Hacker News Reader

Comment by netfire | Hacker News Reader

http://www.amazon.com/b/ref=br_imp_ara-1?_encoding=UTF8&node...

That's a horrible url. Why not use something like http://amazon.com/news/free-prime-same-day-delivery? Is there some significant advantage to identifying the page with non-human readable identifiers in the query string that I'm not aware of?

Otherwise, very cool. I just wish it was in more areas.

View on HN

> Is there some significant advantage to identifying the page with non-human readable identifiers in the query string that I'm not aware of?

Sure. When your organization uses a custom CMS and creating pretty URLs requires approval from multiple departments and takes days (or weeks) and you want to throw up a page all quick-like.

Not sure why you would need approval for a pretty URL and wouldn't need approval for the design and content in general. Couldn't those be lumped together for approval at the same time? It just seems subpar for a company of Amazon's size, especially when most CMS platforms have support for pretty urls.

Not sure that the design for this page went through any sort of approval process though :)

At a big company, you're getting multiple approvals from multiple people in different areas of the company. If it was as simple as "Get the approval from the President" that'd be one thing, but there's most like a team in charge of just URLs and another team designing the app who would need to go to the URL team for approvals. Another team will most likely be the gatekeeper for whether it goes to production or not, and there's probably at least a half dozen other infrastructure teams I haven't mentioned also involved in the process. I haven't worked for Amazon, but I've worked for another large company, so I know what it's like.

Most of the other teams involved that I've seen at places are usually other development teams who are consulted not because they have any specific reason but because as you go up the chain they want to make sure that nobody is already planning to use that same name/pretty url for something to avoid conflicts where one team all the sudden has two days notice to change their branding/url.

http://www.amazon.com/sameday

Which still gets redirect to:

http://www.amazon.com/b?ie=UTF8&node=8729023011

Interesting that they have a pretty url already generated for the page and then redirect to something that's completely unreadable. It seems like you'd want to have a single, readable url to have people sharing with others, if they copy and paste the url from their browser.

Why do that when they can create a unique URL And data mine who you share the link with or where you post it? People call it "dark social" where there are no referrers such as IM, Text, or chat services like Slack, Hipchat, etc.

They do exactly this for their products, if you look in the URL you'll see /dp/something if you do Amazon.TLD/dp/something it takes you to the product. Someone mentioned dark social and that's exactly what it's for according to one of the Edinburgh Amazon devs. :-)

It can be significantly simplified by removing the junk:

http://www.amazon.com/b/?node=8729023011

The rest of the stuff is probably used to track how you got to the current page.

Still completely meaningless from a human-readable perspective (you know, unless you memorize page node identifiers :) )

"That's a horrible url"

I clicked on it and it worked....seems fine to me.

I imagine few people use Safari around these parts, but it shows you only the domain, cutting off the rest of it. So you see just "amazon.com".

Makes you somewhat sad to think the future is going to be just people clicking on links, never actually typing in a full address by hand :)

Posted this below: Why do that when they can create a unique URL and data mine who you share the link with or where you post it? Publishers call it "dark social" where there are no referrers such as IM, Text, or chat services like Slack, Hipchat, etc.

Amazon's had the same garbage URLs for years. They must be consciously insensitive to the URL experience--and/or it's so deeply embedded they don't care enough to change.

It's http://www.amazon.com/sameday

Which immediately gets redirected to:

http://www.amazon.com/b?ie=UTF8&node=8729023011

True. Amazon's CMS is pretty old-school, but let's not blame Java like one of the commenters did earlier.

Reply on news.ycombinator.com