5 ms·

History of the browser user-agent string (2008)

38 points by jtchang 10y ago

yuhong 10y ago
Note that the history of the WebKit/KHTML user agents is more complex than the article indicates.
yagop 10y ago
And then Microsoft built Edge and tries to be Chrome too with UA as Mozilla/5.0 (Windows NT 10.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/42.0.2311.135 Safari/537.36 Edge/12.10136
- glandium 10y ago
  The best thing that could happen is that Chrome would pretend to be Edge too, that would definitely kill UA sniffing for good, because there would now be two distinct browser engines with the same user-agent.
ikeboy 10y ago
This seems to date from 2008 https://web.archive.org/web/20080904080055/http://www.webaim.org/blog/user-agent-string-history/ https://web.archive.org/web/20080904080055/http://www.webaim...
- dang 10y ago
  Ok, we added 2008 to the title.
oulu2006 10y ago
This is great, I often wondered why the user agent strings are so the same between different browsers.
mh-cx 10y ago
Hilarious. To me the similarity to evolutional processes is striking. I mean, didn't biological evolution also give rise to sometimes bizarr creatures? They somehow coevolved with the other inhabitants in their niche. Still they live their (sometimes not so) happy lifes. What a weird world this is.
ErikAugust 10y ago
And to think Larry Page wanted to be a "good net citizen" with his user-agent string: https://groups.google.com/forum/#!msg/comp.lang.java/aSPAJO05LIU/ushhUIQQ-ogJ https://groups.google.com/forum/#!msg/comp.lang.java/aSPAJO0...
lazyjones 10y ago
Since browser vendors are clearly out of their minds, it's up to users to end this mess and save a few bytes of traffic in the process: just set your user-agent string to something short (most browsers allow it without add-ons).
- takeda 10y ago
  Not exactly what you are suggesting, but uMatrix has an option to once in a while switching user agent to random one from the list. This is implemented for privacy reasons. Sadly there still some websites where you want to disable it though, because they block some features otherwise.
jimktrains2 10y ago
I've always thought the UA header should just be removed. Sniffing is a bad thing and analytics on it are mainly used to preemptively decide not to use features that you should just use feature detection for (and for tools like Google Analytics comes from the js lib not the UA header anyway). Plus it's so many wasted bytes.
therealidiot 10y ago
I use an extension to fake my UA to that of other browser+OS combinations. It's silly how many sites serve you different things based on the User-Agent still. Google doesn't even work properly if I don't identify my browser - the results show but clicking them does nothing. I still fake the version and OS though.

38 points by jtchang 10y ago

Note that the history of the WebKit/KHTML user agents is more complex than the article indicates.

yagop 10y ago

And then Microsoft built Edge and tries to be Chrome too with UA as Mozilla/5.0 (Windows NT 10.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/42.0.2311.135 Safari/537.36 Edge/12.10136

glandium 10y ago
The best thing that could happen is that Chrome would pretend to be Edge too, that would definitely kill UA sniffing for good, because there would now be two distinct browser engines with the same user-agent.

ikeboy 10y ago

This seems to date from 2008 https://web.archive.org/web/20080904080055/http://www.webaim.org/blog/user-agent-string-history/ https://web.archive.org/web/20080904080055/http://www.webaim...

dang 10y ago
Ok, we added 2008 to the title.

oulu2006 10y ago

This is great, I often wondered why the user agent strings are so the same between different browsers.

mh-cx 10y ago

Hilarious. To me the similarity to evolutional processes is striking. I mean, didn't biological evolution also give rise to sometimes bizarr creatures? They somehow coevolved with the other inhabitants in their niche. Still they live their (sometimes not so) happy lifes. What a weird world this is.

ErikAugust 10y ago

And to think Larry Page wanted to be a "good net citizen" with his user-agent string: https://groups.google.com/forum/#!msg/comp.lang.java/aSPAJO05LIU/ushhUIQQ-ogJ https://groups.google.com/forum/#!msg/comp.lang.java/aSPAJO0...

lazyjones 10y ago

Since browser vendors are clearly out of their minds, it's up to users to end this mess and save a few bytes of traffic in the process: just set your user-agent string to something short (most browsers allow it without add-ons).

takeda 10y ago
Not exactly what you are suggesting, but uMatrix has an option to once in a while switching user agent to random one from the list. This is implemented for privacy reasons. Sadly there still some websites where you want to disable it though, because they block some features otherwise.

jimktrains2 10y ago

I've always thought the UA header should just be removed. Sniffing is a bad thing and analytics on it are mainly used to preemptively decide not to use features that you should just use feature detection for (and for tools like Google Analytics comes from the js lib not the UA header anyway). Plus it's so many wasted bytes.

therealidiot 10y ago

I use an extension to fake my UA to that of other browser+OS combinations. It's silly how many sites serve you different things based on the User-Agent still. Google doesn't even work properly if I don't identify my browser - the results show but clicking them does nothing. I still fake the version and OS though.