Cool URIs Don't Change (1998)

149 pointsposted 8 hours ago
by Klaster_1

29 Comments

torh

4 hours ago

Not that long ago I clicked on a link that Microsoft provided somewhere in Windows -- could have been the event log, I don't remember. I do remember it was for a specific support article, but it ended up at a generic landing page for something. I am not talking about a link from Windows 95, it must have been Windows 10. But regardless, it is all gone. Then again, it probably wasn't a cool URL to begin with.

It's amazingly many news sites that also seem to scrub their URLs every time they do a redesign. Also uncool.

phillipseamore

an hour ago

Got an MX4 yesterday and it has a printed card with "Go to logitech.com/mx to get started" resulting in a "Forbidden" error page. This product hasn't even been out for a year!

phatskat

an hour ago

I bought an air conditioner last year, some wal-mart “smart” AC that was affordable (we had to get three for the new house we moved in to so budget was a big concern) and it has a QR code for the companion app. I finally got around to getting the app this summer and lo and behold, it’s a dead link. Yay.

sixtyj

3 hours ago

Average time of url is 5 years.

Therefore I try to make a local copy of a document that I link to.

There a plugin for Firefox that can check bookmarked links. You will be surprised how they vanish pretty fast…

Panino

an hour ago

Mailing list archive software is a constant rug pull: today you link to a post about something relevant, and next month the link goes to a different message about a different subject.

beached_whale

3 hours ago

Microsoft has a terrible habit of doing that. KB's used to have a predicable URL, they may again, but they keep changing them. Or any of their docs really. It's terrible as those URL's end up in other documents to be killed. This is not really an enterprise friendly way of supporting customers.

toyg

2 hours ago

Yeah, I always wondered how one of the wealthiest companies on earth could be so poor in linking to their own documentation.

Then I realized that they have a galaxy of consultants and experts that effectively make a living by keeping on top of all issues and fixing them for a fee. So there is little incentive for them to actually be user-friendly: as long as they are "partner friendly" with timely material, that's good enough.

They are not alone in this, btw - pretty much all major platform providers do this, one way or the other.

mikepurvis

5 hours ago

Sadly, though:

    $ curl -I https://www.nsf.gov/pubs/1998/nsf9814/nsf9814.htm
    HTTP/2 404 
    content-type: text/html; charset=UTF-8
    content-length: 54737
    date: Sun, 09 Aug 2026 17:58:44 GMT
    strict-transport-security: max-age=31536000; includeSubdomains; preload
    server: Apache

zibw

5 hours ago

A classic. This keeps getting more credible as it ages. Now it's been at the same URI for 28 years.

firasd

5 hours ago

What this page doesn't mention is 301 or 302 redirects. SEO has made "old URLs staying live" more of a widespread concern than it was at the time. And WordPress etc ship with inbuilt redirects upon slug rename

So to a large extent this has been mitigated and not using the suggestion here, which is to create a permanent URL ontology upfront

That said eventually neglect, removal, reorgs (or simply websites going offline) still happens... but the way the suggested goal has been advanced is due to becoming a business priority and with redirects and CMSes as tech to help

That said one suggestion made here turned out to be very useful and indeed is the default in WP:

   http://www.w3.org/1998/12/01/chairs
If you use date as part of the taxonomy then -- as Tim BL says here:

> A reason for using a topic area as part of the URI is that responsibility for sub-parts of a URI space is typically delegated, and then you need a name for the organizational body - the subdivision or group or whatever - which has responsibility for that sub-space. This is binding your URIs to the organizational structure. It is typically safe only when protected by a date further up the URI (to the left of it): 1998/pics can be taken to mean for your server "what we meant in 1998 by pics", rather than "what in 1998 we did with what we now refer to as pics."

bensyverson

5 hours ago

Yes but how often in practice does the same information stay accessible via 30x redirects after one reorganization, let alone two or three?

EGreg

4 hours ago

Sounds like most URIs have been downright uncool, then!

mooreds

8 hours ago

Great article, lots and lots of past discussion: https://hn.algolia.com/?q=cool+uris

WalterGR

5 hours ago

This post is related to (inspired by?):

"The original URL for this prediction will no longer be available in 11 years (2011)" (longbets.org)

https://news.ycombinator.com/item?id=49228458

186 points | 12 hours ago | 91 comments

sunaookami

an hour ago

Always like it when Hacker News "follows" a topic over days with different related links. Kinda like a bonus for reading HN daily :)

namanyayg

3 hours ago

OP is from 1998 so it's the other way around

WalterGR

3 hours ago

I meant the post/submission itself, not the content.

Swizec

an hour ago

I've been trying to keep all URLs backwards compatible forever on swizec.com and it's surprisingly hard. I _think_ anything from 2010 onward should be good. The domain started in 2005.

Unfortunately I'll never know for sure.

glacials

3 hours ago

If you run a statically generated website, I recommend append-only generation, where you keep your `dist/` (or whatever web root) stateful between builds. This guarantees you don't break URIs even if the static generator changes or the source content is removed.

You can even make an orphan branch and check out that branch into `dist/` as a worktree to keep it under version control.

cxr

2 hours ago

Even for people who don't practice this, if you're publishing a static site, you can make sure you're generating output with rel=canonical links and at any given point, you should be able to run a tool (entirely separate from your chosen static site generator) over the output to extract those URLs and verify at time t2 that the set of links that existed at t1 (where t1 < t2) is a subset of those at t2.

ricardo81

an hour ago

I recall some link rot research from ~10 years ago that estimated it was in the high single digit percentiles per year.

archive.org perhaps have/could have some interesting insights on it.

nicbou

2 hours ago

This is something I work really hard to maintain on my websites. I often refer to government websites as sources for my guides, and these URLs break at an alarming pace. It's as if the German government is moving reference pages and services around just for giggles. It creates a significant maintenance burden on my end.

When I worked for an ecommerce website, we spent weeks making sure every URL worked after a migration.

Nowadays it feels a little quaint because most of my 404s are from LLMs hallucinating URLs that never existed, sometimes for topics I never covered. I wonder what nonsense it attributes to me.

jruohonen

6 hours ago

It is not only about "coolness" but long-term preservation too; that was one of the main motives behind DOIs.

shiandow

5 hours ago

The headers that push a bit into the margin are an interesting stylistic choice, but why on earth does it have to be random? Did something change that affected the layout somehow? (ironically)

al_borland

4 hours ago

I find it works well for helping maintain context. I often can't easily tell the difference between different headers unless they're right next to each other, and even then some of them are difficult to discern. Adding the indentation is a way to make the hierarchy more clear.