That John Martin

Tom Anderson reports on some of the new panel quality tools from Peanut Labs and MarketTools and lays down the gauntlet on unreasonably long questionnaires.

Secondly, with professional respondents gone, who will take the 45 minute surveys some researchers like to field? First of all let me say I think anyone who tries to field a 45 minute survey, only to reduce the data later with factor analysis, should be ashamed to call themselves a researcher. If you know what you are doing you should be able to design a good 15 minute segmentation survey. However, some consultants and end clients seem to think volume of data is more impressive than quality of data. This may be one of the best possible outcomes from this new technology, that no one will ever again be able to field a 45+ minute survey.

Good discussion thread at the bottom of the post, too.

We’re launched. Finally. Say hello to Crowd Science.

It’s been a huge week, equal parts fun and stress, but we did it.

You can read about the company and our new product, Crowd Science Demographics, over on our website or go and register for a beta account.

I’m incredibly proud of what we’ve built and excited about what lies ahead. Onward!

Ricky Van Veen tumbles a Charlie Rose interview with Ted Turner:

I’ll say it again — old rich men with something to say are one of [my] favorite things in the world. They have the money to allow them to talk about whatever/whoever they want and not enough time left on earth to hold anything back.

I just discovered Datawocky, a blog by Anand Rajaraman, cofounder of Kosmix. Anand has an interesting post up on some attempts at the Netflix prize made by some of his students at Stanford. The most successful strategy, at least among this group, was to make use of a second, independent dataset: IMDB.

But the bigger point is, adding more, independent data usually beats out designing ever-better algorithms to analyze an existing data set. I’m often suprised that many people in the business, and even in academia, don’t realize this.

PopSample is growing. We have a great core team that hails from the halls of Apple, Autodesk, comScore, and Cisco, but we need more help.

We’re currently looking for a Software Engineer in Silicon Valley, and an Account Coordinator in Toronto.

If you think you fit the bill, or know of someone that might, give us a shout at jobs@popsample.com. We offer competitive compensation, flexible hours, expensive office chairs, big LCD monitors, and an operating system agnostic workplace.

Paul has shamed me into reviving my blog. And we have a startup to talk about. So let’s see how we go.

I’d love a snap-to-grid option on the OS X dashboard. It bothers me that my three clocks aren’t lined up perfectly, and that the space between my CPU and memory monitors is slightly larger than the space between my unit converter and my calculator.

Sometimes it’s the little things.

John Paczkowski writes the greatest headlines in tech journalism. Today is his last day at Good Morning Silicon Valley. He’ll be missed, but he leaves it in the capable hands of Dr. Murrell.

I’ve been using the FeedDemon 2 beta for a while now and overall I’m really happy with it. The new templates are excellent, as are the improvements to feed management, like hierarchical feed preferences and persistent per feed styles. I do have a couple of gripes though.

The new tree control is a little annoying - only a single branch can be opened at a time, so jumping around looking for something is a hassle. I’m also not quite sure how to manage the news bins. It seems difficult to access and manage individual items.

The other thing that just doesn’t work for me is the attention report. The basic algorithm that Nick uses looks pretty good, and I can see some “worthy” feeds bubbling up, but there is far too much noise from the high traffic feeds.

Shouldn’t the calculated attention value be divided by the number of items in the feed? That way a feed wouldn’t be able to improve it’s attention ranking by sheer brute force. I guess that’s just giving you a different number, and a different type of ranking, but it feels like it would be more valuable.

I’m talking, of course, about the Earthlink ad that gets played across the Gillmor pod properties. I have no real problem with it - actually, that’s not entirely true. I agree wholeheartedly with Dave Slusher, and that condescending tone of voice used by the actors goes down about as smoothly as a bowl of lego. But, it’s a new medium, and they’re giving it a shot, and it’s start, etc etc. I just think that it’s a wasted opportunity. Surely there is enough meat in those podcasts to put together something like The Deck.

The Deck is a micro advertising network featuring just four sites, including A List Apart, 37signals, Daring Fireball and Coudal Partners. They say that they won’t run an ad for a product that they haven’t paid for and/or used themselves. So the ads are highly targeted. And not just in a keywords point-and-shoot kind of way, but in an attention kind of way.

PodShow aside, I could imagine a network consisting the Gillmor podcasts, plus Venture Voice, VentureWeek, and maybe even Om and Niall. Surely the value derived from such a construct on both sides of the transaction is worth the extra effort.

This question about using dial-up security as a security measure has to be the crappiest Slashdot post ever. Despite the generation gap, Slashdot is better than that.

I know I’m late to the party on this one, but Audible’s redesign sucks.

Orson Scott Card follows up his insightful piece about gay marriage with a thorough dismantling of evolution. (Note sarcasm in previous sentence.) As always, The Panda’s Thumb has the medicine.

This link has been both Dugg and Scobelized, so for the remaining thirteen people on the planet that haven’t seen it, check out Tripod performing Make You Happy Tonight. Tripod used to do these Song-in-an-hour bits on the old morning show on Triple J that were classic. (It looks like a bunch of them are available under 2004 over here.)

Zeldman is hilarious and spot in his latest piece titled Web 3.0.

If Steven created the site with Perl and CGI and used tables for layout, this is the story of a boy who made a website for his own amusement, perhaps gaining social points in the process. He might even contribute to a SXSW Interactive panel.

But if Steven used AJAX and Ruby on Rails, Yahoo will pay millions and Tim O’Reilly will beg him to keynote.

Despite being on the other side of the planet, the Sydney Morning Herald is one of my daily reads. They launched a new design recently which I really like - it’s definitely one of the better meatspace newspaper websites out there. But they drop the ball by using a meta-refresh of five minutes. It drives me crazy. I can’t imagine advertisers would be happy with it either.

If their RSS feeds weren’t so skeletal I could do away with the site altogether, but sadly that’s not the case.

I’ll have to hunt for a GreaseMonkey solution to the problem…

Jeremy points to a report at Search Engine Watch that compares the spam filters in Gmail, Yahoo! Mail and SpamCop. I am mostly an Outlook + SpamBayes guy, and have only recently started using Gmail in anger, and I gotta say, the filter performance is just not comparable. Given a clean archive and a few hundred pieces of spam, SpamBayes is near perfect, whereas Gmail is running at about 2 from 3, similar to the findings in the report.

It’s big problem seems to be on the stock/investing type of spam. If it weren’t for Gmail’s excellent keyboard shortcuts it would be a deal breaker.