Businessweek Archives

Cookies: Let's come up with a different name


? Try this for managing MP3s online |

Main

| Blog the Bikes ?

June 05, 2005

Cookies: Let's come up with a different name

Stephen Baker

Lots of feedback on my post about cookies. The common point is that cookies are often extremely useful, and terribly misunderstood. The trouble is that even sites that use cookies in constructive ways avoid talking clearly about them--because they know that users fear them, and will erase them. Here is some original thinking on the matter. I think the term is so intertwined, at least in the people's minds, with bad cookies and spyware, that it's beyond redemption. How about coming up with a new word for good cookies? Suggestions? --Photo morguefile

05:15 AM

advertising

TrackBack URL for this entry:

http://blogs.businessweek.com/mt/mt-tb.cgi/

Interesting suggestion but I fear it will simply lead to the elimination of another term from our useful vocabulary as a new word will immediately be hijacked by the disreputable.

This has already happened with job titles as Gary Trudeau has shown in his recent Doonesbury strips with the waiters at "McFriendly's" all called Vice Presidents...

Posted by: Paul Woodward at June 5, 2005 05:49 AM

This page about what cookies actually are and how they can be used might be educational. I think many people panic just because they don't know what a cookie is...

http://www.htmlgoodies.com/beyond/reference/article.php/3472661

It's a good enough rundown for the average Joe.

Posted by: nortypig at June 5, 2005 06:15 AM

The term "cookie" is a strange hybrid. Usually, technical terms tend to be quite cryptic and unrelated to anything that you or I might experience in the real world. But "cookie" is obviously something that we all recognize. That's part of the problem, because the real-world concept of a "cookie" relates not even slightly to the technical concerns of "client-side web-based interaction state management information packet" (or "HTTP State Management Mechanism"), which is basically what a so-called "cookie" is, roughly speaking.

Cookies are frequently billed as being for "session management", but there would be no need to maintain them permanently, after you leave a web site, after you've closed the browser, or shutdown your computer. Many if not most cookies are used for inter-session state, so web sites can "welcome" you on your next visit.

Shame on anybody at the IETF who seriously believes that cookies were or are the "best" approach to managing user state information. It's the approach that someone *happened* to take in the early days of internet web browsers, and inertia kept it going. All the "reasons" for keeping them as-is are just a heap of excuses, and mostly *bad* excuses at that.

The main problem with cookies is that they *imply* a relationship between the user and a web site that the user may neither have made a commitment for, may not be interested in, and might even be appalled if they knew that a relationship had been established on their behalf. If the browser popped up a plain English message that asked you if you wish to establish a permanent relationship with the selected web site, the story would be completely different than it is today.

Metaphorically, people "visit" web sites the way they visit stores. Most stores don't require that you accept and carry around some kind of "store visit state management device" just because you want to browse around and ask some questions. And if the store needs any info, they *ask* you for it, and you can then say "no" and walk out.

The early browser designers and the IETF *failed* to design "cookies" to operate in a user-friendly manner with respect to establishing relationships. Again, shame on them.

My suggestion for a new name: "burrs". They may indeed peform a necessary function, but they're also very annoying. And if the IETF ever does get around to designing a user-friendly mechanism (which I don't expect to *ever* happen, given their cavalier attitude towards users), then maybe they could be called a "relationship tokens". A user should be able to pop up a list of their "online relationships" (in simple English and user-oriented terms), regardless of whatever bizarre internal technical terminology the IETF has chosen.

BTW, I'm still annoyed that your blog software *deletes* the blank lines I place between each of my paragraphs. Could it be a "cookie" problem? In fact, my browser (IE 6.0) does pop up an error message when I click Preview or Post that says "A Runtime Error has Occurred... Line 121... Error: 'this.bakecookie.0' is null or not an object." Sounds like maybe BW hasn't properly "baked" its cookies. Good Grief.

-- Jack Krupansky

Posted by: Jack Krupansky at June 5, 2005 12:36 PM

I am actually shocked at your suggestion! It's tantamount to re-branding for deceptive purposes. When companies get into trouble, they play these games (think Phillip Morris , now Altria Group). One article I read cited ValuJet , now AirTran (http://seattletimes.nwsource.com/html/businesstechnology/2002039047_namechange18.html).

A clove of garlic by any other name would still stink.

Posted by: Charles Pizzo at June 6, 2005 02:06 AM

Cookies are not the only way to do things but they're convenient - not evil. In answer to walking into a shop I'd answer that yes you're probably on video and if they wanted to find out all of your info your face is a good start, plus the credit card number you gave when you purchased the shirt and tie.

Thinking you go anywhere on the net without it being recorded is very naive. If you don't have a web hosting account go to a friend who does and look in their stats. Your IP is logged. So I don't understand the arguments put forward at all.

When I go to a forum for the 10th time in a day I'm glad they use cookies to remember nortypig...

the username and password are on the server but a small cookie tells them I was there earlier in the day.

If you could tell me how to maintain sessions while the user has gone elsewhere or even shut their computer I am willing to learn though...

The truth - don't go to bad places on the web, don't give your details to strangers when prompted, and don't ever believe any part of the web or anything you do isn't recorded for a life time. It's just the way we live now. So cookies in themselves aren't really anything but a text file to remember stuff..

You can set the time for a cookie to expire, a day, a week, a year.

Posted by: nortypig at June 6, 2005 09:01 PM

Also when you walk into a store you don't know who owns it or who owns the products you're paying for. You should as that is a relationship, as I understand you.

Investigate Nestle and you'll find they manufacture weapons, contribute to 3rd world famine and basically exploit the poor... buying Milo or Condensed Milk gives you an unknown relationship between you and the guy being shot with the gun... Such is life.

Posted by: nortypig at June 6, 2005 09:05 PM

The back cookies problem is a template bug in Movable Type, easily fixed. See:

http://forums.sixapart.com/index.php?showtopic=56838&hl=

Posted by: George at October 29, 2006 11:07 PM

Ok, I have to agree with nortypig here, cookies are useful. I don't want to have to type in my name over and over again when I go to a site. Beside, we have no other alternative so we might as well use them.

Anyways, does anyone know where the term 'cookie' comes from. I did some research and it originates from "magic cookie" which was used in Unix based software, but I can't find any reference to that anywhere. I mean, we know where other computer terms like bug and easter egg come from, but what about "magic cookie"?

Posted by: Jordan at April 4, 2007 01:19 PM


Too Cool for Crisis Management
LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus