The x that’s not the meta-x

I’ve been looking for a prefix to mean ‘not meta-‘ (and I’m not the only one). Of course there’s a lot of confusion about what ‘meta’ means – literally it means ‘after’ (‘metaphysics’ was the section of the learned curriculum that happened to follow ‘physics’). By reanalysis ‘meta-x‘ has come to mean ‘x about x‘, as in ‘metadata’ (data about some data) or ‘metatheory’ (theory of a theory).

(“Metadata” has itself picked up a vile alternative meaning, also by reanalysis, of data about anything, not just data about data.  This is silly since all data is about something, thus “data” and “metadata” are synonyms in this reading. One loses a very useful distinction for no good reason.)

I often find myself wanting, for local rhetorical purposes, such a category, for various x. For example, in discussing data and metadata, one wants ways to refer to the data and its metadata as separate entities, and one can get confused easily because both of them are data.

There are two possible ideas here. One is to have a category (blank-)x that is the complement in x of meta-x. That is, something can qualify as a (blank-)x by not being about an x – perhaps not about anything at all. But I think this is not very useful, perhaps because non-meta-x is not a natural category (sort of non-black-dog). I think my application is more about roles than categories: When you have an x1 that is a meta-x and an x2 that some meta-x (such as x1) is about, x1 is playing the meta-x role and x2 is playing the non-meta-x role in the discussion — even if x2 is about some third x.

E. O. Wilson needed a term to distinguish merely social animals such as birds and people from really social animals such as termites, and coined ‘eusocial‘ where the prefix ‘eu-‘ means ‘good’ or ‘real’. So I’m trying out ‘eu-‘ as the opposite of ‘meta-‘. If I say that a thing is the euthing then it is the thing that is not playing the role of metathing in the current discussion – the euthing is the thing that the metathing is about. (As a category, maybe a euthing is a thing for which at least one other thing is its metathing.)

Example: I search a data repository catalog and retrieve some metadata about a data set of interest. I might say the data I see is the metadata, and the data it describes is the eudata. If the eudata is itself metadata (for some third data), that’s fine, it’s eudata in the first discussion but when we start to look at its relation to the third data, it becomes the metadata and the third data becomes the eudata. Or, the first (meta)data might play the eudata role if I were talking about the HTTP response message that carried it to me – then the HTTP headers are metadata. Etc.

Just a thought.

Protecting resources

This post is for Mark Sheldon: Last summer I wrote this memo on “resource protection” (access control, sort of) on the web. It resembles a brief summary of Tyler Close’s paper ACLs don’t and was input to the W3C TAG (technical architecture group)’s consideration of a draft from the W3C web apps working group on cross-origin requests (basically, the problem of preventing attacks on web page B when web content from A accesses content from B). Later in 2009, Tyler and Mark Miller published “Uniform Messaging Policy“. The conversation continues in the web apps WG.

New opportunities for linked data nose-following

Here’s something I just posted to the TAG blog on how to find metadata and stuff like that:

New opportunities for linked data nose-following