¶ 1 Leave a comment on paragraph 1 0 Tools and platforms are changing all the time, so there’s an awful lot of work for developers to make sure they’re always working. One tool that we quite like is called Paper Machines (http://papermachines.org), and it holds a lot of promise since it allows the user to data mine collections of materials kept within the Zotero reference manager. It was built by Jo Guldi and Chris Johnson-Roberson. It is a plugin that you install within the Zotero; once enabled, a series of contextual commands, including topic modeling, become available to you by right clicking on a collection within your Zotero library. At the current time of writing it can be quite fiddly to use, and much will depend on how your system is configured. However, it’s only in version 0.4, meaning, it’s really not much more than a prototype. It will be something to keep an eye on. At themacroscope.org/2.0/essays/papermachines.html you can see the kinds of things that Paper Machines aspires to do.
¶ 2 Leave a comment on paragraph 2 0 We mention this here to highlight the speed with which the digital landscape of tools can change. When we initially wrote about Paper Machines, we were able to topic model and visualize John Adams diaries, scraping the page itself using Zotero. When we revisited that workflow a few months later, given changes that we had made to our own machines (updating software, moving folders around and so on, and changes to the underlying html of the John Adams Diaries website), it –our workflow- no longer worked! Working with digital tools can sometimes make it necessary to not update your software! Rather, keep in mind which tools work with what versions of other supporting pieces. Archive a copy of the tool that works for you and your set-up, and keep notes under what conditions the software works. Open source software can be ‘forked’ (copied) on GitHub with a single click. This is a habit we should all get into (not only does it give us the archived software, but it also provides a kind of citation network demonstrating the impact of that software over time). In digital work, careful documentation of what works (and under what conditions) can be crucial to ensuring the reproducibility of your research. Historians are not accustomed to thinking about reproducibility, but it will become an issue.