Greetings. A Santa Clara County judge has moved a step closer to forcing Topix to release identifying information -- presumably mainly IP addresses -- of anonymous posters in a long-running defamation case. The implications are both complex and important.

The core of the case seems relatively straightforward. A couple was accused, indicted, tried, and acquitted of serious sexual assault charges. Along the way, even before indictments, large numbers of viciously attacking anonymous comments appeared on Topix associated with the case -- more than 25K comments in all.

The vindicated couple wants to find out who -- they apparently assume it's actually a relatively small number of individuals -- has been posting those comments, so that they can be pursued for defamation.

That's where this all morphs from straightforward to messy and complex, and quickly falls into one of those "just because you don't like it doesn't mean it isn't true" situations that I bring up so often.

Vicious defamation is presumably not a concept that most people would support. But the ability to anonymously comment is important as well -- and U.S. courts have not been terribly consistent on their rulings regarding the appropriate balance between the two.

And as other seekers of presumed online miscreants have learned, an IP address alone doesn't necessarily point you at the right person -- not given public terminals and unsecured Wi-Fi access points.

Yet as several stories on this case pointed out today, large numbers of highly offensive, factually incorrect attacking comments tend to rise to the beginning of search engine query results, in some cases swamping out any retorts or factual pages and comments -- a situation that could theoretically persist indefinitely.

Some Google critics in particular have called for the censorship of search results in situations like this.

I view such suggestions as both impractical and unwise. My basic philosophy is that in most cases the cure for "bad" information isn't censorship, but rather bringing good information to prominence.

However, the manual manipulation of natural search results ordering is not something to be done lightly, and I would argue would be best avoided.

There are possible alternatives -- albeit not trivial and noncontroversial ones.

I have in the past proposed consideration of a limited and controlled "dispute links" mechanism that would provide a means of informing searchers that top search results contain disputed information. To avoid abuse, such a capability might perhaps only be invoked under court order -- for example, to note that the couple being discussed were already found innocent of all charges.

This mechanism would definitely not be simple and a myriad number of complicated issues would come into play. Whether or not in the final analysis it would even be a practical solution is not a foregone conclusion.

But consider the parties in play in the Topix case. On one hand we have the couple declared innocent by a court. For them to be forever saddled with these sorts of attacking -- and now demonstrably false -- comments in top search results without any annotations regarding the court case's outcome seems fundamentally unfair.

On the other hand, anonymous speech needs a high (though few would argue absolute) bar of protection to avoid all manner of anti-anonymity "scavenger hunts" run amok.

So my question becomes, where could "more information" be appropriately and non-disruptively injected into the system to provide the maximal fairness for the aggrieved parties, yet also allow for the appropriate protection of anonymity?

Given that search engines are the means by which most people will find their way into such discussions, it seems at the very least prudent to explore whether there are methodologies that search engines could deploy -- whether my "dispute links" concept or something else entirely -- that could help to defuse these dilemmas without tampering with natural search results determinations nor their associated orderings.

As always, I welcome your thoughts.

--Lauren--