Language in the tools and resources we develop


#1

First, I just wanna say thanks for the folks who have Tinfoil.Press together. It has had some very impressive conversations so far.

Second. Building Umbrella we have tried to focus a lot on the language that we used. Doing our best to make it's language use and tone more accessible to people. We still have a long, long, long, long way to go. In our future development plan we have a big push to really do more in the area. Drawing as much as possible on Flesch–Kincaid and now "Controlled Natural Language," we are trying to experiment with ways to see can we get close to Randall Munroe's "Thing Explainer" while still equipping the average user in the field with security knowledge that they need. Clearly that is a tough task.

I wanted to create a thread to start to share some of the resources that we are using and also to see has anyone else tried to take this approach in their digital/physical security work? Has anyone found technical resources to help with this? (Other than purely for scoring)

"Readability Score"
https://readability-score.com/

"Test your document's readability"
https://support.office.com/en-gb/article/Test-your-document-s-readability-0adc0e9a-b3fb-4bde-85f4-c9e88926c6aa

"Controlled Natural Languages for Knowledge Representation"
http://www.aclweb.org/anthology/C10-2128

"The Controlled Natural Language of Randall Munroe's Thing Explainer"
https://arxiv.org/abs/1605.02457


#2

Agree language is massively important.

But while it's important to use words that people already know, I don't think controlling the vocabulary is a complete solution.

I'd be interested in discussing approaches to user testing of training materials. So far my experience with this has been very informal. But one could imagine an online survey, "did you understand this article?" Or even better, a little button you press next to any paragraph that seems confusing to you.


#3

Thanks for mentioning this, it's something that's been on my mind as well. It's very easy to go down technical rabbit holes, but I hope we'll avoid that on Tinfoil. There's an opportunity here to make it a useful resource for new voices, and for surfacing new questions, but only if we make it an inviting place for new people to talk. We probably need to start with simply being conscious of jargon, and using plain language. But more fundamentally, we really ought to be thinking about how to put ourselves in the shoes of someone who is new to this space.


#4

What are the questions beginners ask most about journalism security? I've been a trainer and I still don't have a very good sense of this. Has anyone ever kept a list?


#5

Using technical language in order to signal status, or using an obscure term when a less obscure term would be equally specific, is a waste of time for everyone.

That said, words actually mean things.

Technical language is not an obscurantist lexicon that exists solely to exclude outsiders (that's continental philosophy). The technical lexicon is vast and complicated because technology incorporates a vast and complicated body of knowledge. In such a vast forest of information, specific coordinates are not just helpful, they are indispensable.

Since the question of "good faith" versus "bad faith/status signaling" uses of technical language come down to a question of intent, it is difficult or impossible to police "good/bad" uses of this language. In order to do so consistently and fairly, one would have to know the state of mind of another person. Clearly, this is not possible.

I would advise:

1.) If you don't understand what someone is saying, assume good faith and ask for clarifications and/or links to further resources

2.) If someone asks you to clarify a term or concept, politely offer them an explanation or a link to further resources. At the very least, useful search terms that would lead them to further reading would be helpful.

3.) Read. Learn. Google is your friend. Academic journals are available to all now:

scihub22266oqcxt.onion

(Tor Browser is required to follow the above link)


#6

Thanks for sharing this approach, Rory! I look forward to exploring these resources further.

In addition to paying attention to accessibility in the language we use, I also wanted to suggest that we strive to make sure the language we use promotes inclusivity, social change and justice. Here's a great example/resource recently commissioned by SumofUs: https://s3.amazonaws.com/s3.sumofus.org/images/SUMOFUS_PROGRESSIVE-STYLEGUIDE.pdf.

Kristin


#7

This is good (I am a big fan of Flesch-Kincaid and other metrics). I would also say that we can treat non-linguistic things as being 'language-like' and gain some understandability benefit from doing so. For example, think of each icon, or color, or shape, et c. in your information design as being 'like' a word. Now you can see that, if using the ten hundred :slight_smile: most common words is a good idea, then using the 10 most common icons (rather than obscure ones) is a good idea.

And make sure that the icon matches the N high-order bits of the related text. It's counterproductive to have an icon that seems to mean X when the next 500 words of text say "Well, X, but not really. It's actually Y..."