15Dec16 Meeting

talking about different threads of narrative, one in the pre-print article, one in the published article, and one embodied in the code itself.

What we want to do is show that we see parallel fragments, parts, these moments when the code and the narrative "touch" or intersect. Joris observed these points in his notes on the other wiki page. Matt says the charts and graphs are other instances of this thing that we are talking about..when the disparate narratives overlap or touch or imbricate or overlap. Is there a term for when rivers touch/cross. we need a fancy term, with solid latin or greek roots.

Joris noted the lack of methodological details in printed article vs. the pre-print. This would be something interesting to talk to Ted about, why are there so few methodological details in the MLQ article? In the process of authoring the MLQ article, Ted and jordan make a series of decisions about the path of the analysis, but what we find interesting is the vestige of a road not traveled. There are traces of these paths (vs. other paths that Ted and Jordan could have taken that have no trace...using something other than Logistic Regression).

What are the implications for these traces, like the binoral_select() in our methodology? Are these dead-ends or potential joins between the narratives (but unrealized - neither good or bad).

Matt has a question about the impact our article would have on subsequent authorship of code by folks like underwood. Will authors be less open about their code because of the potential for scrutiny. How will defactoring change the way scholars write code?

Increasingly code is part of the process of literary analysis, but the current system of scholarly communication (writ large) totally ignores the code elements of this work. We ask the questions:

• how to review it
• how to understand it
• how to theorize about it
• how to publish and distribute it
• how to preserve it as part of the scholarly record?

We need to point out and motivate these questions. We need to articulate why they are important.

We need to put together an outline of the whole article. Here a rough sketch:

• Introduction
• Literature review (includes STS, plus the stuff we have regarding "review of code in DH")
• Method - introduce defactoring as a method. we describe what we are going to do
• Results - This is the descriptive analysis of Ted's code, an instance of defactoring. focus is on working through each part of Ted's code. Describing what it is doing, but not getting too bogged down by findings or theoretical concepts
• Discussion - This is where we discuss and critically unpack some of the things we found in Ted's code in the previous section. Such as the binormal_select the "eyeroll", the parallel narrative and the touch points between them.
• Conclusion - discussion about defactoring and the motivating questions about code in scholarship.