An interview with Darius Kazemi, internet artist

One of the most common feature requests we receive on Android and iOS Wikipedia app teams is to sync the reading list between devices, and so we’re happy to announce that the synced reading list is now available on both mobile apps .


You can also sort, search and reorganize your lists offline. To activate syncing, you just need to log in and enable the sync feature, which will be the first time you log in and pop-up under ‘Settings’. 

You’ll need an internet connection to sync, but all other features work offline, and allow you to download and organize a collection of articles for travel, commute or whenever you need it without additional data costs. If required, it is required to do your research work.


Keeping our lists in sync means storing those lists on our servers. We approached this by looking for the most privacy-friendly way to achieve this. We do not build profiles on users, or cater to your reading interests. When you stop synchronizing, we lock it everywhere in your account and delete your list. Although we store them for you, we treat them as your lists, not ours.

And on the web 

we also know that readers want to save things on their laptops and browsers so that they can read them later or offline. A plugin for desktop Chrome and Firefox browsers is still available, with Safari coming soon, which will allow you to add articles from these browsers to your reading list, and next time you’re connected to the browser, put them in the application Magically display the Internet on your phone. Learn more here

What was your first project that uses Wikipedia?

Oh God. Good question I think it was probably content, forever – it was an extension of a failed national novel creation month prototype. The idea is that it produces a semi-coherent “thinkpiece” type article only by falling down a Wikipedia hole! I wrote a little post about how it works which may be of interest to your readers. It uses a parse action to get JSON formatted content, then parses the HTML clientside using JavaScript and jQuery.

What is your normal process when building a bot after initial thought?

I try to start making bots as soon as I have ideas, and the first thing I do is a manual, non-computational prototype. So if I have an idea for an algorithm, I will write the algorithm and then try to follow it manually a few times to see the result manually. So for @BracketMemeBot, I manually created some brackets to see the results around Wikipedia’s categories to find out what my algorithm might look like. 

This is the research part of the process, and it is helpful because I can determine very quickly whether a bot will be good without writing any code. I leave a lot of projects in the process because the output doesn’t have what I want, but no code is written either so it’s not a big deal.

It is hard to say because I am almost always ready to do a certain thing,

immediately discover that is not possible, and then redesign the bot within the parameters of what is possible. Like my Empire Plots bot, I had to find a way out about the fact that there is no “dead African American actor” category on Wikipedia. But I’ve always been able to find a workaround on MediaWiki or DBpedia – although often, as with content, forever, I need to add HTML parsing / scraping to the mix.

Probably the most specific road block I was running was with the bracket meme bot. It would certainly be nice to be able to get a random range, or at least to be able to get a number of the total number of available ranges and then be able to sequence an arbitrary number in that range. 


Please enter your comment!
Please enter your name here