User:Digitect

=Contributions= Editor at Wikipedia since 2004 and author or major contributor for the following articles, among others:
 * Richard Schultz
 * Eigil Nielsen (footballer, born 1918)
 * United Premier Soccer League
 * Shotcut
 * Little Diversified Architectural Consulting
 * Warren and Mahoney
 * David Chapple
 * Chilled beam
 * Global TransPark
 * Original artist of several US State flags contributed to the Public Domain, developed by Wikipedia, and now referenced elsewhere around the web:

=Off-site Interests=
 * Steve Hall Architecture I'm an architect inspired by the craft of architecture: beautiful design and the details, building science, and craftsmanship to create it.
 * The Cream for Vim text editor
 * Genealogy: Hall (Immigrants from Germany)
 * DodoCAD A customization overlay for AutoCAD

=Exploration into Automated Sourcing=

Background
It is common on Wikipedia for tables of data be derived from surrounding information. It should be immediately obvious that tables are summaries of narrative.

For larger bodies of information, these tables may become quite large. On Wikipedia, perhaps too large to fit within the narrative from which it is sourced, and so, broken out onto a separate page.

The Problem
(1) Isn't it immediately obvious that tables derived from linked pages are sourced by those constituent pages?

Apparently not. I've been bumping in to more and more pages deleted for only this reason. There's actually a little self-appointed deletion gang running around removing pages they don't understand. It's hard to believe that Wikipedia has devolved into such a simplistic mechanism, but I guess it's inevitable that an increasing number of users won't remember back to the approach when it was founded. Back then, we were all simply trying to make it better. You researched and referenced where missing, and edited and improved where lacking. In that context, deletion is just lazy. Nobody did that back in the day except to address graffiti and vandalism. But we didn't keep track of our editing statistics back then, either. The whole place would improve if editing stats were anonymous and nobody could brag about how much they've done. I seem to recall that's why wikis started in the first place, Wikipedia, too. But I digress.

So if cross-linked pages are not enough to imply derived sourcing, what else could we do?

Options
(2) We could cut/paste sources between the narrative and summary pages. Again, this seems obvious to say, but that would be disastrous:


 * Manual repetition of sourcing is risky. It is a likely way to introduce topographic errors.
 * Maintaining two different locations to a single reference is virtually impossible without an automated cross-reference system.
 * Sourcing in narrative connects with a different mental model than for raw data, despite being the same. Over-eager editors reading too quickly could mistakenly disconnect one but not the other.

So...

(3) Could WP develop a more sophisticated method for connecting the dots? An interconnected cross-referencing system for sources would solve everything. It needs to be sophisticated enough that a new editor would understand the many-to-many nature of the process.

Example
This is only one particular example of a large table derived from many well-sourced pages. (The original page was Comparison of Canon EOS digital cameras and was deleted 2020-05-02T03:00:49, citing lack of sources, despite having 1,394,082 total views, 24,458 monthly average lifetime, and 789 daily average the past year. It had 1,349 data points (19 columns x 71 rows) and was developed by users with 162 different logins over 11 years. It's hard to believe a few youngsters thought that kind of contribution didn't weigh enough in its favor, but as I said previously, WP has been under new management for a while and has been off the trail in the woods for a quite few years.)

For a page like this, what would an automated cross-referencing system look like?

(NOTE: This is historical and is being used as test data, do not reference.)