search

Andykmcc / project-graph

A project management tool
2 stars 3 forks source link

Big Idea: How should efforts and their relationships be visualized #33

Open mikeswoods opened 7 years ago

mikeswoods commented 7 years ago

I'm no Tufte, so I don't know the answer to this. Based on the competition #32 , almost all existing software falls into (1) a tabular format a la JIRA (2) a card-based format like Trello.

(@Andykmcc Please refine this as needed) There must another way. Efforts are a fundamental unit of organization and they can be nested. Visualizations that lend themselves to this information structure exist (see Circle packing: https://bl.ocks.org/mbostock/4063530). Is this gimmicky? Is it easy to navigate? Is it easy to extract information from (from a human perspective)?

I think something like the above is useful, if coupled with a linear index displayed on the side: The index can be quickly used to locate efforts of interest, while the nesting diagram displays the effort of interest in relation to other relevant efforts. I need to make a graphic mock-up to better explain what I mean.

vincepmartin commented 7 years ago

I thought about this as well. I think that it could be gimmicky in some ways. However, I think that there is some value in having a view like this that is quite dynamic in nature. For some reason I imagine having a visualization linked to some sort of keyboard based method of navigating could work out. Ex something like vim?

Andykmcc commented 7 years ago

i think i want to avoid d3 (or at least the out of the box shit) because i agree, it is all to gimmicky.

graph data can be shown as a tabular data, and i think we totally need to ability. the "kanbam board" view can actually be really helped for individuals working in the weeds of things. so i think a table and/or card view is totally appropriate for day to day management. it is the longer term planning and relationship, dependency management that happens at the graph level.

managing efforts, or at least creating them should be super easy, insanely quick. making it 100% keyboard navigable is something we should def. do. but that is really only useful for like 20% of software engineers. no one else uses keyboard shortcuts. so i think it is important not to rely on them. instead, build them, put a hint about them someplace and then you let the expert users figure them out.

another very technical thing that i think we need from the get go to be able to build a good UX is good search. a lack of good search forces so many companies to build super complex sorting, filter systems that never really deliver anyways. so... who wants to learn Elastisearch? i know it plays very well with Neo4j