invisible image with a MeFi-specific URL March 7, 2003 4:09 AM Subscribe
Linking to your own site in a comment is kosher, but what about putting an invisible image with a MeFi-specific URL on that site's front page?
(Ok, correction, it's slightly more invasive; normal server logs don't capture data such as your screen dimensions or CPU type. Still, even that is fairly innocuous data.)
posted by ook at 5:05 AM on March 7, 2003
posted by ook at 5:05 AM on March 7, 2003
That was a very different situation, crunchland: in that case the image was placed on metafilter itself, so was tracking every hit to that metafilter page. Evil. This one is sitting on gravlab.com, so the guy is only tracking hits to his own site. Not evil.
Which now that I think about it, is kind of an odd distinction: I don't know why I have such a visceral reaction against capturing information from someone else's site, but find it perfectly acceptable for someone to do it on his own site... I suppose because when I visit site X, I expect a certain amount of data to be captured by site X -- the site earned that information, in a way, by getting me to visit it. To have that same information be diverted to site Y, possibly without me or the owner of site X ever knowing about it, feels like theft. Even though the data, in both cases, is perfectly innocuous information that can't be tied to me as an individual.
Interesting ethical question, that.
posted by ook at 5:34 AM on March 7, 2003
Which now that I think about it, is kind of an odd distinction: I don't know why I have such a visceral reaction against capturing information from someone else's site, but find it perfectly acceptable for someone to do it on his own site... I suppose because when I visit site X, I expect a certain amount of data to be captured by site X -- the site earned that information, in a way, by getting me to visit it. To have that same information be diverted to site Y, possibly without me or the owner of site X ever knowing about it, feels like theft. Even though the data, in both cases, is perfectly innocuous information that can't be tied to me as an individual.
Interesting ethical question, that.
posted by ook at 5:34 AM on March 7, 2003
Let me get this straight. We're worried about someone tracking hits on their own site? We better outlaw WebTrends huh?
"Drop those log files asshole!!! You're under arrest!!!"
posted by y6y6y6 at 5:41 AM on March 7, 2003
"Drop those log files asshole!!! You're under arrest!!!"
posted by y6y6y6 at 5:41 AM on March 7, 2003
I have long suspected that there are invisible linked images on the lens of my glasses that tell me to do bad things...like boost the economy by impulse shopping.
posted by Shane at 6:16 AM on March 7, 2003
posted by Shane at 6:16 AM on March 7, 2003
ah. I guess I didn't even consider someone would complain here about something someone else does on their own website. that seems almost inconceivable.
posted by crunchland at 8:50 AM on March 7, 2003
posted by crunchland at 8:50 AM on March 7, 2003
The third link, I'm lost, as all I see is a white screen. Yet I understand the comments.
skoosh do you feel you were trolled into viewing the site for your info to be collected? See if you are collecting info on your site, but make an excuse for folks to come, I could see feeling baited.
Posting your site outside of the user page makes for slippery slope is what I'm seeing. I did it in a comment and really wished I hadn't but then again linking to my userpage would have accomplished the same thing w/o qualms.
posted by thomcatspike at 9:25 AM on March 7, 2003
skoosh do you feel you were trolled into viewing the site for your info to be collected? See if you are collecting info on your site, but make an excuse for folks to come, I could see feeling baited.
Posting your site outside of the user page makes for slippery slope is what I'm seeing. I did it in a comment and really wished I hadn't but then again linking to my userpage would have accomplished the same thing w/o qualms.
posted by thomcatspike at 9:25 AM on March 7, 2003
PS, I feel baited too now, because I viewed the website to understand this thread, whoops skoosh;P
posted by thomcatspike at 9:27 AM on March 7, 2003
posted by thomcatspike at 9:27 AM on March 7, 2003
i think skoosh was confused because the page they looked contained something that referred to mefi (which they probably found a bit odd - why does this person have mefi-related stuff on their site? some kind of scam?). in fact the mefi reference is generated dynamically on that page from the info that browsers send when they connect (what ook is saying). in other words, if you connected from somewhere else you wouldn't see the reference to mefi.
[and now everyone's going to look at thomcatspike's blog.. ;o]
posted by andrew cooke at 10:32 AM on March 7, 2003
[and now everyone's going to look at thomcatspike's blog.. ;o]
posted by andrew cooke at 10:32 AM on March 7, 2003
was confused because the page they looked contained something that referred to mefi
Still lost in seeing it.
And yours too andrew because they have been said, as the vicious circle starts in explaining then leads. I just took a look at yours last week as you made a comment I agreed on, then thought: hey what more does andrew have to say.
No I was not trying to bait, that is why I did not link the comment that I explained plus I still regret it for personnel feelings on the subject.
posted by thomcatspike at 11:06 AM on March 7, 2003
Still lost in seeing it.
And yours too andrew because they have been said, as the vicious circle starts in explaining then leads. I just took a look at yours last week as you made a comment I agreed on, then thought: hey what more does andrew have to say.
No I was not trying to bait, that is why I did not link the comment that I explained plus I still regret it for personnel feelings on the subject.
posted by thomcatspike at 11:06 AM on March 7, 2003
tcs - if you visit their website following the link above, for example, then that page will (assuming you have a normalish browser) run some javascript (if you look at the source code for the gravlab site, it's where it says "OPENTRACKER SCRIPTCODE START"). that javascript makes your browser (without telling you) send some information about your browser and the site you last visited (in this case metafilter) to the gravlab server. if you're using internet explorer that will all be invisible to you (at least, it is to me using ie6 and default settings). but somehow skoosh noticed this (if you're using mozilla, for example, you can see it under View/Page Info/Media) and started wondering what was happening.
really it's not a big deal. nor is looking at user web pages ;o)
posted by andrew cooke at 12:19 PM on March 7, 2003
really it's not a big deal. nor is looking at user web pages ;o)
posted by andrew cooke at 12:19 PM on March 7, 2003
Thanks andrew, & agree.
posted by thomcatspike at 12:24 PM on March 7, 2003
posted by thomcatspike at 12:24 PM on March 7, 2003
Skoosh - the only "Mefi specific URL" from that "invisible image" is a dynamically generated refferrer URL so I can see where my hits come from. And that code was there long before I posted to Mefi.
And you can see what the perlscript/js dynamically generates here.
posted by jdaura at 2:01 PM on March 7, 2003
And you can see what the perlscript/js dynamically generates here.
posted by jdaura at 2:01 PM on March 7, 2003
So, let me get this straight, you're a web designer and you just found out you're color blind?
posted by yonderboy at 12:47 AM on March 8, 2003
posted by yonderboy at 12:47 AM on March 8, 2003
Sorry for not saying anything for a day; I only use MeFi from home, as a rule. I added a comment right after posting this, to the effect that I wasn't saying it was necessarily bad, but simply pointing it out and asking how people felt about it. Unfortunately, that comment disappeared somewhere along the way.
ook: No, I am not so paranoid as to read the source code of every page I visit. I only discovered the 1x1 pixel image because I have a dial-up connection. jdaura's page took about 15 seconds to load, and in that time, I got to see a placeholder image where the tracking image would eventually be. I was expecting an actual, visible image, and when I didn't get one, I started right-clicking around until I hit the right pixel, and copied and pasted the image's URL to my trusty text editor. At the time, it seemed odd that an image would have a script-generated URL attached to it, and even odder that it had metafilter's name in that URL.
Also, I must admit that I've never felt entirely comfortable with 1-pixel user-tracking images, though I'm not sure that I can explain why. I guess it's because I associate it with third-party commercial marketing companies who gather and compile the information to create user profiles. I've never run my own domain or my own server, and I've never had access to server logs, so I never think about the fact that all sorts of information is routinely exchanged and retained as part of any http request.
Thanks to andrew cooke and jdaura for pointing out that the metafilter.com in the image's URL was script-generated and referrer-dependent, and that the script was not put there just for the occasion of linking from MeFi. I especially appreciate the full disclosure from jdaura.
yonderboy: ditto, eh?
posted by skoosh at 5:06 AM on March 8, 2003
ook: No, I am not so paranoid as to read the source code of every page I visit. I only discovered the 1x1 pixel image because I have a dial-up connection. jdaura's page took about 15 seconds to load, and in that time, I got to see a placeholder image where the tracking image would eventually be. I was expecting an actual, visible image, and when I didn't get one, I started right-clicking around until I hit the right pixel, and copied and pasted the image's URL to my trusty text editor. At the time, it seemed odd that an image would have a script-generated URL attached to it, and even odder that it had metafilter's name in that URL.
Also, I must admit that I've never felt entirely comfortable with 1-pixel user-tracking images, though I'm not sure that I can explain why. I guess it's because I associate it with third-party commercial marketing companies who gather and compile the information to create user profiles. I've never run my own domain or my own server, and I've never had access to server logs, so I never think about the fact that all sorts of information is routinely exchanged and retained as part of any http request.
Thanks to andrew cooke and jdaura for pointing out that the metafilter.com in the image's URL was script-generated and referrer-dependent, and that the script was not put there just for the occasion of linking from MeFi. I especially appreciate the full disclosure from jdaura.
yonderboy: ditto, eh?
posted by skoosh at 5:06 AM on March 8, 2003
That was a very different situation, crunchland: in that case the image was placed on metafilter itself, so was tracking every hit to that metafilter page. Evil.
You ain't kidding.
posted by oissubke at 9:11 AM on March 8, 2003
You ain't kidding.
posted by oissubke at 9:11 AM on March 8, 2003
You are not logged in, either login or create an account to post comments
Seems kosher enough to me. If he wants to track visitors to his own website, that's his business; the information this captures is no more invasive than that gathered by every server log on the net. My opinion, feel free to disagree.
Out of curiosity, how did you find this? Do you routinely view source on every page you visit, or what?
posted by ook at 4:58 AM on March 7, 2003