroft was the Attorney General then. The current incumbent is Alberto R. Gonzales, "sworn in as the nation's 80th Attorney General on February 3, 2005," as www.usdoj.gov informs. The case in question is now ACLU vs Gonzales.

Quest for queries

Catch up with the 32-page document of immediate interest, available on SiliconValley.com, which begins with a letter dated August 25, 2005, from Joel McElvain, Trial Attorney. It is addressed to Google, Inc, and has a subpoena enclosed.

For starters, "A subpoena is an order, issued by a court of law, which requires you to appear in person at a certain place on a certain date and time," as www.floridabar.org explains. "Should you fail to personally appear as the subpoena orders, and you are not excused, you may be found in contempt of court, meaning that you have disobeyed the court. The court may punish you."

McElvain's letter spoke of subpoena duces tecum (Latin for `bring with you under penalty of punishment'), seeking "the production of documents identifying all queries conducted on your company's search engine, and all URLs identified through such queries." The letter sought a discussion with Google on "methods by which a random sample of URLs and queries could be drawn" from the company's database.

The summons demanded information in "an ASCII file on electronic media, such as a CD-ROM disk or a CD/DVD disk." It defined `query' or `queries' thus: "A text string, such as a word, collection of words or other symbols, that is entered into your company's search engine for the purpose of retrieving URLs or lists of URLs, but does not mean any additional information that may be associated with such a text string that would identify the person who entered the text string into the search engine, or the computer from which the text string was entered."

URL means uniform resource locator, or "an Internet address identifying a particular site contained in the Web," explained the summons. And these were to be URLs available on the company's search engine as of July 31, 2005. The time range for queries was June 1 to July 31, 2005.

Google's objections

Google's response is dated October 10, 2005; it's a five-page letter signed by Ashok Ramani, Commercial Litigation Counsel. It refers to the negotiations and discussions with the Department's statistical consultant Professor Philip Stark. "He proposed that Google provide him with an upper bound of the number of stored URLs on each server and the total number of search queries run on the relevant day," writes Ramani. "Professor Stark would return with one million random numbers for the URLs and the search queries, and Google would produce whatever URLs and search queries corresponded with each set of numbers."

Among the many objections that Google raised was this - that the instruction was "overbroad, unduly burdensome, vague, and intended to harass." Google conceded, however, to respond "to the production requests based on a reasonably diligent inquiry." One other objection was that the Government sought information "not relevant to a claim or defence of the underlying lawsuit," which as you remember is about COPA.

Apparently, the Government wants to use "the one million URLs requested from Google to create a sample Web against which to test various filtering programs for their effectiveness." Google objected to the Government's "view of Google's highly proprietary search database - the primary reason for the company's success - as a free resource" to formulate its own defence. "This is not an appropriate use of the federal courts' subpoena power," opined Ramani.

He added that complying with the request would imply that Google viewed its search database as `completely reflective' of the Web. "Google does not hold itself out in this fashion, and in fact resists that notion. It is against Google's competitive interest to be viewed as completely reflecting the Web." The Government could have obtained the information that it wanted from www.archive.org, which "actually holds itself out as reflecting the entire Web," adds the letter.

Yet another objection is on grounds of redundancy. Since the Government had already received URLs from `at least one other major search engine', Google wondered why URLs were needed from it. "Though the search engines doubtlessly have some differences in the URLs they store, what distinguishes Google from its competitors is the sophistication of Google's search engine in locating relevant results and ordering relevant results," explained the letter.

To respond to the request would be an undue burden, stated Ramani. "Google would have to spend a disproportionate amount of engineering time and resources to (i) `number' (even in rough terms) in real time the URLs contained in its search database and (ii) extract based on that initial numbering the URLs selected by Professor Stark."

The final objection reads thus: "Finally, Google objects because to comply with the request could endanger its trade secrets. Dr. Stark's involvement would require Google to disclose the approximate number of queries it receives on a given day, and some details about how it stores these queries, such as the number of servers and server distribution. This information would be highly valuable to competitors, or miscreants seeking to harm Google's business."

10,000 URLs from each of 100 data centres

McElvain's response dated December 23, 2005 spoke of narrower terms: "Google could select at random 100 of its data centres containing URLs and then select at random 10,000 URLs from each of those centres". And, the letter asked for "copies of the text of each search string entered onto Google's search engine over a one-week period".

One learns from the letter that after the service of objections, Ramani had called McElvain to inform that "Google would decline to comply with the subpoena." This position got reaffirmed after a bit of yo-yo of discussions. "I remain hopeful that Google may yet choose to comply with this subpoena," concludes McElvain, adding that in the absence of such a change of position, there could be "a motion to compel."

That did happen. So, we wrap with a 13-page file, available on www.siliconvalley.com, the motion to compel Google to comply with the subpoena. Dated January 18, it begins with a narration of the COPA story from 1998, and the Government's efforts to defend the constitutionality of the law, by factually proving that the Act would be better than filtering software "in protecting minors from exposure to harmful materials on the Internet."

It states that the Government had issued subpoenas to, and received compliance from other entities who operate search engines; "Google thus should have no difficulty in complying in the same way as its competitors have," urges the motion. "Because Google has the largest share of the Web search market, its response to the subpoena would be of value to the Government in its development of its overall sample of queries." It offers, "The Government is willing to compensate Google for its reasonable expenses in complying," and producing `a multi-stage random sample of one million URLs'.

Google has so far resisted the subpoena, but Yahoo!, the Microsoft Network and America Online have complied to a limited extent, reports Tacoma News Tribune in a story `31 minutes ago' on www.ocala.com, with the headline, `Internet subpoenas are cause for worry'.

A mail from Google on January 24 has this comment of Nicole Wong, the company's associate general counsel: "Google is not a party to this lawsuit and their demand for information overreaches. We had lengthy discussions with them to try to resolve this, but were not able to and we intend to resist their motion vigorously."

Keep looking, therefore, for what happens in this case of searching through searches.

Cases@TheHindu.co.in

More Stories on : Internet | Case Sensitive

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
Within reach

Guess who's placing the orders

Play, Simply Play

When the Govt comes calling...

The fun way to learn

A joint effort

Offline Web pages

Clearing msconfig startup

`Built for scalability'

Quiz

In sync with asynchronous transfer