2019 in debate is a far cry from what it used to be. Goodbye paper, expandos, and tubs. Goodbye photocopying and printing. Goodbye asking to look at a physical card to write down the citation. We exist digitally in word documents, virtual tubs, caselists with full speech documents, and repositories like the Open Evidence Project. Evidence has, in many respects, gone from a hoarded commodity to being a community resource. The ironic thing is, as we create more evidence solutions, we create new problems in how it works: poring through files to find a single card; spending an inordinate amount of time formatting files, structuring them, perfectly naming and organizing headers. Printing a single debater’s dropbox now would likely collapse a handcart. Evidence has become more readily available to all, but as the amount of accessible evidence has exponentially increased, the management and organization of this evidence has become a gigantic task.
Arvind Balaji, a 16 year old Junior at Round Rock High School, may have the solution. Arvind is the mind behind debate.cards, an open source, flexible, and free to use evidence search engine. I reached out to Arvind over email to talk debate.cards with him to see where it came from, how it works, and where it can go.
[Our emails have been lightly edited for readability in an article format. No content has been censored or changed in meaning.]
Colin: How long have you been involved in debate? How did you get involved?
Arvind: Third year debating. I knew a few friends that debated in high school, figure I would a give it a shot.. and here I am three years later haha
Colin: Where did you get your interest in coding and when did you start?
Arvind: I've had interest in computers since I was young, my parents are both software engineers which is suppose helped me to foster that interest. I kinda just taught my way to code along the way, from starting out with simple website and working up to full fledged projects like debate.cards
Colin: Wow, so you have a pretty deep history with software development. I saw on your github (snooping) that you have developed a debatetimer and an openevidence download tool among other things. Have you done any large projects as sophisticated as debate.cards?
Arvind: I think debate.cards is definitely the most ambitious project I've taken on so far, but the smaller projects along the way helped me learn a lot.
Colin: Is [software development] what you intend to pursue after you're done with school/is this partly an exercise in building your portfolio?
Arvind: I haven't completely decided yet, but I think I'll most likely be pursuing CS. And yeah, I definitely think debate.cards is a good contribution to my personal portfolio.
Colin: What programming languages are you most familiar with?
Colin: A lot of people draw parallels between debate and other activities (usually something they personally find interesting. Do you see a relationship between coding and debate?
Arvind: For me personally, coding and debate are both things that I'm passionate about, so being able to work on a project that combined the best of both worlds is something that incredibly fun and rewarding.
Colin: There are some examples of individuals attempting to monetize software solutions in debate. Why open source and not keep it private?
Arvind: debate.cards was created with intention of keeping it free. The goal of debate.cards has always been to make evidence more accessible, I feel that monetizing it ultimately detracts from that purpose.
Colin: That's an admirable attitude. Developing something that can be used as a public good should have low or no barriers to access it. Debate thrives on individuals willing to do selfless work for the benefit of others.
I have a couple of follow up questions here:
Why do you think that accessible evidence is necessary?
Arvind: I think that more accessible evidence is a good way to help close the resource gap between small and large school debaters. I also think it just increases the overall quality of debate, having access to more specific and well-research evidence allows for debates to be more educational and in-depth.
Colin: There are plenty of people in debate who have disliked some technological changes in how debate works, whether it's the use of laptops or paper, use of internet in debate (more in a regional high school context), using cites or docs, or even using the caselist at all. Most commonly there's a fear that debaters will become lazy and retreat from having to learn independent research skills. An example from [my coaching experience] is that we usually have to go through a period of forcing novice or JV debaters to stop relying on the search feature in their virtual tubs so they actually have to learn their files. You're offering a paradigmatic shift in how people can access evidence in various places online. What effect do you think debate.cards will have in this regard? Do you think this is unfounded or do you just think the benefits outweigh?
Arvind: I think that criticism of tech in debate is inevitable. I'm sure the same was said for paperless, open evidence, and open source. While yes, there is a risk that software like debate.cards will lead to lazy debate practices, I don't think it's as big of deal as it's made out to be. I don't think it's an issue that's intrinsic to debate.cards and yes, the aforementioned benefits are probably more important.
Colin: In simple terms, how does it work (both front and back end)? What makes it easy to use? Where might it have some user interface challenges?
Arvind: At its core debate.cards is pretty simple. First, Word documents are fed into a parser, the parser splits up the individual cards in the document and does a little bit of guess work to get information such as the author and date out of the card.
All of the parsed data is then indexed, the indexer allows for quick and powerful searches of all the stored data. Think google, but for debateevidence.
The then the last part is the actual user interface that allows for searching through the data, and downloading it back into a word doc.
I think the super simple interface it what makes it appealing for users. Search for what you want, then download it. No frills.
Ironically, I think this is the cause of its biggest shortcoming. The site currently lacks features that would allow for more advanced searches and better evidence discovery, features that would make it an even more useful tool for debaters.
Colin: I noticed that the search bar says "Search for a cite or tag..." So the keywords search are for just those 2 parts? When evidence is parsed through, are cites and tags lumped together for keyword searches or separately?
Arvind: I guess that isn't completely true anymore, it searches through the cite and tag along with the heading levels. All three of those things are indexed separately but there isn't currently a way for the user to pick which fields to search in. That functionality will be added in the relatively near future though.
Colin: How extensive is its library? What kinds of places can it search? How easy would it be for people to configure their own to crawl other places?
Arvind: Currently debate.cards pulls data from the past 7 years of the Open Evidence Project and open source documents from this year's High School Policy wiki. Short term plans include adding the LD and College wiki, but my eventual goal is to open it up for users to directly contribute files.
There is currently a section on the GitHub readme file on setting up your own instance of debate.cards which would allow people to add in their own custom data sources, which might be an interesting challenge for those that are tech savvy.
Colin: What could custom data sources be? Locally hosted for an enhanced vtub search tool? Or does it have to be web-based?
Arvind: It could in theory be anything, the application is written in a way that is modular. So people could in theory write their own modules that would add in new data sources. But to be clear, this definitely isn't something I'm recommending to the average debater or team, or even officially supporting it - it's just an idea. Maybe in the future I can look into adding a more user friendly way to do this.
Colin: At the college level there has been some discussion of consent to data collection, even for willingly provided open source documents for disclosure purposes. How feasible is it for you to exclude specific caselist pages from debate.cards? Just as relevant, are you willing to take requests to do so?
Arvind: This is definitely something I've been thinking about recently. When I initially added open source docs to the wiki, I took the liberty to assume that those that are willing to open source their documents would also be willing to let their documents be used for something like debate.cards - both of which share the same goal of making research more accessible to debaters. However some of the backlash over oodebate has made me question that. I still think that for the most part that is a fair assumption, especially since debate.cards is not a commercial product. However, I think it's definitely important to respect the feelings of the debaters that put these files out in the first place. There is some technical work that needs to be done first, but once that's ready I'll definitely add an opt out option.
Colin: Debate.cards has a lot of potential as it is. What kind of development do you see in the future of debate.cards? Is there an end goal vision of what you think it could be?
Arvind: I think the coolest part about debate.cards is the fact that, for the first time, there is a structured way to store and retrieve debate evidence. Instead of being confined to folders full of shoddily formatted word docs, having a more structured and semantic way to store individual debate cardshas enormous potential for the future of technology in debate. I'm frankly not sure what the direction of debate.cards itself will be in the future, but I'm confident that the technology behind it will end up being useful for a lot more than just debate.cards
A public API for debate.cards is in the works which would allow other developers to use the data from debate.cards in their own applications and projects.
Colin: So do you think that the traditional hierarchical structure of debate files should be dropped in favor of individual card searches? Or am I reading too much into that?
Arvind: No not really. I just think that technology in debate is inevitably going to progress, and word docs are just not a very convenient format to work with it. Being able to represent the contents of a speech doc in semantic way just makes developing technology for debate much easier.
Colin: What kind of data could be called through the API?
Arvind: All of the features of debate.cards (searching, retrieving a specific card, downloading as a word doc) will definitely be available. Beyond that, I'm not completely sure yet. I think that once I get the first version of the API running, I can probably expand the feature set based on community demand.
Colin: A lot of nerds in debate at the college level are highly interested in data. Are data analytics and/or visualization somewhere on your checklist?
Arvind: Creating analytics is not something I currently have plans for, but the data for others to do so is something that can probably make available through the API.
Colin: Alright, last question for you. You've done this by yourself so far, do you want people to help out? How can they do so?
Arvind: Yeah help is always appreciated, that's part of the reason why debate.cards is open source. For those that feel like they might have the ability to help out but don't know where to start, just shoot me an email and we can talk.
Even if you don't have technical skills to directly contribute, things as simple as making bug reports when you find and issue or providing feature suggestions help out a lot.
If you are interested in learning more or helping contribute, you can visit debate.cards or find the source code on Arvind’s Github here.