Archive for the 'Web' Category

I have a tradition every Sunday morning: no matter how much work I need to catch up on, I always start the day with the New York Times crossword puzzle. I almost always finish it, too. (Thanks, Brandeis University Department of English and American Literature. It seems a liberal arts education has a purpose after all.)

Now and then — very rarely, I assure you — I find myself stumped. In such cases, I swallow my pride and run a web search to see if the internets hold the elusive answer. That happened today: 85 Across asked for “Carnaval sur la plage” artist and I was at a loss. I’d never heard of the painting, so I had no idea of who the painter might have been.

I went over to Google, ran the search, and there was my answer within the snippets of the second and fourth results on the SERP. I didn’t even have to click through to find out that it was that famed Belgian painter, immortalized by They Might Be Giants.

But here’s the really interesting thing: above the regular results, Google had inserted links to a couple of “searches related to carnaval sur la plage artist”:

Erm… “Ron Guidry nickname?” “Mario Puzo sequel?” Those searches are related to this painting?

I don’t see any pitchers or gangsters in there. So how can those searches be related to the search I ran?

Well, if I cheated more, I might have noticed right away, because if you go to today’s puzzle, you’ll see that the clues for 94 Down and 21 Across are “Nickname for Ron Guidry” and “Mario Puzo sequel.” They’re related by the fact that they’re in the same puzzle.

I tried running searches based on other clues in the puzzle to see how much of the puzzle the engine had determined to have this relationship:

• 4 Down: “My Fair Lady” composer returns related searches for Mario Puzo and Bad Moon Rising — the clue for 49 Down is “__ bad moon rising” – 1969 song lyric
• 21 Across: Mario Puzo sequel returns related searches for Mario Puzo books and balalaikas — the clue for 22 Across is relatives of balalaikas
• 22 Across: relatives of balalaikas returns no related searches
• 54 Across: Chinese dynasty before the Shang and 103 Across: harmonica-like instrument both return related searches, but without any connection to the crossword
• 42 Across: Chief city of Moravia and 48 Across: “Revelations” choreographer return no related searches

So apparently, Google hasn’t indexed the content of the puzzle and related every clue to it. Rather, it looks like it has detected a trend: someone searches on some of the clues, someone else searches on the same clues, someone else searches on some of those and a few others, and this all happens within a few hours, so Google determines that the searches are related to each other based on that, so when I come in and search on one of the clues, Google offers up some of the other searches that were run today by other people who ran that same search.

Interestingly, I’m finding that a search on the exact clue for 94 Down, Nickname for Ron Guidry, isn’t returning any related searches. That only comes up when I search on Ron Guidry nickname, which is the more likely search syntax, considering the absence of the stop word. (And by the way, the related searches now coming up for that one are now Mario Puzo sequel and state flowers — the latter being the theme of this week’s puzzle.)

Apparently, this is something new. A search for [google related searches crossword] didn’t give me any useful information, but I did locate a post from just over a week ago on the Official Google Blog: Two new improvements to Google results pages. Here’s some of the text from the post:

Starting today, we’re deploying a new technology that can better understand associations and concepts related to your search, and one of its first applications lets us offer you even more useful related searches (the terms found at the bottom, and sometimes at the top, of the search results page).

For example, if you search for [principles of physics], our algorithms understand that “angular momentum,” “special relativity,” “big bang” and “quantum mechanic” are related terms that could help you find what you need.

…

We are now able to target more queries, more languages, and make our suggestions more relevant to what you actually need to know. Additionally, we’re now offering refinements for longer queries — something that’s usually a challenging task. You’ll be able to see our new related searches starting today in 37 languages all around the world.

There’s nothing in there about tracking groups of searches over a short period of time and relating them to each other, but that seems to be what’s going on here.

Two posts in a day! (And they’re my first two posts all month.)

I was watching CNN this morning, and right after Howie Kurtz spanked Sports Illustrated for failing to be seriously journalistic about its bikini photography, they cut to an advertisement for the Florida Keys — perfectly normal tourism stuff. At the end of the advert, the announcer suggested checking out the tourism board’s website at “FLA Keys.com,” that is to say, what I heard was “eff ell ay keys dot com,” which one would expect would be written out as “flakeys.com.” Of course, one would probably read such a domain name as “flakeys dot com,” which is maybe not the image the Keys want to promote. That would explain why the domain name that appeared at the bottom of my screen was “fla-keys.com”. But the announcer had made no mention of the hyphen.

Naturally, that got me wondering. I was online at the time, and I might not have been looking at the eye of hell while the ad was running, so I would have thought the site was located at flakeys.com. What would happen if I tried to navigate to that address?

Sure enough, a request for flakeys.com redirects (via a 302 instead of a 301 — stupid IIS) to fla-keys.com, so those flakey Floridians at least have that covered.

But couldn’t they have gotten a domain name that would both allow them to easily say and spell the address the same way (that is, without the much-despised hyphen) and not require a redirect? That question led to some quick exploration:

• floridakeys.com, which looks like it could be the official site of the local tourism board, complete with a “Florida Keys” logo, is owned by a company called Cooke Communications, which apparently publishes some local newspapers and magazines.
• thefloridakeys.com is the home of “Best of the Florida Keys,” which is kind of archaic looking. Dig the animated gif of the hurricane on the home page!
• flkeys.com is a local realtor.

And what of the hyphenated variations?

• florida-keys.com is apparently for sale and is currently hosting one of those parked made-for-AdSense “search” pages.
• the-florida-keys.com is registered to some guy in New Jersey, but nothing is published there.
• fl-keys.com contains a meta refresh that sends the user to fl-keys.com/floridakeys.htm, which then fires a script that redirects to fl-keys.com/search/load.php, which in turn fires another JS redirect to fl-keys.com/search/re.htm, where you’re hit with yet another script that send you off to fl-keys.com/search/index.htm (wheeeeeeeeee!), which is another sponsored search page.

It seems that unless they’re willing to buy what could be a pretty costly domain name, the folks at the official tourism council got beaten to the punch. Flakeys indeed.

I have a client who is a Minneapolis interior designer. (Yes, that’s a plug for them. As if you wouldn’t do the same thing.) Earlier today, I was checking out the progress of their rather new site using Google’s webmaster tools, and was pleased to see them starting to move up, and even get some traffic, for searches like [interior designers minneapolis] and [interior design firm minneapolis]. They’ve got a ways to go, but they’re definitely moving in the right direction.

Google also listed them at #4 for [interior designersminneapolis] (note the missing space). That’s a little troubling. They rank higher for the search with the typo than without. I took that to mean that either someone’s linking to them with the typo in the anchor text, or they’ve got the typo somewhere on the site. I couldn’t find either one, though.

I was further confused by Google’s attempt to correct me:

Google wants to know if I meant “interior designers minneapolis” and if I meant to search for “interior designers minneapolis.” Wha? Obviously, if I meant it, I meant to search for it. What else would I be doing at a search engine?

And the two searches being offered to me are indeed identical. The first one links to
http://www.google.com/search?hl=en&sa=X&oi=spell&resnum=0&ct=result&cd=1&q=interior+designers+minneapolis&spell=1 and the second to
http://www.google.com/search?hl=en&sa=X&oi=spell&resnum=1&ct=result&cd=1&q=interior+designers+minneapolis&spell=1

So what’s the point?

As it turns out, this is not Google’s doing. I run a Greasemonkey script called GoogleMonkeyR. I use it to number G’s results. If I disable it and run the search again, I only get the “Did you mean: interior designers minneapolis” line. I have no idea why it’s adding that extra “Did you mean to search for” bit.

So, problem solved, for the most part. I’ll write to the script’s author about the bug, and I’m no longer freaked out about tripping over a bug in the all-knowing Big G.

It’s still a bit odd that my client is ranking for that search containing a typo, though.

Anyway, it gave me a decent excuse to link to them.

This site is probably going to be almost as busy as the NYFT today. Of course, since little of that traffic is likely to be coming from the Amarillo area, it’s not likely to make the owner rich.

This isn’t the Joe the Plumber who was referred to some 25 times during last night’s debate. It’s just the lucky domain name owner of the day.

A friend of mine wrote to me today asking if I was familiar with Guy Maddin‘s film My Winnipeg. He’d seen it on an Air Canada flight, and as he put it, “Just the fact that Air Canada would put such a weird film on their movie list makes me want to immigrate there.” And he’s an immigration attorney. I think I’ve seen three of Maddin’s films — Dracula: Pages from a Virgin’s Diary, Cowards Bend the Knee (or The Blue Hands) and The Saddest Music in the World, and it would be fair to say I’m a fan.

My friend told me that the film is available at Netflix, so I popped over there. While I still consider myself a cinephile (cineaste, movie buff, whatever you want to call it), the fact is that I don’t see nearly as many films as I used to. Consequently, I don’t log in to Netflix that often. Before I checked on the Maddin film, I figured I’d look over some of their suggestions for me, based on my ratings — the famed “Movies You’ll ♥” page. I had thought that the recommendations were based on ratings from friends in addition to my own ratings, but apparently it’s all me.

Some of the recommendations and the reasons behind them seemed sensible enough. Samurai Rebellion was recommended because I liked Seven Samurai, Yojimbo, and The Hidden Fortress. No problem there.

Ace in the Hole (also known as The Big Carnival) was recommended because of my positive ratings for 2001: A Space Odyssey, Annie Hall and Blue Velvet. I’ve seen Ace in the Hole, and I recall liking it, but it was a long time ago. It was one of the films we studied in my very first film studies class. The class was an English department elective in my high school, and basically involved memorizing the glossary at the back of Louis Giannetti‘s Masters of the American Cinema (a book I ended up buying for a film class in grad school some six years later), and watching films while the teacher, the inimitable Miss Lily Achille, would shout out in her Eleanor Roosevelt/Margaret Dumont voice, “The framing! Look at the framing!”

Where was I? Oh, yes — Ace in the Hole. I appreciated being reminded of it, and may in fact add it to my queue. But I have no idea what it has to do with those three other films.

At the bottom of the recommendation page was a big surprise, with a few surprises attached. Under the heading Sports and Fitness Suggestions were the following three titles:

Now, just for starters, it doesn’t make a lot of sense to recommend anything to me from the Sports and Fitness category, unless you happen to be my doctor. I am, after all, a fat, lazy slob. This just isn’t my thing. But look at the films I liked, which led Netflix, in its infinite wisdom, to offer up these suggestions.

Lost in Transition is recommended because I enjoyed Lost in Translation

What do these two films have in common, apart from the similarity in their titles? Well, here’s the blurb for the recommended film:

Take an unforgettable journey in this action-packed film shot in multiple locations around the world, documenting snowboarding’s top athletes at the peak of performance. Facing off on some of the planet’s steepest slopes, featured boarders include Mark Landvik, Frederik Kalbermatten, Dave Downing, Eero Ettala, John and Eric Jackson, Markku Koski, Jeremy Jones, Mads Jonsson, Dave Downing, Benji Ritchie and more.

It’s just the similarity in the titles.

Crunch: Candlelight Yoga is recommended because I enjoyed Lilya 4-Ever

If you’ve never seen Lilya 4-Ever, you really should, but only if you think you can handle it. It’s very, very unpleasant. For someone in my condition, yoga (by candlelight or not) would also be very, very unpleasant, but I don’t think that’s the connection. I recall that there was a yoga show on PBS when I was a kid, called Lilias! Yoga and You, and the show’s host, Lilias Folan, is still at it. Maybe she’s on the disc, and Netflix made a Lilias/Lilya connection… Nope, that’s not it. The host of the yoga video is one Sara Ivanhoe (yeah, right). Lilya does find herself in a number of uncomfortable positions over the course of the film, but that just couldn’t be the connection. I’m at a loss on this one.

Finally, the most confusing recommendation of them all:

Billy Blanks: Bootcamp: Ultimate Bootcamp is recommended because I enjoyed The Short Films of David Lynch

What? Does Lynch do Tae Bo? Do the Ultimate Bootcamp workouts somehow involve running sores, or fear of sexuality and biology? Maybe during breaks they bring on a couple who bark like dogs while abusing their tuxedo t-shirted child.

With all due respect to Amy Poehler, Rosemary Watson is the Hillary Clinton impersonator.

Check out her site or her YouTube page, That Hillary Show.

I wrote some thoughts last month about how we choose to communicate URLs to people — what’s essential to getting the message across, what’s not, and what may be changing with time.

Tonight, while waiting for Gary Indiana to finally give the Clinton campaign the coup de grâce (Updated: OK, not quite, but really close), I saw a spot for a certain flower company. It is almost Mother’s Day, after all. “Which flower company was it, qwerty?” I hear you ask.

It was this one. And there’s the URL, right in the logo (assuming you ignore the flower growing out of the dot). You can read it out: “one dash eight hundred dash flowers dot com.” But you watch the commercial and there’s the founder of the company and his daughter (shades of Frank and Jim Perdue?) referring to it as “one eight hundred flowers dot com.” No dashes. Why are there no dashes? In my line of work, I advise people not to get domains with dashes, and a big reason for that is because it’s harder to communicate. But in this case, the dashes are there in the branding. Are they assuming people will think of the logo and know that when they think of “one eight hundred flowers dot com” they should remember that there are dashes in there? No, that can’t be it.

This is from the site’s About Us page:

Jim McCann, founder and CEO of 1-800-FLOWERS.COM, INC., opened his first retail store in 1976 and successfully built his own chain of 14 flower shops in the New York metropolitan area. In 1986, he acquired the 1-800-FLOWERS phone number and continued to grow his business under the 1-800-FLOWERS® name. His immediate focus was to create a reliable brand name built on trust, and over the next few years he achieved this through his understanding of his customer base and market. The next logical step was expansion, and McCann successfully expanded his business into other retail access channels-going online in 1992 and opening a web site in 1995. Today, 1-800-FLOWERS.COM® has a well-known web site (www.1800flowers.com), and maintains strategic online relationships with a number of online services, including America Online, Microsoft Network (MSN), and Yahoo!.

So whatever they were called from 1976 to 1986 doesn’t matter much. In ’86 they branded the company based on that phone number, 1-800-flowers.

The way we communicate phone numbers has certainly changed over the years. I believe everyone in the US has been on a system that requires them to dial 1 at the beginning of a long-distance call since the 1970s. We no longer say “area code” before announcing an area code. And it’s probably been thirty or forty years since one would identify an exchange (the first three digits after the area code) with two letters and a number, which was itself a streamlined version of naming an exchange for a place, followed by a number. Wikipedia notes that the Ricardo’s phone number on I Love Lucy was MUrray Hill 5-9975, which would later be referred to as MU5-9975, and later 685-9975. I remember this commercial from my misspent youth (spent in the glow of the eye of hell):

There was another hotel with ads that included a phone number with a full on “Murray Hill” exchange, but I couldn’t find that one.

But I digress. The point is, we know how to say a phone number. Even with all the changes that have taken place over time, we’ve never had to actually say “dash”. The hyphens were in there just to make the printed version of the number easier to parse. So if we know that 1-800-flowers can be read “one eight hundred flowers,” the people marketing the web site feel that we can look at that logo with its hyphens and hear the domain name spoken as the phone number with a “dot com” tacked on the end, and not have our heads explode.

Beyond that point, here’s the real point, courtesy of my pal the WebBug:

If I request http://www.1800flowers.com, the server returns the following:

HTTP/1.1 302 Object moved
Location: http://ww21.1800flowers.com/

A “temporary” redirect, but I end up at a domain without any dashes.

If I request http://www.1-800-flowers.com, the server returns the following:

HTTP/1.1 200 OK
Date: Wed, 07 May 2008 05:31:28 GMT
Server: Apache
Cache-Control: no-cache=”set-cookie,set-cookie2″
Set-Cookie: JSESSIONID=0000UzUSX5tGcqe8AriacpxKyf7:120mbebeh;Path=/
Set-Cookie: ShopperManager/enterprise=d7b1c7dc-1bf6-11dd-b18b-cbe10af70195;Expires=Mon, 25-May-2076 08:45:35 GMT;Path=/
Cache-Control: no-cache
Pragma: no-cache
Expires: Thu, 01 Jan 1970 00:00:00 GMT
X-Powered-By: 1800Flowers web server
X-AspNet-Version: 1.21.366
Connection: close
Transfer-Encoding: chunked

No redirect. So both versions get me to the site, but excluding the dashes from my request does require the server to take an extra step. Should I take that to mean that the version with the dashes (the one that the logo looks like the domain name would be) is the default and the one without them (the one that you hear when they say the name) was set up to catch errors and keep the competitors from getting control of mistyped traffic?

One last thing to note, now that Clinton’s “apparently” won Indiana, according to MSNBC: In the footer of the site’s home page they’ve got links to other sites they run. Among them are 1-800-Baskets and 1-800-Greetings, with links anchored by “Greeting Cards” and “Gift Baskets” respectively. And the targets of those links? 1-800-Baskets is at http://www.1800baskets.com/ and 1-800-Greetings is http://www.1800greetings.com/. And just for fun, what happens when you add the dashes so the URLs match the company names on those two?

http://www.1-800-baskets.com/ returns a 301 redirect to http://www.1800baskets.com/ and http://www.1-800-greetings.com/ returns a proper 200, but the news isn’t quite as good as it appears. If I request http://www.1800greetings.com/ I get a 301 to http://1800greetings.cardways.com/cp001/clientinterface/creategiftcard.asp?cltid=121&returnid=0 (an affiliate deal, apparently), which looks like this:

And yes, a request for http://www.1-800-greetings.com/ returns a 200, but it looks like this:

The wrong site.

When you talk about a URL, what do you say? Actually, before we even get to that, how do you say “URL”? I’ve already given my own answer away by writing “a URL” (yoo-arr-ell) rather than “an URL” (erl). Similarly, I make my living as an SEO (ess-ee-o) rather than a SEO (see-o).

Honestly, this interweb business is young enough that a lot of the terminology isn’t standardized, and similarly, many of the acronyms don’t have a set pronunciation. That will change with time, I’m sure. After all, nobody pronounces “scuba” (ess-see-yoo-bee-ay) — at least I hope not.

So, back to the question at hand: when you say a URL, just what do you say? Do you include all its parts, using the official generic syntax of scheme, authority, path, query, and fragment? Don’t be silly, of course you don’t.

As more and more people go online and become accustomed to these things, it’s become pretty standard to leave out the “http://”. Even the “www” (which really shouldn’t be necessary) is left out most of the time. Just look at an ad in print or on the eye of hell. The web site is usually just represented as domain.com. And you really can, in most cases, just type that into your browser’s location bar. The browser will go with the default protocol of http, and the server, when reached, will add a trailing slash and sometimes the www as well.

Of course, there will always be exceptions, like this guy:

LETTERMAN: Can I just take a second here, Larry — I’m sorry, I don’t mean to interrupt — to give our World Wide Web address. If people want to e-mail us, we’re on the World Wide Web as well.

KING: You are too? What is it?

LETTERMAN: wwwww.comcomcom — comcomdiggitydiggitydiggitydank.com –comdiggitywww.comdave.com. So give us some of that e-mail…

What about the directory or the file name? If you want to send someone to a document other than the home page, you have to go further than just the domain name, right? Nope. There are ways around that too.

CNN keeps their political news in a /politics/ directory, but apparently they were concerned that telling people to go to cnn dot com slash politics (which would end up taking them to http://www.cnn.com/POLITICS/ after the browser and the server had their way with the request) was asking too much of them, so they’ve made it easier. You can simply go to cnnpolitics dot com.

And what happens when you request that URL? Your browser adds the http://, so a request gets sent to the server at http://cnnpolitics.com. From there, you figure the server would tack on the www and the trailing slash, right?

Wrong again. The server responds with a status code of 302 (“found”) and redirects the request to http://www.cnn.com/politics/. So why tell people to go to the cnnpolitics.com domain? I suppose part of the reason is that they registered the domain name in order to keep anyone else from getting it, but I think it’s pretty safe to assume that they chose to use it and redirect it to make things easier for the user. Apparently, it’s easier to communicate cnnpolitics dot com than cnn dot com slash politics, especially when it’s spoken rather than printed.

What can we take away from this? Apparently, it’s that slashes are problematic. Maybe that’s because a standard keyboard has two different kinds of slashes: the forward slash and the backslash. When a person says “slash,” they almost invariably mean the forward slash, but I suppose it could still cause some confusion. When I say “guitar,” do you think of an electric guitar, an acoustic guitar, or what? Certainly, when the electric guitar was new to the world, “guitar” meant acoustic guitar, just like one used to be able to say “television” to mean a black and white television, then at some point one would specify “color television” for the new technology. Eventually, there would be a point at which you couldn’t just say “television” because you might have been referring to either color or black and white. Now, just about every television is color, so one can say “black and white television” and just “television” for color. And we’ll go through this again with digital, high definition, etc.

But with slashes, it seems that despite the fact that there’s a general understanding that “slash” means “forward slash,” it’s not understood widely enough, and some have decided to find ways around it. Hence, if you own enough domains, you don’t mind setting up redirects every time you publish a new document, and you don’t care about how hard your poor defenseless server has to work, you can just feed people domain names to direct them to your pages.

There’s another option out there, but it pretty much relies on people understanding what slashes are for and which ones are kosher on the web. Shell has been running spots on the eye of hell of late, promoting how hard they’re working to clean up the environment and find alternative, clean sources of energy. As if. Basically, it’s one of those “please don’t hate us, we need those enormous profits more than you know” campaigns.

At the end of the spot, they tell you to go to shell.com/us/realenergy for more information. That’s the URL you see on the screen, but the voiceover says, “shell dot com [pause] us [pause] realenergy.” Is that helpful? If you’re not looking at the screen when you hear this, will you know what to make of those pauses?

And what happens when you try to go to shell.com/us/realenergy?

HTTP/1.1 301 Moved Permanently
Server: Sun-ONE-Web-Server/6.1
Date: Mon, 07 Apr 2008 20:10:07 GMT
Content-length: 0
Content-type: text/html
Location: http://www.shell.com/us/realenergy

You get redirected to the URL with the www. From there…

HTTP/1.1 301 Moved Permanently
Server: Sun-ONE-Web-Server/6.1
Date: Mon, 07 Apr 2008 20:11:40 GMT
Content-length: 0
Content-type: text/html
Location: http://www.shell.com/us/realenergy/
Connection: close

You get the slash added at the end via a redirect. After that, you get a 200 response at the new URL, hit a bunch of JavaScript and then…

<meta http-equiv=”refresh” content=”1; URL=http://realenergy.shell.com/?lang=en&page=homeFlash”> </meta>

A meta refresh to a subdomain of shell.com, without the www, and with some personalization, based on a flash sniffer and most likely my IP address.

As it turns out, they could have told me to go to realenergy.shell.com (look mom, no slashes!), and I’d have ended up in the same place. But I suppose subdomains are even harder to communicate than any of that other stuff.

It’s true, you kids. After making us wait almost two years, Fafnir an Giblets have returned, and one can only hope the Medium Lobster is right behind them.

And if this is one of those first of April pin-a-fish-on-my-back-and-douse-me-with-mayonnaise tricks, I’ll be very, very upset.

There’s an Easter… um… Passover… well, call it “Spring” edition of one of my favorite swarms going on: Blog Against Theocracy. You should go check it out.

I’d have written an entry myself, but I’ve just been too busy procrastinating about work.