DEBRIS.COMgood for a laugh, or possibly an aneurysm

If you know anybody who owns an iPhone, you’ve probably seen this at the bottom of their emails:

Advertising via the email “signature” is nothing new; it was probably invented by Hotmail in 1996, and is used widely today. Lots of webmail and message-platform vendors promote their products this way.

With that in mind, check out the signature on the email I just received from my little sister, a Gmail user:

Is that not brilliant? Apple bought the signature line from Gmail to promote the iPhone!

That single line of text beautifully captures the sense of lust the iPhone inspires.

I was in awe of this little line of text. Was Ridley Scott in on this?

But no, it’s not a real Apple campaign. Not yet, anyway. My sister made it up, in sarcastic protest to all her hipster friends who can afford iPhones.

It’s too good not to be real, though, so (with my sister’s blessing) I’m offering it to Apple. Give TBWA the afternoon off. We got your iPhone campaign right here.

But please send an iPhone to my sister.

This blog has historically had a relatively high PageRank value, a benefit I attributed to the fact that the software that drives it was created to meet what everybody understood to be best SEO practices in 2000-2001 — for example, each article has a short and eternal URL (mouse over any article title to see its permalink), a unique and descriptive page title, and so on. I’m not sure whether SEO was an industry in 2001, but certainly it was a skill of webmasters and web engineers to build indexable websites, and SEO guidelines were formost in my mind as I architected this website.

My site’s PageRank changed in early 2006. I was reading about SEO, ironically, and I learned that sites answering to multiple URLs risk getting penalized in Google’s index due to Google’s new (early-2006) duplicate content filtering algorithm. Although I’ve only ever published my own blog URL without the ‘www’ hostname, the server hosting this site responded to both forms of the URL: both debris.com and www.debris.com. Every page on the website was available at both addresses.

I followed Matt Cutts’ advice in early 2006 to use a 301 redirect to automagically forward everyone who was trying to get to www.debris.com (or any page there) to the equivalent page at debris.com. Shortly after that, my PageRank disappeared — the Firefox PageRank plug-in reported it as “n/a;” multiple-datacenter survey tools mostly reported it as blank, with a couple datacenters inexplicably still showing 4-5.

At first I thought this was a temporary condition resulting from the 301 redirection. I waited a month or two to see if a subsequent PageRank push would reveal the effect I’d intended, that the previous PageRank of 5 for www.debris.com and 4 for debris.com would consolidate to a solid 5 or maybe even a 6 for the canonical domain. Alas, this never happened.

In May, 2006 I created an XML sitemap to seed Google’s crawlers with the newest content from this site, thinking that perhaps this would cement in the crawlers’ collective mind that, despite inbound links from 3rd party sites to formerly duplicate-content URLs, everything was happy and canonical and uniquely addressed on the server, using the technique advocated by Google’s own webspam master, Matt Cutts. The sitemap reduced traffic from the crawlers — reflecting my on-again, off-again publishing style — but unfortunately didn’t correct the site’s missing PageRank.

By early 2007, I had waited more than six months. I wondered if my site had been inadvertently penalized, for its PageRank never came back.

So, I did the thing I should have done last March, as soon as my PageRank disappeared — I filed a reinclusion request. Honestly I didn’t really need to be “reincluded,” as my site was still in the Google index, and did turn up in searches for which my site is authoritative. But it was the only trigger I had left to pull.

A week or two later I took an additional step, inspired by a blog post by Rogers Cadenhead. He described that his blog software showed his entries at multiple addresses — the home page, the category pages, the tag pages, and on each item’s permalink page. This is how debris.com works as well. Rogers describes this as a “huge mistake.” I personally disagree; this seems to me to be a service to the user. And I have to point out that Matt Cutts’ blog shows full posts at multiple URLs, all of which turn up in Google’s index.

But as Rogers points out, the best case outcome of this seems to be that Google shows an italicized message at the bottom of results pages indicating that duplicate results have been omitted. I followed his lead and sequestered all but the permalink URL from Google’s crawlers, via “nofollow” attributes on category-page HREFs and robots.txt entries for those pages. Also I dropped the category pages from the XML sitemap.

I’m not sure these category-page changes made any difference, but the reinclusion request certainly did; this site is now showing an all-time high PageRank of 6.

And it only took 30 days. Or 13 months, depending on how you count it.

I haven’t wasted this much time since I saw the earth sandwich video from Ze Frank — and spent a couple hours over the next few days watching every previous episode of The Show.

My friend Pete, former recording engineer to the stars (and to my band too) sent me a link to an episode of the death-metal comedy cartoon Metalocalypse. It’s been on the Cartoon Network since August, so this may not be news to you unless you also live in a cave. Fortunately, either way you can see most of the episodes online.

Update 2008-01-07: The bleenks.com site appears to have gone away. Sorry, no more episode archive. But see http://www.diefordethklok.org/ for plot summaries and stills.

Some of the characters are all but unintelligible, but they’re funny anyway.

Update: whoa, things get dark midway through the first season. Or maybe I just have a limited appreciation for severed-limb humor.

Firefox had been giving me the spinning beach ball of doom for weeks, every time I downloaded a file. Whether a 5k GIF or 50MB FLAC, the browser hung for 30-40 seconds before finally popping the Downloads window open.

You’d think it would take a pretty serious computing challenge to redline two 2GHz G5 CPUs, but no — just try downloading some album art from Google Images. Apparently there is some kind of CPU-cooking fractal math in the progress-bar code in the Downloads window. Or maybe Firefox is calling into the SETI@home project on the sly, scanning a few parsecs of some remote slice of the universe for intelligible radio signals before ultimately beginning the admittedly pedestrian task of saving a few hundred bytes of image data to my hard drive.

Thanks go to Karl Pietri for suggesting I click the “clean up” button in the Downloads window. Erasing the past year’s worth of download history cut the download startup time from 30+ seconds to a reasonable .05 seconds.

More info, and a screenshot, can be found here: Firefox Download Clean Up

After spending a year considering how easy it would be to add tags to this journal software, I spent about 90 minutes actually adding tags to this journal software. See taglinks below.

It’s a pretty basic implementation; the editing interface allows me to add any number of keywords to each journal item. At the moment, the tags are rendered as links to Technorati, allowing readers to discover other blogs covering the same topic. In a couple weeks, or, honestly, about 14 months, I’ll add a local tag-search feature that allows easy browsing of all items with a particular tag (or tags).

In terms of functionality, there’s a bit of a collision with the existing categorization scheme I’ve been using for years. But this is a classic case of ontology vs. folksonomy — of top-down, prescriptive categorization versus free-form, infinite-number-of-buckets tagging. Smart money says the latter is superior for organizing large corpora.

My old categorization scheme isn’t exclusive — every journal item could appear in multiple categories. Categories are tags, really, although in my case not very interesting or descriptive ones (e.g. auto, misc, travel). Who would go to Technorati, search for blogs about “auto,” and click through to debris.com? According to my server logs, exactly nobody, despite my creative and occasionally insightful writing about the automotive industry.

So I can imagine one day replacing my categorization system, which by the way took a couple days’ worth of coding, what with the nested and hierarchical display and fancy editing interface that I’ve used all of about seven times. Once I’ve built a browse-by-tag feature, the categories will be pretty much useless. Such is evolution… that long list of category links in the right column is basically a digital appendix. Or, maybe, a male nipple.

Why now? Why add tags when I’m not even writing that often?

A couple reasons: One, because I’d like to join the distributed community Derek Powazek spoke about at Etech. Tighter integration with Technorati is a solid first step. Two, because I’ve been doing a lot of thinking about the atomicity of information, and how to best collect and present many bits of related information in multiple contexts, and I needed more firsthand experience with tags. More on this later, but probably in another context.

This kills me… GMail’s much-discussed, much-maligned context-targeted banner ads don’t have a lot of message text to target within my GMail account’s typically empty “Spam” mailbox, so all the ads are for Spam recipes!

Spam sushi, anyone?

One thought was unmistakably on everyone’s mind as they strolled around AT&T’s reception at Web 2.0, with a premium drink in one hand and a plate full of high-concept snacks in the other, like Gazpacho Shrimp Shooters, or Dungeoness Crab Tataki on Nori Chips, or Ahi Tuna Tartare with Ginger and Lime, or Wasabi Mashed Potatoes with Sesame Grilled Salmon, Crispy Leeks, and Black Caviar (served in a martini glass), while Zigaboo Modeliste’s Aahkesstra jammed onstage and the entertainment budget had room not only for a photographer but a flash guy to follow him around.

Robert Kaye of MusicBrainz, eyeing the spread with some amazement, summed it up: “is it just me… or is the Internet boom back?”

The energy at the conference is palpable. New deals are happening every day. Products are getting better. The browser war is being un-lost. Companies are making money. Stock options are — dare I think it?? — emerging from a five-year bath.

I’m not counting my money yet, but I am eating the hell out of the five kinds of Dim Sum at the buffet. And I’m looking forward to the next 12 months.

Jason Fried of 37signals gave an eye-opening presentation at Web 2.0 today: less is more! This was the best, most directly useful presentation I’ve heard so far.

Less is a competitive advantage. Trying to one-up everyone else is a “Cold War” mentality. Instead, try to one-down the competition. Beat them with less.

There are five things software developers need less of:

1. Less money
2. Less people
3. Less time
4. Less abstractions
5. Less software

Less money
Taking on a big investment used to be the way businesses would get started. But all that investment gets you is debt. Debt doesn’t make you build better products. Debt doesn’t make your customers happy.

Money used to buy hardware, but now hardware is cheap. It used to buy software, but now software is cheap. Money could never buy time or passion, and passion is the one thing you need in quantity.

The one thing money could buy is salaries, meaning people. But maybe you don’t need those either…

Less people
You need only three: a designer, a programmer, and a “sweeper” to do everything else (including marketing).

Don’t scale the team up to fit the vision or the feature set. Instead, scale the vision down to fit the size of the team.

Less time
Having less time is a good thing, a huge competitive advantage. Most of the time spent in most companies is wasted: meetings, overbuilding, overanalysing. Having less time forces you to spend it smarter. Spend it only on the important things — building the software.

Less abstractions
Just build things. don’t write about them. Don’t draw pictures. And whatever you do, never never never make a “functional spec.” This is the biggest abstraction of all, and there is nothing functional about it. A long text document has nothing to do with a great piece of software.

Functional specs are political documents. They’re “yes documents” — every feature is a bullet point; it’s easy for people to say “yes, add that.” Worst of all, they create the illusion of agreement. You spend three months hammering out the functional spec, but three months later you find that everybody disagrees about what’s in the functional spec, and the finished product (if and when it is ever finished) looks nothing like the functional spec anyway.

Instead, just build the product. Build the UI first. Then wrap that with the technology. Have the technology fill in the back end. Work on the customer experience first; that’s the most important thing. Have the interface screens be the functional spec.

Building software is an iterative process. Iterate over the built product, not the abstractions.

Less software
Build less (complex) software. You’ll need less support… less help text… less documentation… less sales effort…

There are a million simple problems that need to be solved. Solve those first — but nail them. You won’t succeed at solving the complex problems anyway.

More of what?
The only thing you want more of is constraints. Less time, less money, fewer people, these are all constraints. They’ll squeeze you. But the result will be a higher-quality product.

Update: Jason’s notes are here. And Robert Kaye’s article is here.

Continuing Web 2.0 coverage…

I’m fascinated by Technorati, the original blog search engine, for a couple of reasons:

• The technical challenge they’ve taken on is astounding — they’re facing a Google-class scaling problem. (See statistics below.)
• and perhaps due to the first point, the performance of the system has, in my experience, been poor. Searches take 30-60 seconds. I’ve never used the site because I don’t have the patience for it.
• Even so, the recent public abandonment/shaming of Technorati by brand-name bloggers strikes me as somewhat unfair.
• New entries into Technorati’s blog-search market, e.g. Google’s blogsearch site, introduce a huge challenge to Technorati’s long-term success, especially if it’s true that Technorati is not executing well.

Technorati’s archive (a retrospective search architecture, btw) is tiered into 3 layers based on recency. They’ve made the assumption that the most recent hits for a search are more interesting than older hits, and of course in many cases that’s true. But the data i/o requirements, of simultaneously pushing new data into the front of each archive while expiring it out the other side (and from there into the next-most-recent archive) would seem to be unmanageable.

I learned this week that the write to read ratio within technorati’s database is 5:1. That presents a horrible scaling problem, because writes are slow and can’t be easily parallelized. Compare to Tribe.net’s write to read ratio: 1:99. Tribe’s is probably an extreme example (in fact I suspect that figure needs to be qualified) but in any case if your write rate is bigger than your read rate, you’ll be hurting under load. E.g., searches will take 30-60 seconds.

Part of Technorati’s problem is volume. Check out these “state of the blogosphere” stats from Dave Sifry:

• There are now 18.9 million blogs
• 70,000 new blogs appear every day, about one per second
• 55% are still active 3 months later
• 13% of blogs are updated weekly or more

It’s a big pile of data to read, store, and index.

October 5, 2005 - SAN FRANCISCO - Today, Google announced the launch of Google Ice, an innovative new product in the beverage-cooling market. Google Ice is available exclusively at the Web 2.0 Conference in San Francisco, but will soon be rolled out within after-hours tech confererence geek lounges nationwide, except perhaps in Redmond.

“Google has already made tremendous strides in making access to information on the web a reality for users across the globe, but we’re still in the Internet’s early innings,” said Google CEO Eric Schmidt. “But at least now you can see what you’re drinking.”

“We hope this product makes quite a… splash,” added Google founder Sergey Brin. “Heh, heh, heh. Heh. Ahem.”

Added Stewart Butterfield, founder of Flickr and employee of Google rival Yahoo, “5000 Ph.D.’s and they come up with this?!”

The BBC has published a Flash (!) version of the old Infocom text adventure based on the Hitchhiker’s Guide to the Galaxy.

That was the first and only Infocom title I ever bought. I’d poked at a few pirated copies of the Zork series and maybe Planetfall, but never played seriously or with any intention of finishing. Although I did download the instruction set for winning Zork I in the fewest possible moves, because I thought that was an important thing to have handy. (Geek trivia note: as it’s largely a series of one-letter instructions and short commands, the whole thing fit easily onto a couple lines of text on the BBS screen I downloaded it from.)

So I played the H2G2 game for an hour or two… got off the planet, got the fish in my ear, and got frustrated. I don’t think I ever ran the program again. I didn’t have much patience for games, much less games that lie to the player.

But now, thanks to the BBC, I can relive all my adolescent frustration for free, online, with pictures!! — a real change of pace considering the game’s ancestry.

Now I’ve spent about 20 minutes at it, and I can’t even stop the bulldozer. Clearly I was a cleverer guy at age 17 than I am now. Note to self: all those jokes about killing brain cells in college may have, in retrospect, been less funny than originally believed.

The BBC site also contains an interview with Steve Meretzky, the programmer who worked with Douglas Adams to create the Infocom game. (The interview alludes to the great difficulty of the Babel Fish puzzle, suggesting that, 20 years ago, I downloaded the solution to that from the BBS, too. Hmph.)

This is the funniest thing I’ve seen in a long time:
Steve, Don’t Eat It!

Following is an illustration of the only way to win an Ebay auction any more: wait until the absolute last second, and bid 29¢ more than than the next guy.

This auction closed at 12:16:50, one second after my bid was received. I nearly missed the window because I decided 3 seconds before close to revise my bid. I wasted 25¢, as it turns out.

Bidding in the last moment of an auction is called “sniping” and it is annoying as hell, unless you happen to be the winner. I’ve lost numerous auctions by $1.00 or less, in the last 10 seconds of action. And I’ve won numerous auctions the same way. In fact, most everything I’ve bought on Ebay in the past year — or failed to buy — has come down to a last-moment bidding war. I’ve saved a bit of money this way, and, probably, shaved 18 months off my life due to sniping-induced stress.

ZabaSearch drives another nail in the coffin of personal privacy. Type in a name and state and get the person’s full name, birthdate, address, and phone number. It’s like a time machine, too; their data goes back at least 15 years, retaining records of every place anyone has ever lived. Sigh.

David Lazarus did some nice investigative work for the Chronicle, in a column called It’s impressive, scary to see what a Zaba search can do. The article explains an unexpected and bizarre connection between the founders/owners of Zaba and the Heaven’s Gate suicide cult.

See also Matt Haughey’s recent writeup on the issue: Strange, Troubling Privacy