Wikidata:Lounge/Growing items

One of the core principles of Wikidata is to identify concepts that are relevant for a large group of people and represent them as "items". However, it has to be noted that items have no clear boundaries, the only boundary is what we agree it to be, or what it has been agreed by some external body. Even statements have no clear limits and are subject to interpretation, that is why we spend effort building community consensus and reaching agreements about what is relevant, how to model data, which properties are needed, and so on.

Also when modelling the sources that provide us data, there is no single way of doing it because, like in a fractal, there are ever-growing levels of granularity without beginning or end. In this essay, I (Micru) present some visual metaphors to understand better what is like to model an item, based on my experience and understanding. This text represents only my view, others might have differing views. As an example I will do so with a possible representation of the Iliad (Q8275).

Story of an item

The item is created...

The Iliad, work item

After finding some agreement of its importance or with the need of supporting other items, an item to represent The Iliad is created. This item represents all the versions of the Iliad, all the different contents, all the representations, etc. We call it a "work (Q386724) item", which is an abstract term to refer to an intellectual creation. It is "information" which has a meaning depending on the observer or agreement with other observers. How to imagine a work item? On the right you can see a possible representation of the work Iliad.

Upstream you can see all the influences, all the existing oral traditions that were there before the Iliad came into existence, all the events that affected its autor and that materialized in the first version of the Iliad. This first version can be imagined as one of the little islands seen inside the flow of the river. With the pass of time, that island has disappeared, but the effects of its existence propagated downstream.

If you look downstream the islands appear bigger because they are in our present, we can know much more about those materializations of the Iliad, we can ask the editor why he translated this verse this way and not another way, but again, there are no clear limits of where one ends and the other begins.

The "work item" means everything and nothing, it all depends on the meaning that the reader or the editor gives to it, and how much data is put into it. You can use any item of the work item tree.

...it starts becoming...

The more attention you put into something, the more aware you are of its details. Same for items. The Iliad is a popular item and as such it gets more attention, and this attention makes it grow with sourced or not sourced statements. The first thing that it is noticed is that the river has many islands, sometimes groups of them. For books we call them "edition items".

Again, they are abstract items with a subjective meaning, but they are useful to cluster data related to a specific translation, or rendering. You can put details about the shore of the islands, or their relation to other islands, and all that is fine.

...gets real...

After a some time adding statements to these items, a new question appears which is, how does an abstract entity like a work gets a physical support? A river island gets sediments or is flooded around with water. Matter is processed by printing presses, or electrons arranged as bits, or sound is generated as words. All those events are "manifestations" (also known as "embodiments" or "materializations"), which represent how matter and information relate to each other for the set of real items that share those characteristics (class).

Or you can pay attention only to the matter-information interactions in just one occurrence, and then you have an "instance" or "exemplar". In this case you can describe it in terms of specific physical properties of matter like weight, size, etc. and how this matter relates to classes (group of elements that share characteristics). The manifestation tree offers some idea about man-made classes that represent a mix of matter and information

...and dissolves back into the flow

Items can grow without end, because there is no end to the data you can gather about them, but it is also possible that some parts become no longer relevant, and then they are deleted. Wikidata will eventually disappear too, hopefully not too soon. So do not get obsessed with it, and enjoy contributing :)

What we really do on Wikidata

• 
  Building consensus
• 
  Creating items
• 
  Growing models of reality
• 
  Analyzing and importing sources
• 
  Maintaining conceptual trees
• 
  Wondering about reality

Some guiding principles

• gather data that can be used now in any Wikimedia project and use it
• think about the needs of other people using the data
• establish limits about how much detail you want to add
• grow first the items on the base before growing the items on the leaves
• use fluid and wise criteria, not clear-cut rules
• be aware that natural language is polysemous and without clear boundaries, split concepts whenever needed
• predict growth and structure accordingly, for instance if you are entering details about one of many editions, create two items, one general for the "work" and the other for the "edition"
• "good enough to be used now" is "great"
• "good enough to be used now, and sourced" is "wonderful"
• "good enough to be used now, sourced, and used in a Wikimedia project" is "perfect"