It's that last question that's bothering me; from an idealistic perspective I would say that it is never ok to exclude anyone, but from a more pragmatic perspective I would answer that question with another question: how big a part are we talking about? Only with the answer to that question can you make a sound decision.
But before I present you the details of my own investigation, let me ask you the question:
Let's see if the outcome of this poll matches my result:
Now the function setJsCookie was defined in one of our external js-files:
document.domain is always explicitly set to 'tweakers.net' to allow cross-subdomain cookies and trackingpixel.php (not the real name) is a script that validates the user's session (based on a session cookie - excluding bots) and writes an entry to the database incrementing either a 'js'-field or a 'nojs'-field based upon the presence of the 'js-enabled'-cookie. I used the session id as a primary field and also logged the user-agent string and whether it was a logged-in user or not. This was done randomly for 1 in every 100 pageviews during the timeframe of 10:30 to 20:30 (so 10 hours in total).
My first question would be: is this a sound and safe approach? The results only showed a very small fraction that reported both 'js' and 'nojs' on the same session id, so I disregarded those from the results. Still, the results were stunning:
Update 28/10 11:25
So I use noscript. I block the domains which the big ad's come from, and allow the domain of the site I'm on. Which means I get some ad's, but seldom annoying, and if they're annoying I generally leave the site.
Buildhow do you built this feature
[Comment edited on Thursday 28 October 2010 09:08]
For the record, I don't disable JS, because it reduces my experience of most websites... and probably because I work with a set of websites that depends on JS to function properly (as in, it uses JS to submit forms to do things like order a product. Rancid.)
It appears to be a data-mining service, tracing your every move on Belgian sites...
After i blocked it, even though i almost exclusively visit english/american sites; the metriweb filter has slightly less hits than */ads* or *pagead2*
It's in my top-three right now with 14.311 blocks/hits (!)
I don't know what éxactly that metriweb does, but noscript and adblock sure make the the web a little more private and safe
The problem in the above described method is that there's two dependencies, instead of one. You're not just checking if JS is enabled, you're also dependent on the agent accepting cookies.
As said I will rerun this test using a different approach to see if I get matching results
[Comment edited on Thursday 28 October 2010 10:16]
@ZpAz: Well, the last time I checked CSS wasn't executable code responsible for almost all security issues in web browsers.
Je kunt als je de site veel gebruikt natuurlijk ook een abootje nemen.Battle Bunny wrote on Thursday 28 October 2010 @ 11:29:
Ik heb helaas op Tweakers.net ook weer adblock plus aan (stond in de whitelist), maar als je van die flitense flash banners krijgt waardoor je een artiekel niet meer kunt leze nga je imho te ver als site.
[Comment edited on Thursday 28 October 2010 12:10]
Volgens mij zijn er veel mensen die er puur ads mee blocken maar zijn er ook verdacht veel mensen welke het uit hebben voor privacy gedoe enzo.
[Edit] ik ken een aantal mensen die niets van computers en Windows afweten, maar wél Firefox hebben ingesteld om aan het einde van de sessie de cache enzo op te ruimen. Als dat soort mensen iets lezen over een browser lek en NoScript enzo, zullen het waarschijnlijk ook installeren).
[Comment edited on Thursday 28 October 2010 14:26]
Maar dat zal inderdaad wel niet de belangrijkste doelgroep zijn voor de meeste sites.
Comments are closed