I agree with everything you say except: "The information gathered is largely a database of lies. "

I know with 100 percent certainty that is not true.

I've run two registration sites and in both cases, the data provided by 96 percent of local users was truthful and accurate.

We did multiple checks on the data and were surprised to find how honest our local, loyal users were.

That said, there are better ways to get to user data than blind registration.

No one knows with 100 percent certainty that what is in their registration database is true. It's all volunteered information. Perhaps you actually had reporters track down these registrants and verify. I don't know what checks you're talking about.

Still, I admit it is a poor choice of words to say "largely" a database of lies since I don't know how much is made up either.

In a post by Jay Small ages ago, he said about 2 to 3 percent of users entered obviously inaccurate information. That's just the obvious stuff. The amount of inaccurate information is inevitably, significantly more.

When a Web site asks a person for their ZAG (zip code, age, gender), there is little incentive to enter the honest info. The real incentive is to get it done ASAP.

Allowing ourselves to believe that so much of the information is real is what keeps registration around. Call me skeptical.

The main point here is: Registration is on its way out for a lot of reasons.

- Steve Outing applauded Jacksonville dumping its registration wall a while back.

- Adrian Holovaty is on record as saying registration treats readers with "disdain."

- Steve Yelvington says registration data can be misleading.

Howard, you were a big lover of registration from the start. So I think it would be interesting for you to share what caused a change of heart. (At least it seems you've had a change of heart based on your comment above.)

First, Jay Small is right about the small percentage of obvious stuff.

Second, in Ventura (several times) and in Bakersfield (once) we ran the local addresses against CASS and then list matched against the subscriber database (in Ventura, both past and current subscribers, so we had a lot of confirmed data to match against registration data).

To me, the evidence is pretty compelling that the vast, vast majority of people register with truthful information for their local news sites. So I was mainly reacting to your "database of lies" statement. I'm confident that if you grabbed any local newspaper registration database and tried to find evidence to support that statement, you couldn't do it.

As for my slightly changed position, I'm liking the idea of unencumbered articles acting as a lure for participation, which can lead to registration. I've just decided I would like to experiment with offering more value in exchange for registration.

Also, most newspapers have not yet delivered on the promise of registration. Advertisers aren't demanding it and don't get it, and to really make it work would take far more effort than most newspapers seem willing to put forth. I think there are higher priorities for sales people and developers to work on.

I'm not anti-registration, the way Outing is, just ready to move on to other experiments.

I'm willing to accept that the "vast, vast majority of people" give honest answers. Most people are generally good.

But I don't believe for a moment that only 2 to 3 percent of people make up the answers on registration forms.

Vast, vast amounts of honesty might be happening. But even slight amounts of undetected dishonesty will lead to bad business decisions based on those fibs.

A database that is even only partly lies is entirely un-useful. If every registrant underwent the confirmation process you described, then the data would be reliable. And, as you said, there are more important demands for developers' time.

So let's agree to scrap registration for the many reasons we've discussed. As you said, advertisers don't request registration data. So why does it still exist?

If we're trying to create e-mail marketing programs, there are better ways to do that. If we're trying to get market research, there are better ways.

Instead of nixing registration altogether, why not give users a better reason to deluge their information?

Better yet, why not use that information to give groups of users content in how they interact or how information is presented to them?

It's obvious the old-school methods need to change. But lets use it as an opportunity for innovation!

(BTW: I work for Scripps. Registration is still being used, but just for folks that want to post comments or receive information.)