humanitiesplusdesign/athanasius-ui-ux

athanasius-ui-ux's Issues

Network Layout

Have you seen the last document in the wiki? What do you think?
I'm almost finished the redesign of the main structure of the interface but I would like to understand what do you think about the network layouts first.

Info about a node

Which information you want to see when I select one node?
All the data avaible about him or a predefined list of information that work for all the nodes.
If we go for the predefined list which are the info? Only name, date of birth and death, place of birth and death?
Another option could be that the "basic" list is there by default and in the settings of Knot he can add other info and costumize the "info panel".

Implementation

I've seen that old prototype of Knot is not improving (I'm talking about the network features especially).
Is someone still working on it? Who's gonna develop the "redesigned" KNOT?

Data for the mockups

While I was waiting for your reply i started working on the network with the real data from the D'Alembert dataset to work on the filters and layouts.
This network data have two main problems that makes the mockup design pretty hard:
1 - Inconsistency in the "occupation" column. Now there are 172 values for 592 nodes (213 are unknown) and most of them are really inconsisten (e.g. "litterateur" VS "littérateur published moral and critical essays". Is it the same occupation or two different ones?). Is there anybody working on that?
2- The network is basically a egonetwork. If I remove D'Alembert from the network the other nodes are not connected to each other. Is it ok for you to work on this kind of network for the mockups?

Graph portal to the datastore

Check out Moritz Stefaner's elastic lists. A version of this could be a very helpful way to quickly explore the data set.

Diagram

This is beautiful, Giorgio. Dan and I will review it carefully and get back to you soon.

Node list in Knot

In the current iteration of Knot, clicking on a name in the list of nodes hides the node. This is not very useful. A much more useful function would be: 1. MOuse-over to "find" the node on the canvas; 2. click the name to "select" the node on the canvas

History of Choices

The original diagram includes an exportable history of actions. I don't know what this would look like, but to turn a number of queries and sub-selects into human readable steps would seem very difficult.

What would be helpful in the UI is a visual cue for selections while the user is using the tool. This could take the form of percentages, shifts, and selections. Examples will be posted to HDlab site.

As a future enhancement, tracing a history of choices, with the option to "roll back" would be very helpful. The original mockups for Knot showing the linear path of selections, would be an ideal model for this.

Data Sources and "schemas"

What do you think if the user can select in the settings which data sources and schema to be used?

Unknown Value VS No Value

If i have nodes on the canvas from different "collections", and one of these collections have the attribute "occupation" while the other hasn't, the nodes that have no value for "occupation" (cause they don't have a column "occupation") are displayed like the nodes that have an "Unknown" value in the column "occupation"?

Add attribute

What is this function about? The added attibutes, once applied, are on the same level of attributes that are in Athanasius or we consider them as temporary tags?

baseball card

The baseball card view as it stands now is too literally a baseball card. In baseball, the stats are clearly bounded and defined: You have an active profession period and agreed upon measurements.

Instead, we need to represent these cards not as people but as real entities about which we have limited information. We need to work on showing what we don't know about an individual as much as what we do know.