Flipping Out

Flickr is somewhat unique in that it uses a code repository with no branches; everything is checked into head, and head is pushed to production several times a day. This works well for bug fixes that we want to go out immediately, but presents a problem when we’re working on a new feature that takes several months to complete. How do we solve that problem? With flags and flippers!

Feature Flags

Canadian Flag

Flags allow us to restrict features to certain environments, while still using the same code base on all servers. In our code, we’ll write something like this:

if ($cfg.enable_unicorn_polo) {
    // do something new and amazing here.
else {
    // do the current boring stuff.

We can then set the value of $cfg.enable_unicorn_polo on each environment (false for production, true for dev, etc.). This has the additional benefit of making new feature launches extremely easy: Instead of having to copy a bunch of new code over to the production servers, we simply change a single false to true, which enables the code that has already been on the servers for months.

Feature Flippers


Flags allows us to enable features on a per environment basis, but what if we wanted to be more granular and enable features on a per user basis? For that we use feature flippers. These allow us to turn on features that we are actively developing without being affected by the changes other developers are making. It also lets us turn individual features on and off for testing.

Here is what the Feature Flip page looks like:

Feature Flipper

Continuously Integrating

Feature flags and flippers mean we don’t have to do merges, and that all code (no matter how far it is from being released) is integrated as soon as it is committed. Deploys become smaller and more frequent; this leads to bugs that are easier to fix, since we can catch them earlier and the amount of changed code is minimized.

This style of development isn’t all rainbows and sunshine. We have to restrict it to the development team because occasionally things go horribly wrong; it’s easy to imagine code that’s in development going awry and corrupting all your data. Also, after launching a feature, we have to go back in the code base and remove the old version (maintaining separate versions of all features on Flickr would be a nightmare). But overall, we find it helps us develop new features faster and with fewer bugs.

5 Questions for Gustavo

Following our second interview Kellan snuck in 5 Questions for Paul Mison before I managed to tap Jim’s suggestion of Gustavo.

I think Drift Words sums Gustavo up nicely with “Although he’s far too clever, he makes up for it by using his polymath powers for good.” … and charts and stunning graphs. It’s always a pleasure to see what Gustavo comes up with.

And on that note …

1. What are you currently building that integrates with Flickr, or a past favorite that you think is cool, neat, popular and worth telling folks about? Or both.

Gustavo: First of all I must say: I’m not really a flickr tool developer. The only real tools I created are the quite minimalistic FlickRandom and Contact Crossing (very kindly hosted by Jim Bumgardner). Rather, I use flickr’s API to mine flickr’s database for interesting information, usually leading to some visualizations or at the very least a bunch of graphs.

The FlickrVerse, April 2005 poster: flickr's social network

One such visualization led to a browser for exploring related groups. Not a full-fledged tool since it explores static and outdated data.

In other words, what I do with flickr’s API is not tool development but analyses, trying to figure out the structure and dynamics of flickr’s social network, group content and participation, etc. The most recent analysis I performed attempted to understand how exposure to photo content flows through the contact network. Starting from a basic pattern combining three types of network relations, I quantified the importance of a user’s social network in determining his or her exposure to content of interest.

2. What are the best tricks or tips you’ve learned working with the Flickr API?

Gustavo: Cache, cache, cache. Depending on what you want to do, you might find yourself retrieving certain bits repeatedly, and you’ll definitely want to build a good local cache. Be careful though, as it’s all too easy to underestimate the size and complexity of the data. Your cache might unexpectedly morph from a speed asset to a sluggish monster.

Another potential problem from underestimating the size and complexity of the data is that you might suddenly discover that your code issued many more API calls than you expected. Remember to play nice, and keep an eye on your API key usage graphs.

Expect unexpected responses: It’s always good to make the code smart enough to realize the server’s response is not exactly what was expected, be it in completeness, format, special characters in free text, etc.

Finally, still on the theme of unexpected responses: Grow a thick skin, as there will always be someone who will misinterpret your motives for developing a tool that “uses” their information.

3. As a Flickr developer what would you like to see Flickr do more of and why?

Gustavo: Let me suggest two: one very simple, the other very complex.
The simple one is a “no special perms” authentication level. For some purposes, as a developer I’d like to have the user authenticate for the sole purpose of knowing who they are – I might want to give certain results for their eyes only, or someone else might want to allow certain actions to happen just once per user, etc. At the moment, the minimal level of authentication requires that the user entrust the developer with access to private data. Many users will rightfully decline giving such access, and as a developer I’d rather not have to even request and disclaim, when I don’t need it.

The more complex one requires a bit of background. Almost four years ago, there were various discussion threads on the topic of finding interesting content. Back then I took a stab at using network information to discover interesting photos – first looking for people who post similar stuff, then looking for people who share faved photos (which striatic called neighbors) and finally, actual photo suggestions.


As an aside, the algorithm worked quite well even for people with a hundred or so favorites (I had just 21 faves!); people today have many thousand favorites – a very rich data set to start from. I went on to produce suggestion lists for many people, but that was always me, manually running scripts. Since then, other developers created interesting tools for content-driven content discovery, including Flickr Cross-Recommendations, inSuggest and flexplore.

flexplore 2.1 beta

The reason I never tried to create a stand-alone tool for this is that I quickly realized that to make it work reasonably fast, I essentially needed to replicate flickr’s database. If you take all my favorites, then list all the people that faved those photos, and finally enumerate all their favorites… and you then repeat this exercise for whoever requests it, you either need to use many thousand API calls per visitor (and wait!), or you need to create a huge cache, covering unpredictably disparate segments of flickr’s database. A massive cache that would grow stale very quickly unless people used the tool continuously. If you try to include extra information into the scoring scheme (say tags, color data, group membership…) the network use and storage requirements grow even worse. In other words, this project appears to be completely unsuitable for a developer without direct access to flickr’s database… and flickr’s content suggestion system (Explore) doesn’t provide any personalization tools: we all see the same.

The only effective solution I can suggest (beyond flickr developing a personalized Explore, or flickr sandboxing third party developers) is the creation of high level API calls, embodying some method of complex querying into the database. For example: A higher level API call could accept a list of users as query, and return a list of photos sorted by the number of people (from the query list) who faved them. With the current API, one has to retrieve the full favorites list for each person, collate the results, and discard the vast majority of the information that was transferred. A higher level API method could make such intermediary results invisible and save much bandwidth, latency and data replication.

Hey, you asked. ;)

4. What excites you about Flick and hacking? What do you think you’ll build next or would like someone else to build so you don’t have to?

Gustavo: There’s a number of things that attract me to flickr hacking. There’s of course the vast and ever-growing amounts of data, and the fact that it’s not “more of the same”: The diversity of data types (photos, users, metadata, groups, etc) and relations make this complex system fascinating to study. There is also the human side: People really care about their stuff. People get excited, for example, when they recognize their user icon in a visualization of the social network, and immediately want to explore around. Last but not least, I really appreciate the openness of the flickr API. I’m amazed that such wealth of information is shared so freely!

Free association, first page

As for “what next”, I have a couple of ideas. I’ll only hint that the word association analysis was a fun first step. :)

5. Besides your own, what Flickr projects and hacks do you use on a regular basis? Who should we interview next?

Gustavo: dopiaza as Utata‘s architect.

Dan: Thank you, Gustavo. Next up (unless Kellan gets in first!) dopiaza.

Images from GustavoG,
striatic and malanalars.

5 Questions for Paul Mison

One of the key goals for this blog when we launched it 9 months ago was to be a channel for the voice of the Flickr development community. Most importantly all the amazing developers building on our APIs. Which is by way of introducing our third developer interview here at code.flickr, our first was with Neil Berkman of Photopholow, while our 2nd, and first in the 5 questions format was with Jim Bumgardner a few weeks ago.

We’ll be posting an interview from GustavoG (as tapped by Jim) soon, but in the mean time I want to post an interview from Paul Mison aka blech. Paul is an active participant in the Flickr API group, and I’ve personally been using and loving his new project SnapTrip. And he likes rainbows, which makes him good by us.

A Trip Underground

1. What are you currently building that integrates with Flickr, or a
past favorite that you think is cool, neat, popular and worth telling
folks about? Or both.

Paul: My current main project is snaptrip, which works with Dopplr, a website for sharing personal travel information, as well as with Flickr. It helps you to find photos associated with a trip listed in Dopplr, to label them so that Dopplr can find them easily later on, and also lets you geotag them (if you haven’t already). I’ve also written a Greasemonkey user script so that you can see when other people’s photos were taken during a Dopplr trip.

Flickr photos in snaptrip

In the past I’ve mainly written command line tools; I used a script to migrate images out of my previous home-brewed image hosting to Flickr, and I have another that applies machine tags for EXIF properties (since tags make it easy to find photos; here’s all my photos with a focal length of 50mm). My great lost project was a web application called groupr that would let you see all the photos in groups you’re a member of, but unfortunately the platform I built it on was withdrawn, and I should rebuild it elsewhere (possibly on Google’s App Engine, like snaptrip; possibly not).

2. What are the best tricks or tips you’ve learned working with the
Flickr API?

Paul: The API Explorer is wonderfully useful, and it should be everyone’s first port of call when developing an application. Beyond that, I’d advise picking an API framework that does the tricky things (notably, user authentication – I’ve handcoded it myself, and it can be fiddly), but otherwise gets out of your way.

snaptrip uses Beej’s Python Flickr API, and I’ve previously used flickraw (for Ruby) and Flickr::API (for Perl). All three are nice and minimal, so that when new API methods are added, you don’t have to update your library. flickraw uses JSON internally, too, which is very nice if (like me) you lean towards dynamic, rather than static, languages.

For Greasemonkey scripts, this post from mortimer? about using API calls in GM scripts is great.

3. As a Flickr developer what would you like to see Flickr do more of
and why?

Paul: Well, a relatively minor request would be for JSON in the API Explorer. More seriously, there’s an entire class of methods I wish existed for groups. For example, the only way to track conversations at the moment is the group RSS feed, which isn’t segregated by thread. There’s no way to find out a group’s administrators or moderators. A final example is that the queue of photos awaiting approval isn’t exposed to the API. While I’m not sure I’d have used all of these in groupr, some of them would have been very handy.

Another small request would be for more extras, especially a few method-specific ones. In particular, I’d love to see ‘favedate’ on flickr.favorites.getPublicList.

4. What excites you about Flickr and hacking? What do you think you’ll
build next or would like someone else to build so you don’t have to?

Paul: For all that my answer to the last question was a demand for more methods, Flickr is exciting both because of the wealth of photography there, and the richness of the methods of getting at it. The geographical data, and access to it, that has emerged over the last year is really interesting, and I’d love to do something with it (and intend to, somehow, in snaptrip). However, for a complete new project, I’ve been poking for a year or so at making your Flickr favourites look a bit nicer, and maybe within another year I’ll actually have something out publicly. (I really need a simple job queue; anyone?)

5. Besides your own, what Flickr projects and hacks do you use on a
regular basis? Who should we interview next?

Paul: Well, as a Mac user, I’m a fan of Fraser Speirs Flickr Export for getting my images onto the site in the first place, and if I upgrade to an iPhone I look forward to playing more with Exposure. On the web, Dario Taraborelli’s Group Trackr is very nice, and I use fd’s Flickr Scout every now and again to see if anything’s hit Explore.

However, I think the most impressive recent project I’ve seen is a desktop app written using Adobe’s Air, called Destroy Flickr, by Jonnie Hallman. There were a lot of subtle UI techniques to hide the latency inherent in talking to a network service, so I think he’d be a great choice for an interview.

Kellan: Thanks Paul! And y’all be looking out for an interview with Jonnie and his oddly named project.

5 Questions for Jim Bumgardner

We’ve been keeping a careful eye on our Sister Blog to see what they’re up to. Something that’s particularly caught our eye is "5 Questions ", asking the same 5 questions to the Flickrverse, with the last question being who we should ask next. And so, we hope, it goes on and on.


This is our version, asking questions of those that develop, hack and fiddle with Flickr in new and interesting ways. Of course we couldn’t start with anyone else but KrazyDad (aka Jim Bumgardner).

Jim founded the Flickr Hacks group back in the day, a great place to hang out and ask question if you want to learn how to bend Flickr to your will. In 2006 he also coauthored the Flickr Hacks book for O’Reilly and happily for us he hasn’t stopped tinkering with Flickr yet.

So, without any further ado, 5 Questions for Jim Bumgardner:

1. What are you currently building that integrates with Flickr, or a past favorite that you think is cool, neat, popular and worth telling folks about? Or both.

Jim: It seems like I’m always building something that integrates Flickr. A recent favorite is this interactive mosaic that shows the most interesting photos of the week.

The photos are arranged to form a spiral, a form that appears quite frequently in my work.

Coverpop: Most Interesting Photos of the Week

I have a "cron job" which runs on one of my computers at home, which updates this mosaic every week, so the photos in it are always fresh. Incidentally, I prefer to call this process a "cron joy." Oh, nerd humor…

2. What are the best tricks or tips you’ve learned working with the Flickr API?

Jim: I think every Flickr hackr should have access to a powerful high level
graphics library. My library of choice is ImageMagick combined with the Perl programming language (it also works nicely with Ruby), but the GD library, which works with various languages, and PIL, for Python, are also good.

I not only use ImageMagick for building mosaics and graphs, but also for "under the hood" kinds of things, like measuring the average colors of photos for the Colr Pickr (see below).

3. As a Flickr developer what would you like to see Flickr do more of and why?

Jim: One of the very first Flickr hacks I made was the Colr Pickr


…which allows photos to be selected by color. Since that appeared, I’ve worked on, and seen some fancier variations on the concept, that allow larger quantities of Flickr photos to be selected using multiple colors. But all these systems, require that thousands or even millions of thumbnails be downloaded and analyzed for color. This is because Flickr does not supply "average color" information in its APIs, and cannot provide the color search functionality that this data would enable.

flickr Colr Pickr

I would like to see Flickr provide, via it’s APIs, the three most common colors in each photo (using cluster analysis), and provide a way to search for photos which match one, two, or three colors. These parameters, similar to geocode searches, would need to be combined with some other search parameters, such as tags, to narrow
the field down.

A feature like this would be a godsend to designers. I’ve got sample code for the color analysis, if anyone’s interested… :)

4. What excites you about Flick and hacking? What do you think you’ll build next or would like someone else to build so you don’t have to?

Jim: One thing that excites me is the ability to access large quantities of photos that contain valuable metadata, such as the time the photo was taken, or the geocoded location. I used the ‘date taken’ data to construct this very cool graph of sunsets:

A year of sunsets

While most digital cameras store the time within photos, these days, not enough of them automatically store the location. We have to rely on photographers adding the geocoded information manually, and sadly, not enough of them are geeky enough to do it. I’m looking forward to the day, a few years from now, when most of the new photos on Flickr will also contain geocoded information, as this will enable me to make apps which enable a kind of instant photo-journalism of heavily photographed events, such as rallies and parades. We’re seeing the beginnings of these kind of apps now, but we’re barely scratching the surface.

5. Besides your own, what Flickr projects and hacks do you use on a regular basis? Who should we interview next?

mc-50 map of FlickrLand: flickr's social network

Jim: GustavoG has made some amazing graphs which exploit and illustrate the Flickr social network.

Dan: Thank you, Jim. Next up for our thrilling installment of 5 Questions, GustavoG .

Images from krazydad / jbum , earthhopper and GustavoG.

Twitter API updates, FireEagle and new Flickr API fun

Last night twitter released their next batch of API improvements, of course the one that caught my eye was …

“[NEW] /account/update_location.[xml|json] – sets the location for the
authenticated user to the string passed in a “location” parameter.
Nothing fancy, no geocoding or normalization. Just putting this out
there so developers can start playing with how geolocation might fit
into their Twitter applications.”

saving woeids in the location field

… which is nice as it’s just thrown in there as a ‘what if’ type of thing. There’s no direct reason for twitter to have location stuff, (well no more than Flickr I guess) but everyone knows that everyone wants it.

It’d be great if you didn’t have to update twitter yourself and there was something else out there that could do it for us.

Read the rest of “Twitter API updates, FireEagle and new Flickr API fun” for more on Twitter’s location API, FireEagle, and Flickr’s not-a-geocoder.