Broken link experience on Downloads page due to tracking


#1

Was trying to click through a couple of links on the Downloads page:
https://downloads.chef.io/chef-dk/mac/

Links to the docs site to “learn more!”, and some other offsite links are wrapped in a JS call to onclick='trackOutboundLink'..., which is a Google Analytics tracker, as far as I can discern.

When using browser extensions that prevent such tracking, the user experience is poor, as clicking the link goes nowhere, and the user loses the wonderful happiness of being able to visit another site.

I’m no web dev, but I think the call to the tracker could be wrapped in a JS conditional, something like:

if ('ga' in window) {
    ga('send', 'event', 'outbound', 'click', url, ...

Thereby allowing an off-site click to work when GA isn’t found/loaded.

The downloads.chef.io site does not appear to have an open repo, so I was not able to submit a PR for this change.

-Mike


#2

This is a weird one for me. On a personal level I am bothered by the premise. What I am reading is that you are actively doing something to interfere with the normal operation of the site and upset that it doesn’t work as a result of blocking said normal functionality… As a community of sysadmins, developers, and operations engineers, it is interesting how we have made it culturally acceptable to interfere with ability of our counterparts at other organizations to collect site metrics. It provides a real interesting discussion around how we make decisions.

With my personal thoughts about metrics and stuff out of the way, I did go ahead and create a ticket on the internal system.

Thanks for the report!

– cwebber


#3

To flip the concern, the way we’ve made it culturally acceptable for organizations both private and public to track our every move on the web, leaving draconian measures the only resort for users.

Looking forward to having this corrected.


#4

Chris,
Thanks for filing it internally.

I’m sorry this bothers you - I don’t see it as me interfering with the
normal operation of a site, rather the site not handling degradation
gracefully.

In the event an external script cannot be downloaded, external hyperlinks
do not function, which breaks the user experience.
How I go about preventing a tracker is my concern, but a site continuing to
operate in the face of a third-party’s problem is pretty important in my
book. Consider if the tracker code goes offline for an outage. Should your
site be adversely impacted, harming your user experience?

I realize that we can go heavy into philosophy around privacy, tracking,
and more. The practice of using methods to protect oneself online from
undesired tracking (cookie blocking, ad blocking, ad nauseum) are prolific
enough to warrant ensuring that core services provided by a web site -
namely a href links that have been around since the start of the web -
continue to function.

Again, I’m less concerned with the tracking than I am with the broken user
experience.


#5

Mike,

Love the response. I think that is why I actually went on a limb to discuss it. It is such an interesting systems topic that is frequently overlooked. Because there are actually a few interrelated discussion… One of them is very much the thing you are concerned about, does the site respond well given a failure scenario, another around what I was bring up dealing with the ethics of actively disabling someones ability to do basic metric collection, and yet another around the societal impacts of the massive data collection that is going on.

The cool part is that we can actually break this out and understand implications. Thank you for expanding on your point. I will definitely bring that into the conversation. As we (the proverbial we) discuss more and more about the systems we run, it is interesting to see the intended and unintended consequences of actions.

The most interesting part for me is to see how the reframing of the question changes the response.

Thanks again for understanding my point of view.

– cwebber


#6

Mike,

I had a few to mess around with this, but cant seem to replicate it when blocking GA using Ghostery in Chrome. Is there some other mechanism you are using to block GA? Would love to get a repro case.

Thanks!

– cwebber


#7

Hi,

I’m able to reproduce this with uBlock Origin 1.2.0 in Chrome; however, I haven’t figured out what the minimal set of checkboxes in the options are required.

Cheers,

Steven


#8

The behavior can be triggered using NoScript in Firefox by blocking the googleanalytics.com domain. Temporarily allowing that domain for JS allows the links to work.


#9

We shipped a change that should fix this. Can someone confirm that, that is the case?

– cwebber


#10

I have uBlock Origin running in my browser. Earlier today, the link within “Learn more about ChefDK on [the Chef Docs site]” did not work with Google Analytics blocked. The link works for me now, still with GA blocked.


#11

Thanks, everyone. It appears working for me as well.