Unsung heroes: Flickr's URLs scheme
unsung.aresluna.orgFlickr deserves a lot of praise for a number of technical advances that I wish had seen wider adoption. Their API was one of the first and honestly still one of the most enjoyable to actually use as a developer. It's still full of incredibly interesting API calls that you wouldn't expect from it unless you read carefully. Did you know, for example, that flickr API will provide you with the bounding box co-ordinates of different types of places? From a neighbourhood all the way up to a continent?
They implemented the Where On Earth ID (WOEID) which was a super useful way of disambiguating different places that shared latitude and longitude (for example, being able to disambiguate the Sydney Opera House, Circular Quay and Sydney Harbour which all can potentially share the same lat/long co-ords).
They implemented machine tags which are tags in the form of -
namespace:predicate=value
Which, when it was implemented by other sites with machine tags allowed you to get and group all kinds of interesting combinations of content.
Yeah, honestly flickr had some incredible tech the was so much fun to explore and use. That their vision of what the web could be wasn't the one that won is one of the great losses of the web IMO.
Don't forget their "interestingness" algorithm that would determine which photos get to the top position in groups and other shares views. Way before ML ranking.
Funnily enough, I think it was also one of the early warnings of what's coming. The problem with the Flickr algorithm was that it ended up promoting a very predictable style of photos - mostly HDR landscapes - which in turn encouraged more photographers to emulate it. In its early years, it was a fantastic place to learn photography. In the "final" years, it was extremely homogenous and, frankly, boring.
Yes, I remember gaming the algo to get to the top page! HDR was much harder to produce back then lol
You should write all that up! (I've been around for a while and I never used Flickr except as a casual, but I did know about a lot of mashups and even spent time using FlickrLickr for Wikipedia; yet I've never heard of... any of that.)
> Flickr deserves a lot of praise for a number of technical advances that I wish had seen wider adoption.
Flickr + MechanicalTurk == ImageNet
The priceless legacy of Web 2.0 golden age
They were also publicly in "beta" for a few years, which provided a great example that it was okay to roll out experimental features to the public.
> I would also try to add a human-readable slug at the end, because…
No? Because what would it be based on and if you edited the thing that it's based on then the URL would either change, or get out of sync which woudl suck. You could ignore the suffix meaning flickr.com/mwichary/sets/72177720330077904-<everything-past-the-previous-dash-is-ignored> I'm not sure if that would be a positive, although I guess S.O. does something like that. The issue is other sites really want to know if it's a link to the same resource or a different resource. And while you could redirect to the new one that just makes more work for everyone.
> I would get rid of /photos
I wouldn't because then you'd had have https://flickr.com/settings but that would not be a user named "settings" and the same for every other alternate purpose URL
> The issue is other sites really want to know if it's a link to the same resource or a different resource.
Thats what the canonical link is for, isn't it? [0]
RFC 6596 introduced it in 2012. Other websites, like search engines or social media, have been using it for a while.
[0] https://developers.google.com/search/docs/crawling-indexing/...
Yes, and? That’s a boil the ocean solution. Every app and site needs to look up urls and update where as they didn’t before.
Most apps and sites will have a thumbnail for the site. So they're already loading the page.
Other sites grab contact information or verification headers. So they also load the page.
Seems like the only people who wouldn't want to load at all, would be those places where deduplication doesn't matter in the first place.
> You could ignore the suffix meaning flickr.com/mwichary/sets/72177720330077904-<everything-past-the-previous-dash-is-ignored> I'm not sure if that would be a positive, although I guess S.O. does something like that.
That's usually how people do it.
Yes, I saw that multiple times in blog engines, with the entry title being the optional part.
I've seen these types of things break because of characters used after the dash separator would be deemed illegal by some part of the chain. However, if you delete everything after the separator just leaving the ID, the page would load. So some URL parser is choking because someone forgot to encode the URL somewhere upstream.
Just another reason to say who cares to human readable bits in the location. Most browsers hide that data anyways.
What I typically end up doing is just recalc the slug and see if it matches the provided one. If it doesn't redirect to the most up to date slug matching the id. Though who knows if those old SEO patterns still matter these days...
I don't know how much Flickr is used these days, but I remember it was quite popular some 15 years ago. I decided to search around, and discovered that it is a treasure trove of photos from the period 2005 - 2015, and incredibly easy to search / filter.
Internet archeology is something I've always found fascinating, and I don't think people realize how much data has been lost after we moved to the modern "big tech" internet of today. So many data hosting services disappear back in the mid/late 00s, and with that, the data too. After social media exploded, many just stared storing all their photos there.
Interestingly, Washington state DOT uses it to post pictures for construction progress updates to this day. It’s great!
The government in British Columbia, Canada does as well - https://www.flickr.com/photos/tranbc/
It still has plenty of active users (myself included) but it feels much smaller than the 2010s. The groups are great if you can find ones that match your taste, and it’s way more fun to do that and share you photos with human-curated communities than with the algorithmic feed.
I agree 100% with the author, clean, easily readable and well structured URLs make the web a better place. URL is a hierarchical structure as introduced in the RFC1738 by a guy you might have heard, Tim Berners-Lee, the inventor of the World Wide Web :-) https://www.rfc-editor.org/rfc/rfc1738
Easily readable URLs is something I learned in the 90s and I still try to enforce in everything I create.
This was a big thing around the time of Flickr and, if my memory is still working, del.icio.us. There was a push for “url is the new command line” which I wish had gotten more traction because there were some fun things happening. I feel like Yahoo Labs was involved in a lot of it, including Pipes and their ahead of its time JavaScript framework. It’s a strange sliding doors moment for me because I turned down a job there right around this time.
I'm a big fan of writing little bits of code into my URL routers that check for pages and try to correct typos. So if someone types https://some-awesome-site.org/jhon-davisdon it will check and correct it to /john-davidson. What's nice is always delivering the "canonical" link rel when you serve the correct page that way, too. I make the assumption that people still try to type links, sometimes ;)
Sounds like mod_speling for Apache: https://httpd.apache.org/docs/current/mod/mod_speling.html
Hah. Cool. Similar. But I implement it usually by having the whole page hierarchy in a SQL table and using wildcards.
Concise URLs deserve more praise.
Also, when you look at a site and see URLs like /wiki/index.php/MyPage it tells you about the skill level and care of the site administrators.
And a bit about the skill of whoever made the wiki software, they need better documentation and automation to help even less-skilled admins to have clean URLs
They do have a guide for it
Ironically XWiki doesn't use their own short URLs: https://www.xwiki.org/xwiki/bin/view/Documentation/AdminGuid...
We have one internal IT team who leaves things on non-standard ports, like https 8443 with no redirects on the standard ports. Drives me mad.
Unfortunately things are going in the opposite direction with media platforms creating an encoded blob impossible to edit by hand so that you (or a tool) cannot strip tracking etc.
> (Alternatively, I would consider getting rid of numerical ids altogether and relying on name alone. Internet Archive does it at e.g. archive.org/details/leroy-lettering-sets, but that has some serious limitations that are not hard to imagine.)
I could try to imagine these limitations and how the Internet Archive overcomes them, but I'd prefer reading about it.
I'm always surprised that archive.org's URL scheme is still used. To this day uploads will have URLs like `/details/hey` and IIRC the only thing to distinguish them from other submissions with matching names is the auto appending of a digit. It feels like `New Folder (1)` for URLs.
Edit: I misremembered. They append the date to the URL to avoid name conflicts, like `/details/hey_20260122`.
I think Youtube got it right early. Make it short enough but random, so users won't be tempted to manually type it out and make mistakes but if they really do need to manually type it the length is reasonable. It also made sharing via SMS/Twitter limits more feasible.
Flickr was a hero, then yahoo/smugmug killed it. It's still there, but along the way all changes reduced it to an also-ran. It's still a nice tool, but I just don't see myself using it again. The URL scheme, as neat as it was, I never noticed or cared to hack at. I just wanted to upload photos.
> I just wanted to upload photos.
Flickr is an unsung hero in this. I uploaded photos back in 2011 when I purchased my first DLSR camera, and had forgotten about them until this day, and seems they're still up! Did some other checks for content I uploaded back then, and seemingly only my YouTube and Vimeo videos are still up, everything else I spot checked from the same period seems to be gone by now.
Kind of neat for a free photo hosting platform.
> then yahoo/smugmug killed it
I understand the Yahoo part, but what do you mean with smugmug? My impression was that they bought it and "revived" it but I might misremember the history there.
They bought it but the revival was just the corpse moving around from heavy handed cpr attempts. It still lives thanks to the guy from Pinboard, but only as an archive really.
Are you confusing Flickr with Delicious (Which was what Pinboard guy bought from Yahoo) or am I missing something?
you are confused :) you are thinking about del.icio.us
flickr is alive and well
Revived might be too strong a word, it's still a shadow of what it used to be. Besides, wasn't it Smugmug that introduced the 1000 photo limit for free users? I too could be misremembering here.
I'd say that's hardly the fault of the acquisition but more the change of the landscape for photo communities. Everyone moved to Instagram and online storage space alone isn't such a big selling point any more as everyone has their image in Google Photos / iCloud already.
That's like saying Tumblr/MySpace got ruined by acquisitions, in reality people just moved on to other platforms like it always happens with social networks over time.
> I just wanted to upload photos.
You can absolutely still do this. I'm still a Flickr Pro subscriber since 2015, and I still regularly upload photos to Flickr. I don't think there was a set of changes that reduced Flickr to an also-ran, the entire market shifted. First, there was a shift away from photography being focused on what I will short hand as "quality" towards being focused on what I'll short hand as "moments" with services like Instagram, which had 100m users by the end of 2013 and continued growing exponentially from there, which was deeply interconnected with the introduction of reliable fairly high-quality phone cameras built into smart phones.
Flickr was, and continues to be, a place where people who use actual camera equipment post photos that are taken not just to capture a moment, but to express a scene, using technique and artistry to do so. That type of high quality photography doesn't really get much traction in more contemporary social media, because the photos of moments shared on Instagram weren't about the photo, they were about the moment. It was about proving that you were in a place or experienced a thing, and the place or thing giving you social value. "Pics or it didn't happen."
Instagram has now largely been supplanted by TikTok, because short video is now much more of a common, engaging, and desired format than photography, and thankfully this means Flickr in 2026 is once again a refuge of die-hard photographers sharing their works, and not seeing much attempt to change it into Instagram 2.0. Many (maybe most) of the photos on Flickr are now taken with smartphones, but there is still an expectation from the community to focus on expressing a scene using technique and artistry, and modern smartphones now have good enough cameras to do just that without detracting from what you're trying to express.
The shift from URLs accessing resources on file systems to more abstract resources (implicitly HTML unless the headers said otherwise) occurred around 1999/2000. Suddenly we were all doing it once we’d figured out the necessary Apache directives. It wasn’t just Flickr, although it and its APIs were a good example of clean URL design
However, getting rid of the /photos prefix would be a terrible improvement.
Having the /{username} at the root of the routing logic means that every URL should either query the user database for a match or use /{username} as a catch-all fallback if no other patterns match. But this makes resolving real 404 pages much more expensive.
If you delete the photos path element they do actually have a fallback resolver that redirects you to the appropriate path under photos.
E.g. flickr.com/mwichary/sets/72177720330077904 does take you to flickr.com/photos/mwichary/sets/72177720330077904
The flip side is enumerable IDs. Back when I was scraping a site for a side project, sequential photo IDs were basically a free sitemap. YouTube's random-ish IDs aren't just branding — they at least make bulk harvesting annoying.
You cannot afford to "bulk harvest" YouTube.
Bulk harvesting doesn't necessarily mean downloading every video, there is still useful metadata. But in this case non-enumerable IDs are on purpose, it would defeat the "unlisted video" system.
I thought Photobucket did it better, with the exception of having to know which server an individual's bucket was on.
It's been a long while, but if I recall, the url schema was something like a00.photobucket.com/albums/username/someimage.jpg
But what was really cool about it was that you could change someimage.jpg to someimage.png and Photobucket would serve a PNG instead. Or you could change someimage.jpg to th_someimage.jpg and Photobucket would serve a thumbnail of the picture. It was very cool.
I use a couple of JS libraries to serve image galleries, and they conform to the Flickr _[a-z] type of naming to indicate resolution. The library will load the appropriate size based on the available real estate.
The auto serving of the chosen output extension is interesting. Wonder if they were automatically created when making the various image sizes, or only on the first time it was requested. And how many formats were supported? The one that got me was hitting an API and tagging with xml|json to get the response as needed instead of sending it as a dedicated query paramter
> Alternatively, I would consider getting rid of numerical ids altogether and relying on name alone. Internet Archive does it at e.g. archive.org/details/leroy-lettering-sets, but that has some serious limitations that are not hard to imagine
They don't rely on title alone, it's a separate identifier. You can set it to anything and you can't change it afterwards but you can change the title.
Good post on URL design (GitHub design team): https://warpspire.com/posts/url-design/
Isn’t the /photos kind of necessary since usernames are UGC? What if a username is “about” or “contact”?
I would be very confused if flickr.com/contact went to a user page.
YouTube had the same problem.
The oldest account have a direct URL in the form of youtube.com/username. Newer account had youtube.com/user/username (I think most account have both URLs).
Then YouTube was bought by Google and they introduced "channels", so some channels had youtube.com/channel/username.
Then YouTube wanted to become like TikTok and they forced at-usernames some years ago, so now accounts have a URL like youtube.com/@username
I really like tiktok's solution to this problem. Every user page is prefixed with a @-sign.
tiktok.com/@about
vs
tiktok.com/about
This has been around for way longer than tiktok but with ~ instead of @.
I wonder why we left ~ for @.
Maybe become some foreign keyboard layouts don't have the ~ symbol?
The ~ character for home directories was an old convention that dates from the ADM-3A (1976) terminal used by some early Unix users. The keyboard on that terminal happened to have the cursor control word "Home" on the "~" key. This shorthand was adopted by shells like sh/csh and emerged in HTTP urls as /~user/ being the shorthand for a user's personal web page on a site.
Much later in history Twitter popularized the form "@user" to refer to a personal identity. I'm not sure if they invented the usage or not. This is distinct, but probably somehow cognitively related, to the use of "user@host" for email addresses after bang paths fell out of favor.
For reasons I can't quite put my finger on @user seems a much better sigil than ~user to me, so I'm not bothered that it's become popular.
It makes sense in a chatroom if you direct a message @someone (at someone), or if you direct a tweet @someone. So I guess the natural progression of that is @someone becoming the identifier.
> For reasons I can't quite put my finger on @user seems a much better sigil than ~user to me, so I'm not bothered that it's become popular.
I think this makes sense if you pronounce the action. On Twitter you'd tweet [at] user(s). I think it made even more sense back in the Twitter via SMS where you had to send a message to Twitter's number but direct at a particular user.
Wikipedia claims it was invented as an ad-hoc convention by some Twitter users, and eventually it became so popular Twitter started turning @username into links.
I remember signing up for an AI inference provider, noticing my profile page was "/<username>" and changing my name to logout which worked but caused some amusing errors. Oops!
GitHub manage to do it. Most URLs you'd think of are either redirects to other bits of the site, or accounts owned by GitHub themselves. It just takes a bit of planning.
> It just takes a bit of planning.
Haha, no it just takes forcing user account name changes.
github.com/copilot, github.com/claude, github.com/models, basically everything you can think of for the last few years has been through this approach.
"Hey, remember the username you've had for twenty years? Yeah we want it now"
It's what you get for being a tenant rather than owning your own site.
And yet it's worth it for the network effects, sadly. These companies should be regulated, the moats are just too deep.
I like that the EU recognised this and did start to regulate that these companies need to interoperate.
I wonder how they prompted these users to change their usernames. Was it just a "we need your username for our business, so comply" notice?
https://web.archive.org/web/20210702114132/github.com/copilo... some proof this was an actual user
Looks like he's now https://github.com/ogcopilot
Some more discussion on similar things: https://news.ycombinator.com/item?id=38380344
You can plan every "top level" path you'll ever want on the site from now until forever? Or do you mean planning as in plan to force account name changes on users when someone's username conflicts with a new feature name?
You could probably get away with banning all common english words as usernames if you wanted to.
Even if you don’t ban all words, there are some you should filter:
Or put all user pages under some top level path and then you never need to ban anything as this problem becomes completely moot.
Until you invent a product or concept name that is taken by one of the users ;-)
In that case it's not really different if the usernames are prefixed to be clearly identifiable as usernames - most companies still wouldn't want any user accounts that look like official company accounts.
Strange, because I always remember Flickr having horrible UX. You could never just open an image file directly; if you tried, Flickr would always redirect you to a page which obscured the image behind an invisible layer which obscured pointer events such as right-click.
I learned quickly to avoid Flickr links.
Maybe it was like that for a while? But flickr allowed image downloads, there was a dropdown in the UI with the available sizes for years. And it had an API (+stable URLs) to download images.
It's possible they did not allow the way you tried to access images directly, to enable control of the downloads for the photographer. But I think you misjudged the behaviour back then, they were pretty open.
I believe that is to prevent hotlinking, which isn't the purpose of the service.
Can't say enough good things about flickr. Those people nailed it in 2004 (I've been a paying subscriber since their first year) and everyone else has been making bad copies ever since. Tagging, friends (pretty much inventing social media without any of the diabolical dark patterns), full-resolution archival storage, a solid API, all over two decades ago. I'm frankly embarrassed for things like Instagram, it's like they're not even trying.
oh yeah i remember as a kid into webdev and php how some sites would have these CLEAN urls. seemed like magic to me.
Aww, they finally depreciated the .gne file extension? It was supposed to never end!
from the article : > This might seem silly. The user interface of URLs? Who types in or edits URLs by hand? But keyboards are still the most efficient entry device.
Can someone let Apple know this ? Safari URL bar is a disaster. If I edit a URL say to remove a part and hit enter 9/10 times it searches the internet.
It seems to forget that it's a url it's displaying despite it cutting the front off, even after telling it to use long urls in address bar. It's so annyoing I actually use another browser since I often need to paste in or modify urls for the work I do. Safari sucks hard. Any solutions to disabling 'search' in the address bar ? I want it to be a URL bar only, and anythign typed in it should be resolved and not searched.
You can let Apple know by not buying their products
I've gone back to flickr for my photo sharing, I've had a pro account for, checks account, oh jeez, 20 years. I stopped using facebook and meta, and it's a solid photo sharing service. I can send links and people just get the photos, imagine that!, no ads, clean interface, lovely.
In one of their transitions, they seemed to somehow deleted most of my photos. I don't know which one exactly, maybe it's my fault somehow.
I just know I didn't log in for 10+ years and now when I do, most of my photos are gone. Oh well
They're also holding the photos that are left hostage; unless you pay them, you can't download the photos you have left there.
I thought it's that and they have them behind a paywall but no - they just somehow don't have them
maybe I will check again
edit: yep they have only my latest 50 photos (from 2014).
edit2: ah they did that in 2024 apparently. oh well whatever.
What's so special about "https:"?
i don't understand the heat under ?&
?set=2546&pic=8597 is much easier to decipher than /2546/8597
From my experience, query param is usually optional, also `?foo=1&bar=2` and `?bar=2&foo=1` should open the same page.
If it's a permanent reference to a record wouldn't /sets/2546/pics/8597 make more sense?
> No parameters with their unpleasant ?&= syntax.
I'm sorry what? URL params are just a thing.
TIL Flickr still exists
Stopped caring about them when they cancelled their 1TB free storage after 5 years, company which can't be trusted long term with your data. Plus the UI was horrible anyway.
Free storage expectation registered