Please keep prefetching

Liz is the latest to write on the dangers of prefetching. Basically, if you use a prefetch capable browser (like Firefox), it grabs specified links and caches them so that when you click on them the download appears instant. It utilizes the time you spend on each page to speed up the time it takes to get to the next page. Although it trades bandwidth for convenience, it is, to my mind, a “good thing.”

Google has started specifying links for prefetching on their results, and people are upset that their browsers may be downloading things they are uninterested in. Liz, following on Cameo Wood’s posting, suggests that this messes with server logs and makes it look to the owners (and to your company, if they are monitoring you) like you have been where you have not been.

And this is a bad thing?

At least in the latter case, it means that I now have complete plausible deniability. “That ‘XXX Hedgehog Love’ site? No, no, it was just prefetched.” Of course, whoever was monitoring my surfing would find such a claim indefensible if not for the fact that 80% of the company’s employees are prefetching.

So, I encourage you to keep the default pre-caching *on*. Not only will it make your surfing experience more effective and enjoyable, you are acting as a shield for those of us who are visiting nefarious sites on the sly.

This entry was posted in Uncategorized and tagged . Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.


  1. Posted 4/4/2005 at 3:44 pm | Permalink

    Joi Ito noticed this too… I found it there just moments ago. I like your angle on it though… sweet.

  2. Posted 4/4/2005 at 11:42 pm | Permalink

    As someone who pays for bandwidth, I say this is a terrible idea.

  3. Posted 4/5/2005 at 10:02 am | Permalink

    Everyone pays for bandwidth. I definitely see where this is an issue for those who pay a *lot* for bandwidth (e.g., in lesser-developed countries). But my guess is that anyone that shows up as a first hit on Google can afford an extra number of page loads. For folks who are blogging, this will be inconsequential when compared to the bandwidth (and processing) eaten up by spammers.

    On the other hand, it does put an interesting spin on a form of DOS. If one wanted to bring down a site, as a group, it would be possible to create an invisible flash mob merely by collectively putting up prefetched links to them.

    So, if, as a group we wanted to bring down an al Qaeda site (this example will, I hope, keep the FBI from my front door), we would all just link to it, and include the prefetch rel. Killing them with kindness.

  4. Posted 4/11/2005 at 12:52 pm | Permalink

    Which is to say, prefetching is a form of DOS.

  5. Posted 4/11/2005 at 12:53 pm | Permalink

    Or rather, can be.

    With all the effort that’s put forth by web authors to optimize JPEGs (30 quality, not 60!) and code (tables are soooo bad!), I really can’t support something like this.

  6. Alex
    Posted 4/11/2005 at 1:03 pm | Permalink

    I guess it is in the same way that (e.g.) flash mobs are. In other words, I suppose intentionality has something to do with a dDOS.

    I was under the impression that most most web authors optimize their sites not to save themselves on bandwidth charges, but to improve the end user’s experience. (Note that I wouldn’t have any way of knowing — this page is something like a 1 Mb download and isn’t even valid HTML.)

    If that’s the case, then most of those who spend lots of time making sure their site is fast-loading would, I assume, support a technology that allows their site to be instant loading.

    The ideal surfing situation lets me flip through websites like I flip through pages in a book, with each page popping up instantly.

    More proximally, I suppose google could promote a kind of “crawl me but don’t pre-fetch me” meta-tag.

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>