When the Web fails, what went wrong? Regular people can rarely figure it out.

Continued">

You go to a Web site, perhaps for a large company or retailer. When you get to what you think you’re looking for, it just stalls and nothing seems to be happening. What went wrong?

The nature of the web and the internet makes that a very hard questions to answer.

Just off the top of my head, the answer could be any of the following:

  • Your device has stop functioning.
  • Your device is still working, but can’t respond to input from the user.
  • Your browser stopped working.
  • Your WiFi isn’t working.
  • The WiFi is working, but you’re in a spot where it can’t deliver a consistent signal.
  • The internet service wherever you are (possibly at home) isn’t working.
  • You typed the name of the site wrong.
  • You clicked on a link that doesn’t actually link to the site you’re seeking.
  • The site is down.
  • The site is working, but something within its internal workings isn’t functioning properly, like a database that’s not communicating with the Web interface.
  • The site is working, but the page you were trying to see isn’t working.
  • The site or page isn’t compatible with your browser.
  • The site or page doesn’t render properly on your device.
  • The site didn’t update its security certificate, so your browser says it’s unsafe.
  • The site was hacked or is under attack, such as a DDOS attack.
  • The site’s host, such as Amazon Web Services, is down.
  • Some element embedded in the page but served from some other place, such as an ad, isn’t able to render properly.
  • It actually is working, but you didn’t wait long enough.

That’s 18 different possible reasons, all of which generate the same result — a site that doesn’t seem to be responding. And it’s far from a complete list.

You may know to use tools to check some of these issues. Point your browser at fast.com to see if your connection is working and how fast it is. Go to downdetector.com or isitdownrightnow.com to see if the site is up. Do a web search to see if there are outages happening at your internet service or in the infrastructure that powers sites.

But for less sophisticated users, the lack of clues must be mystifying. Even sophisticated users waste a lot of time debugging these probems.

Why can’t the Web be like a car?

Just as with the Web, most drivers of cars couldn’t tell you how the car actually works. But we’ve gotten pretty good at basic diagnoses. For example:

  • Car turns over slowly when starting, then gives up: probably a battery problem.
  • Temperature gauge rising, smoke coming up from under the hood: Radiator.
  • Car revs unevenly when you step on the gas: transmission.
  • Loud noises from the back of the car: muffler.
  • Dark smoke coming out back of car: engine or exhaust issue.
  • Car very bouncy or squeaky: suspension.
  • Uneven bumping sounds that speed up when the car goes faster: tires.
  • Won’t stop or makes squealing noises when trying to stop: brakes.

But the clues the Web gives are too subtle to allow this sort of discrimination for the average user.

Is there any way we could make our devices, browsers, and sites give better clues about what’s not working?

I have no idea. But I’d love to hear from people who do.

One response to “When the Web fails, what went wrong? Regular people can rarely figure it out.

  1. I’m a software developer. In my experience (little as it is, I’m a young guy), developers are too busy with trying to get it working than communicating failures properly.

    When you go to a web page, you are sending a request to a website which can involve 10 or 100 different steps, and each can go wrong for some reason or another. In a previous project, we were asked how many errors could there be, and we said: each line of code can result in an error. There’s not enough budget to cover those situations, so the solution was to ship the product and improve the handling of errors as they started happening and we got complaints.

    Internally, each request generates a response which has some code for success or failure. 200 means success, 400 user failure, 500 server failure. In this case, 503 means the service is unavailable, which probably means the website is down. Maybe the browsers could read this and display the errors?

    There’s also some things the user can do, such as trying to browse a different website (means the site is down, not their connection), refreshing the page, logging out and in.

    Reference: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.