Google’s Got Plenty of Compute Cycles

by on August 3, 2008 · 36 comments

There are apparently people who believe that it’s some kind of technological Faux pas to type a website’s URL into the search bar. As Joe Weisenthal points out, this is completely nonsense. There are a number of good reasons to use the search bar even if you have a pretty good idea of a site’s URL.

Beyond the specific reasons Joe gives, there’s a more fundamental issue of cognitive economy. URLs have to be exact, and so remembering them takes a non-trivial amount of cognitive effort. If I want to remember the Institute for Humane Studies website, I have to remember that it’s theIHS, and that it’s a .org rather than a .com or a .net. But if I type “IHS” into Google, the Institute for Humane Studies is the third search term. If I type something a little more descriptive, like humane studies, it comes up as the first result. Search terms don’t have to be exact, and so they tend to be much easier to remember: type something in the general vicinity of what you’re looking for, and Google will find it for you.

The point isn’t that I couldn’t remember theihs.org. Rather, it’s that remembering the URLs of all the websites you visit is a waste of cognitive energy in exactly the same way that it would be a waste to remember IP addresses rather than domain names. Technically speaking, the IP address lookup would be faster, but the difference is so trivial that it’s swamped by the fact that the human brain isn’t as good at remembering 32-bit numbers as it is at remembering well-chosen domain names. By the same token, even if the search bar isn’t the “right” place to put URLs, it will, in practice and on average, be the quickest way for actual human beings to get to the sites they’re looking for.

This is an example of a general attitudinal problem that’s distressingly common among geeks. Geeks have an tendency to over-value lower-level layers of the technology stacks based on the misguided belief that higher-level technologies are unnecessarily wasteful. Many geeks’ preference for text over graphics, command lines over GUIs, text editors over word processors, and so forth seems to too often be motivated by this kind of false economy. (To be clear I’m not claiming that there aren’t good reasons for preferring command lines, text editors, etc, just that this particular reason is bogus.) What they miss is that human time and attention is almost always more scarce than the trivial amount of computing power they’re conserving by using the less complex technology. The 2 seconds it takes me to remember a website’s URL is worth a lot more than the tenth of a second that it takes Google to respond to a search query.

  • Joseph Price

    The “Awesome Bar” of Firefox 3 obviates much of this dilemma–at least, if you’ve bookmarked the page or visited it recently.

    But your calculus regarding time & attention vs. computer power is good insight.

  • Joseph Price

    The “Awesome Bar” of Firefox 3 obviates much of this dilemma–at least, if you’ve bookmarked the page or visited it recently.

    But your calculus regarding time & attention vs. computer power is good insight.

  • desm0nd

    take a look at peers which does somekind of preview for all things typed into the searchbar. It saves some more time.

    https://addons.mozilla.org/en-US/firefox/addon/6589

  • desm0nd

    take a look at peers which does somekind of preview for all things typed into the searchbar. It saves some more time.

    https://addons.mozilla.org/en-US/firefox/addon/

  • http://bennett.com/blog Richard Bennett

    People who deal with numbers and other engineering concepts tend to be very comfortable with them, Tim, and people who do use specific options of programmers’ tools need a way to pass their preferences to their tools. Commmand line interfaces are a technical mechanism for technical people, and ordinary civilians need not worry their pretty little heads over the inscrutable ways of the Geek Overlords.

  • http://bennett.com/blog Richard Bennett

    People who deal with numbers and other engineering concepts tend to be very comfortable with them, Tim, and people who do use specific options of programmers’ tools need a way to pass their preferences to their tools. Commmand line interfaces are a technical mechanism for technical people, and ordinary civilians need not worry their pretty little heads over the inscrutable ways of the Geek Overlords.

  • Adam Thierer

    I want to second Joseph Price’s comment about the “Awesome Bar” in Firefox 3 obviating much of this dilemma. It is indeed awesome and really improves one’s web-surfing productivity. I can now easily find previous sites and articles that I stumbled upon and then want to retrieve again at a later time when writing a blog post or looking for an article to cite in a footnote for a paper.

  • http://www.techliberation.com Adam Thierer

    I want to second Joseph Price’s comment about the “Awesome Bar” in Firefox 3 obviating much of this dilemma. It is indeed awesome and really improves one’s web-surfing productivity. I can now easily find previous sites and articles that I stumbled upon and then want to retrieve again at a later time when writing a blog post or looking for an article to cite in a footnote for a paper.

  • http://www.thestalwart.com Joseph Weisenthal

    Thanks Tim. I think your broader elucidation of my point is spot on. This kind of mentality probably goes a long way to explaining why various open source/geeky projects haven’t really taken off or proven as disruptive as they should be.

    Firefox is obviously an exception to this, but it’s not surprising, given stuff like the “Awesome Bar”, which is all about making browsing simple enough for a total moron. Eventually I’ll have to upgrade to FF3, but I tend to loathe any upgrades, so I’ll wait awhile.

  • DSDan

    A potential issue in using the search bar that should be mentioned is that you are now providing Google with a complete history of the websites you visit, not just the terms you searched for.

  • Alien

    +1 on DSDan… that is why I bookmark and made my own list of daily links in an HTML page to remember for me. So Google doesn’t know which places I frequent or repeatedly visit.

  • Eshan

    I’m using Internet Explorer with the Quero toolbar. It extends the regular Ctrl-Enter behavior (which just adds www. and .com) to do a Google “I’m feeling lucky search”. So you can type in “institute for humane studies”, hit Ctrl-Enter, and you go straight to http://www.theihs.org/

    It’s pretty nice to access things by the name I use, and not a domain name.

  • http://www.thestalwart.com Joseph Weisenthal

    Thanks Tim. I think your broader elucidation of my point is spot on. This kind of mentality probably goes a long way to explaining why various open source/geeky projects haven’t really taken off or proven as disruptive as they should be.

    Firefox is obviously an exception to this, but it’s not surprising, given stuff like the “Awesome Bar”, which is all about making browsing simple enough for a total moron. Eventually I’ll have to upgrade to FF3, but I tend to loathe any upgrades, so I’ll wait awhile.

  • DSDan

    A potential issue in using the search bar that should be mentioned is that you are now providing Google with a complete history of the websites you visit, not just the terms you searched for.

  • Alien

    +1 on DSDan… that is why I bookmark and made my own list of daily links in an HTML page to remember for me. So Google doesn’t know which places I frequent or repeatedly visit.

  • Eshan

    I’m using Internet Explorer with the Quero toolbar. It extends the regular Ctrl-Enter behavior (which just adds www. and .com) to do a Google “I’m feeling lucky search”. So you can type in “institute for humane studies”, hit Ctrl-Enter, and you go straight to http://www.theihs.org/

    It’s pretty nice to access things by the name I use, and not a domain name.

  • http://blaynesucks.com Aaron Massey

    While I agree with your overarching point that there are serious practical reasons to searching rather than using the address bar to specify a URL, I personally don’t know that it’s faster, even cognitively, than remembering the address and typing in the address bar for commonly visited sites. I think there are two situations:

    (1) Worst case: Searching from google.com itself.

    Steps:
    1. wait for google.com to load
    2. enter search terms
    3. click “Search” button
    4. wait for results to load
    4. scan results and use mouse to click on appropriate link
    5. wait for the site to load

    (2) Searching from a search bar on the browser.

    Steps:
    1. enter search terms and press enter
    2. wait for the results to load
    3. scan to find the right link and use mouse to click on the appropriate link
    4. wait for the site to load

    In both cases, you have a cognitive switch from keyboard to mouse mode, which is bad from an HCI standpoint and may be unnecessary if you knew the URL. Also, in both cases you have at least one additional page load. From a cognitive standpoint, page loads are really bad because they essentially build in time that allows people to get distracted. (Ever searched for something and then wondered why you did it when you got the results?) They are analogous to a “page cache miss” in operating system terms.

    Google recognizes this and optimizes fanatically for speed, but even with these optimizations a person still has to watch to make sure the search was entered and the page is actually loading. Then they have to recognize that the page has completed loading so that they can begin a new page load by finding the appropriate link and clicking on it.

    Also, advanced searches are hard without memory of the URL. For example, if I wanted to find an article on copyright that I recently read on the Technology Liberation Front, I could perform a much more accurate Google search by entering search terms AND the URL like this:

    “copyright site:techliberation.com”

    If I don’t know the URL, then I could get tons of other stuff on copyright that I would have to sift through before I found the specific article I wanted.

    I agree with the previous posters about the Awesome Bar, and I certainly don’t have a problem with people wasting their own time with searches, but I am willing to devote some cognitive power to save time when loading commonly visited sites. (And I think the market for selling “good” domain names shows that a lot of entrepreneurs see this as a key advantage.)

  • Tim Lee

    Aaron,

    I’m not going to claim that you should always use the search bar, which is just as stupid as saying you should always use the location bar. If an address is short and you know it off the top of your head, by all means you should use the location bar. Of course, for really commonly-visited sites, you should probably create a bookmark so you don’t have to type it in at all.

    However, I think your HCI analysis is flawed. When i want to do a Google search with safari, I type apple-option-f, type my search term, and hit enter. The top result is almost always what I’m looking for, and I have my mouse hovering over it by the time the page loads (The top Google search result’s approximate location on the page has long since become a mater of muscle memory). When you consider that I’d be shifting my hand to the mouse anyway in order to use the scroll wheel, the only real difference from using the location bar is the latency to Google’s servers. On a fast network this is a fraction of a second.

  • http://blaynesucks.com Aaron Massey

    “Of course, for really commonly-visited sites, you should probably create a bookmark so you don’t have to type it in at all.”

    I only started using bookmarks when the Awesome Bar came out. I certainly made them before that, but I didn’t used them for my most commonly visited sites because they simply aren’t faster than typing. I used them for those specific pages where I found myself wanting to refer to something and couldn’t remember the site URL or any search terms that would bring it up on the first few pages.

    “The top result is almost always what I’m looking for, and I have my mouse hovering over it by the time the page loads (The top Google search result’s approximate location on the page has long since become a mater of muscle memory). When you consider that I’d be shifting my hand to the mouse anyway in order to use the scroll wheel, the only real difference from using the location bar is the latency to Google’s servers.”

    This is great — for you. It’s always the “human” in human computer interaction that makes HCI generalizations difficult. :-) Personally, I avoid the mouse as much as possible. Keyboard shortcuts are almost always faster for me. (I have never used the file menu to open a new tab in Firefox.) This is why computer interfaces have both.

    My goal wasn’t to get into a debate about HCI though, so I must have not written very well in my previous comment. My goal was to address this part of your original post:

    “This is an example of a general attitudinal problem that’s distressingly common among geeks. Geeks have an tendency to over-value lower-level layers of the technology stacks based on the misguided belief that higher-level technologies are unnecessarily wasteful.”

    Let me give the two-sentence geek side of this argument”

    There is a general attitudinal problem that’s distressingly common among the technically illiterate. Non-Geeks have a tendency to under-value the use of lower-level layers of technology stacks based on the misguided belief that higher-level technologies are just as fast.

    The key is striking the balance between investing cognitive energy to optimize the common use of a piece of technology (by learning a URL) and accepting that it may simply be faster to do it the basic “dumb” way (by just googling).

  • http://blaynesucks.com Aaron Massey

    Also, your muscle memory is really just another form of cognitive energy required for optimization. But, I guess that gets back to the HCI debate…

  • http://www.blaynesucks.com Aaron Massey

    While I agree with your overarching point that there are serious practical reasons to searching rather than using the address bar to specify a URL, I personally don’t know that it’s faster, even cognitively, than remembering the address and typing in the address bar for commonly visited sites. I think there are two situations:

    (1) Worst case: Searching from google.com itself.

    Steps:
    1. wait for google.com to load
    2. enter search terms
    3. click “Search” button
    4. wait for results to load
    4. scan results and use mouse to click on appropriate link
    5. wait for the site to load

    (2) Searching from a search bar on the browser.

    Steps:
    1. enter search terms and press enter
    2. wait for the results to load
    3. scan to find the right link and use mouse to click on the appropriate link
    4. wait for the site to load

    In both cases, you have a cognitive switch from keyboard to mouse mode, which is bad from an HCI standpoint and may be unnecessary if you knew the URL. Also, in both cases you have at least one additional page load. From a cognitive standpoint, page loads are really bad because they essentially build in time that allows people to get distracted. (Ever searched for something and then wondered why you did it when you got the results?) They are analogous to a “page cache miss” in operating system terms.

    Google recognizes this and optimizes fanatically for speed, but even with these optimizations a person still has to watch to make sure the search was entered and the page is actually loading. Then they have to recognize that the page has completed loading so that they can begin a new page load by finding the appropriate link and clicking on it.

    Also, advanced searches are hard without memory of the URL. For example, if I wanted to find an article on copyright that I recently read on the Technology Liberation Front, I could perform a much more accurate Google search by entering search terms AND the URL like this:

    “copyright site:techliberation.com”

    If I don’t know the URL, then I could get tons of other stuff on copyright that I would have to sift through before I found the specific article I wanted.

    I agree with the previous posters about the Awesome Bar, and I certainly don’t have a problem with people wasting their own time with searches, but I am willing to devote some cognitive power to save time when loading commonly visited sites. (And I think the market for selling “good” domain names shows that a lot of entrepreneurs see this as a key advantage.)

  • Tim Lee

    I certainly made them before that, but I didn’t used them for my most commonly visited sites because they simply aren’t faster than typing.

    Safari allows you to bind user-defined short-cuts to Apple-1 through Apple-9, so I set those for my absolutely most-used sites.

    It’s always the “human” in human computer interaction that makes HCI generalizations difficult.

    Well right. I wasn’t claiming that everyone should do things the way I do. I was just criticizing the original RWW that claimed that it was somehow wrong to do things the way I do.

    There is a general attitudinal problem that’s distressingly common among the technically illiterate. Non-Geeks have a tendency to under-value the use of lower-level layers of technology stacks based on the misguided belief that higher-level technologies are just as fast.

    In many cases what we’ve got here is an up-front cognitive investment that will pay off over time if you spend a ton of time doing certain very technical operations on a computer. Once I know how to use a command line, I can perform certain file and text operations a lot more efficiently than I could with a GUI file browser. For the 95 percent of people who rarely perform those kinds of operations, it’s not worth the investment, while for the 5 percent of people who do, it is.

    I don’t have any quarrel with the people who have made that investment and find it to have been useful. (I’m in that category myself–I still edit text files with vi) I just think geeks over-estimate the fraction of the population for which this is true. Most people are justified in remaining rationally ignorant about the low-level operation of their computers. What I’m objecting to is a tendency toward a kind of geek machismo, in which you’re some kind of sellout if you aren’t doing everything as close to the bare metal as possible.

  • Tim Lee

    Aaron,

    I’m not going to claim that you should always use the search bar, which is just as stupid as saying you should always use the location bar. If an address is short and you know it off the top of your head, by all means you should use the location bar. Of course, for really commonly-visited sites, you should probably create a bookmark so you don’t have to type it in at all.

    However, I think your HCI analysis is flawed. When i want to do a Google search with safari, I type apple-option-f, type my search term, and hit enter. The top result is almost always what I’m looking for, and I have my mouse hovering over it by the time the page loads (The top Google search result’s approximate location on the page has long since become a mater of muscle memory). When you consider that I’d be shifting my hand to the mouse anyway in order to use the scroll wheel, the only real difference from using the location bar is the latency to Google’s servers. On a fast network this is a fraction of a second.

  • http://www.blaynesucks.com Aaron Massey

    “Of course, for really commonly-visited sites, you should probably create a bookmark so you don’t have to type it in at all.”

    I only started using bookmarks when the Awesome Bar came out. I certainly made them before that, but I didn’t used them for my most commonly visited sites because they simply aren’t faster than typing. I used them for those specific pages where I found myself wanting to refer to something and couldn’t remember the site URL or any search terms that would bring it up on the first few pages.

    “The top result is almost always what I’m looking for, and I have my mouse hovering over it by the time the page loads (The top Google search result’s approximate location on the page has long since become a mater of muscle memory). When you consider that I’d be shifting my hand to the mouse anyway in order to use the scroll wheel, the only real difference from using the location bar is the latency to Google’s servers.”

    This is great — for you. It’s always the “human” in human computer interaction that makes HCI generalizations difficult. :-) Personally, I avoid the mouse as much as possible. Keyboard shortcuts are almost always faster for me. (I have never used the file menu to open a new tab in Firefox.) This is why computer interfaces have both.

    My goal wasn’t to get into a debate about HCI though, so I must have not written very well in my previous comment. My goal was to address this part of your original post:

    “This is an example of a general attitudinal problem that’s distressingly common among geeks. Geeks have an tendency to over-value lower-level layers of the technology stacks based on the misguided belief that higher-level technologies are unnecessarily wasteful.”

    Let me give the two-sentence geek side of this argument”

    There is a general attitudinal problem that’s distressingly common among the technically illiterate. Non-Geeks have a tendency to under-value the use of lower-level layers of technology stacks based on the misguided belief that higher-level technologies are just as fast.

    The key is striking the balance between investing cognitive energy to optimize the common use of a piece of technology (by learning a URL) and accepting that it may simply be faster to do it the basic “dumb” way (by just googling).

  • http://www.blaynesucks.com Aaron Massey

    Also, your muscle memory is really just another form of cognitive energy required for optimization. But, I guess that gets back to the HCI debate…

  • Tim Lee

    I certainly made them before that, but I didn’t used them for my most commonly visited sites because they simply aren’t faster than typing.

    Safari allows you to bind user-defined short-cuts to Apple-1 through Apple-9, so I set those for my absolutely most-used sites.

    It’s always the “human” in human computer interaction that makes HCI generalizations difficult.

    Well right. I wasn’t claiming that everyone should do things the way I do. I was just criticizing the original RWW that claimed that it was somehow wrong to do things the way I do.

    There is a general attitudinal problem that’s distressingly common among the technically illiterate. Non-Geeks have a tendency to under-value the use of lower-level layers of technology stacks based on the misguided belief that higher-level technologies are just as fast.

    In many cases what we’ve got here is an up-front cognitive investment that will pay off over time if you spend a ton of time doing certain very technical operations on a computer. Once I know how to use a command line, I can perform certain file and text operations a lot more efficiently than I could with a GUI file browser. For the 95 percent of people who rarely perform those kinds of operations, it’s not worth the investment, while for the 5 percent of people who do, it is.

    I don’t have any quarrel with the people who have made that investment and find it to have been useful. (I’m in that category myself–I still edit text files with vi) I just think geeks over-estimate the fraction of the population for which this is true. Most people are justified in remaining rationally ignorant about the low-level operation of their computers. What I’m objecting to is a tendency toward a kind of geek machismo, in which you’re some kind of sellout if you aren’t doing everything as close to the bare metal as possible.

  • David

    I vaguely recall some analysis of common search queries in both Yahoo and Google search engines from a year or two ago. The analysts were surprised to find that the most common search term in Yahoo was Google and vice versa.

    Seems like people have one or the other as a homepage and don’t realise it… seeking the opposite as a search location.

  • David

    I vaguely recall some analysis of common search queries in both Yahoo and Google search engines from a year or two ago. The analysts were surprised to find that the most common search term in Yahoo was Google and vice versa.

    Seems like people have one or the other as a homepage and don’t realise it… seeking the opposite as a search location.

  • Pingback: amnesia a machine for pigs free download

  • Pingback: foods that burn fat

  • Pingback: 180 37.5 mg phentermine tablet

  • Pingback: adipex for sale

  • Pingback: Camfrog

  • Pingback: Dentist Camberley

  • Pingback: prix de l'immobilier

  • Pingback: More about the author

Previous post:

Next post: