The notion of <https://www.theboohers.org/2015/03/...
# thinking-together
o
The notion of provenance gets used a lot where I work, it’s a great term which (roughly speaking) refers to the lineage of data. Provenance documents the inputs, entities, systems, and processes that influence data of interest, in effect providing a historical record of the data and its origins. The term is probably worth a discussion on its own, but my question to y’all is this: Where provenance charts the history/past of data, is there a term that refers to its future, such as possible transformations, movement, etc, etc. My mental model here is something like a light cone where provenance defines a specific path within the possible histories of the data, its counterpart would define the possible, or probable future path. I’m mostly just looking for a good term because I think it would be useful, I’ve been trying to mush different etymologies to find something that sounds okay but nothing great has come out of that yet. I’m not expecting there to be much work on really developing the notion concretely, but if there is I’d certainly love to know about it!
❤️ 2
t
Service designers have a process called "futures thinking" which contains a diagram very similar to what you just deacribed
o
^ Cool! Different kind of field but I can definitely see some useful stuff there.
t
I heard of this through "Annika Hamann. Lead UX & Service Designer, and Futures Thinking at Futurice." I am sure she would be happy to share some references
Oh sorry. A more concrete thing is reachability analysis in (e.g. Markov decision) processes which deals with the state distribution becoming fuzzier over time. I don't know anything so fitted to data though
1
o
^ this is a great example, thanks 🙂
n
fate? Or, you could use the terms source and sink, or dependencies and dependents.
t
Now I thought of Markov chains describing the path of passwords... it made me laugh. An important property is " The stationary distribution of a Markov chain describes the distribution of Xt after a sufficiently long time that the distribution of Xt does not change any longer. " It's when the chain forgets its initial conditions. For a password this is when it has reached the darknet and haveibeenpwned! It's doesn't matter about your security, all passwords will eventually reach this state regardless of which database you originally store it in!
😄 2
k
Sounds like a good topic for a scientific paper: "The fate of passwords and its relation to the heat death of the universe" 😉
To be a bit more constructive: various domains of science dealing with dynamical systems and/or causality have discovered this concept, but use different names for it. In statistical mechanics, it would be the accessible phase space. Reaching even farther for analogies, the null space of a matrix in linear algebra is similar as well, when you see the matrix as describing constraints. Bayesian inference is all about this, in a way, the overarching question being: given what I know, which constraints does it impose on what I don't know?
d
state
o
@Daniel Krasner I feel there’s a lot of truth to this actually. It’s kinda what I’m trying to get at. And it is a key distinction between provenance and its inverse. Provenance is about ‘immutable’ values, reasoning about possible futures/process/transformation requires something like a notion of state. At the very least, reasoning about these things requires us to include the environment and context, where provenance can largely ignore any context that isn’t deemed important for provenance. The opposite is true in this case, as the entire environment must be considered because we can’t know ahead of time what may influence it or what may be important.
j
It has an existing meaning in finance that you might need to overload, but I would go with prospectus.
k
@Orion Reed
provenance can largely ignore any context that isn’t deemed important for provenance
If you want complete provenance tracking, you can only throw away what you can prove to be unimportant. Which in today's common software environments isn't very much.
💯 2