Zar2gar1

Howdy!

Oh, There are User Templates

This user is a WikiHobbit.

e^{i\pi \ }

This user is a mathematician.

Big list o' reorg ideas

Merger suggestions

Bureaucracy Civil service (Society)

Trust (business) Trust company Corporate group Holding company (Society)

Academic study of Western esotericism into Western esotericism (Society)

Already discussed and seconded on the article talk-page
The content can probably be pruned significantly before / after moving into the primary page

Classification Classification (general theory) Categorization (Science Basics)

May be a tricky one, but all 3 have significant overlap, even if technically different
Also see Classification of the sciences (Peirce) and the general Typology article (currently a disambig)

Primary alcohol into Alcohol (chemistry) (Chemistry)

This is an easy one :-)

Primary carbon, Secondary carbon, Tertiary carbon, Quaternary carbon all into Carbon-carbon bond (Chemistry)

Another easy one

Ice II Ice III etc. (Chemistry)

Essentially all phases of ice except maybe Ice Ih
Two possible approaches:
- Squeeze them into the current Phases section on the Ice article
- Split out the Phases section into its own article, then consolidate there

Aqueduct (water supply) Navigable aqueduct Aqueduct (bridge) Canal (Tech)

This will be a challenging one that takes some thinking through
Some of these articles should definitely remain separate, but there's a lot of overlap across them

Assisted-opening knife into Switchblade (Tech)

Technically slightly different, but the same idea, with a lot of redundant content

Manchester Baby and Manchester Mark I into Manchester computers (Tech)

A borderline case so ask for consensus first at the article
Definitely a lot of redundancy though, especially in the Background sections

MEMS Micromachinery Nanoelectromechanical systems (Tech)

Even if keeping the micro- and nano-scale articles separate, there's a lot of redundancy
Also check links on the MEMS article for others that could potentially be absorbed

Solid geometry into Three-dimensional space (Math)

Symmetry Symmetry (geometry) (Math)

At first glance, makes sense they're separate
A close reading shows the Symmetry article is still overwhelmingly mathematical though

Engineering optimization into Design optimization (Applied science)

Engineering studies into Science and technology studies (Applied science)

Solar fuel into power-to-X (Tech)

Engineering research into Applied science#Applied research (Applied science)

New article ideas (or expand in existing articles)

Benediction sign (Religion)

Noticed no direct explanation of benedictio latina & benedictio graeca on English wiki
- Especially relevant to art history
Can migrate over article from French wikipedia: fr:Signe de bénédiction
Will need to disambiguate with:
- Hand of benediction (medical condition with related etymology)
- Benediction (a full ritual, not just the gesture)
Related concepts include Mudra and (parts of) Priestly Blessing
- Include intro sections and out-link to main articles?

Event notification (Tech)

Currently just a redirect to a (pretty generic) subsection on Event (computing)
Noticed while working on Reactor pattern that there's no detailed discussion of the mechanism
Many specific instances have their own articles:
- Select (Unix), kqueue, epoll, IOCP
- libevent is also related, though it's more of an abstraction layer

Improve separation of related articles

Gossypium vs. Cotton (product)

Already split, but could improve cross-links and hatnotes
Also check for redundancy between articles

Sugar vs. Sugarcane vs. Sugar (chemistry) vs. chemical types like Sucrose vs. product classes like Brown sugar

Another product vs. source one, but this one gets messy really quick
Current suggested course of action
- Add Sugarcane to hatnote on main article and disambiguation page
- Move Sugar (chemistry from current redirect (Carbohydrate) to its own page
- Migrate very specific details from Carbohydrate to Sugar (chemistry)
- Migrate chemical details from main article to Sugar (chemistry)
- Consolidate / re-orient specific types like Sucrose to Sugar (chemistry)
- Consolidate / re-orient specific product classes towards the main article
- Consolidate cultivation parts of production from main article onto specific source crops

Palaquium gutta vs. Gutta-percha

One more product vs. source one
Improve hatnotes and cross-links, then consolidate redundancies
Possibly add disambiguation page?

Western esotericism vs. Exoteric

Already discussed this some on talk for Western esotericism
Essentially, eso- and exo-teric have two historically related but distinct contexts:
- In the loose sense, esoteric doctrines vs. more mainstream ideas
- More technical in philosophical scholarship, when a philosopher's works are believed to be written for select students vs. a general audience
Consensus seems to be for the following course of action:
- Rename Exoteric to Esoteric and exoteric
- Move content on the scholarly context from Western esotericism to the new page

Surface vs. Surface (mathematics) vs. Surface (topology) et al.

Need to discuss and get consensus; no clear course of action yet but consider the following
Migrate out details from Surface (mathematics) to more specific articles, such as:
- Algebraic surface, Coordinate surfaces, and Solid geometry
- This has already been done to an extent for Surface (topology)
Migrate out generalities from Surface (mathematics) to the main page
Re-evaluate Surface (mathematics) page
- Make a redirect to the main article section if minor enough at that point
Re-evaluate specific articles for further consolidation with each other
- E.g. Coordinate surfaces vs. Solid geometry

Simple template & module ideas

Here are a few ideas I've had that maybe I'll get around to someday. Unless someone else wants to beat me to the punch:

Improve VA link template

Template:VA link is used a lot on the VA discussion pages, and people seem pretty fond of putting it in the header. However, this results in unstable section anchors. How about...

Update the underlying module at Module:Vital article to accept a dummy control flag in the VA link function, but still default to false

Make the dummy flag functional to inject a plaintext marker ("VA §") instead of the VA bullseye icon

Create a second VA link template, with safesubst, to invoke the module with the plaintext option

This should minimize any disruption to using the current template while the new one gains traction

Create a custom user.js widget to replace the plaintext marker with the icon in browser

Have it filter on namespace too, especially in the off-chance of collisions in articles

Update the template & module docs to indicate usage

Report the new template on the VA talk page and update VA instructions to indicate usage

Systemize industrial infoboxes

We have infoboxes for |products, companies, and even industrial processes.

However, there doesn't seem to be a clean schema connecting them together, and there actually aren't more general infoboxes for industries and technologies (the link Template:Infobox technology is actually just a redirect for industrial processes)...

Create a general industry infobox

Create a general technology infobox

Refactor the existing infoboxes a bit

Seed 10 articles with the industry infobox

Seed 10 articles with the technology infobox

Update 10 articles each with the other refactored infoboxes

Preliminary research: VA code and data

The VA project is especially starting to pick up at level 5, which is at a whole different scale. Cewbot already does a lot, but I'm interested in trying something new and maybe taking up a bit of the load:

Preliminary research and planning

Want to try doing an initial version in Lua even though it doesn't have a bot framework
Shouldn't be too bad though if I keep the logic clean and Mediawiki API calls simple
Can always fall back onto Python / PyWikibot if necessary

Check I won't be stepping on Cewbot's toes

Spoke with Kanashimi who said a 2nd bot would be good
Cewbot's code is available if I decide to fall back onto JS and reuse it

Settle on vitality metrics & figure out sources

Quarry is good for basic queries & testing
However, the DB replicas impose a lot of limits
- Particularly in regards to views & indices (and therefore potential joins)
- By creating a user DB on ToolDB, one can have much fuller control
Many items will need to be pulled from content though
- Probably via the Wikimedia API

Vitality metrics

After playing with Quarry some, I've determined I probably will need to create a user DB on ToolDB. However, the table-based metrics should still be easier to gather than the Mediawiki API ones to start:

Task #1: Compile DB vitality metrics

Get account setup on ToolDB

Get enwiki_p as a user clone

Configure all tables, views, & queries

Collate the following result set for VA articles only:

DB-sourced Metrics
Metric	Frame	Expected dynamics	Breakout?	Other comments
Creation date	Historical	Stable		See Lindy effect
Last revision date	Current	Unstable		Primarily to filter out stale articles
Edit density	Moving average (MA)	Cyclical and fluid	3, 12, & 36 month MAs
Languages	Current	Sticky
Interwikis	Current	Sticky
Wikilinks	Current	Sticky	In-, out-, total, and ratio	Article namespace only

Task #2: Create Mark I model

It may not be pretty, but I'll probably just download the results and load them into a spreadsheet to start.

Then I'll try building up a few models. The key points to keep in mind:

Try each factor twice, one raw and another logarithmic (may follow a power law)
- Set the objective to the VA level, viewed as a log (VA5 is 1 point, VA4 is log_10(5), VA3 is log_10(50) ...)
Don't forget to randomly assign VA datapoints to training & validation sets
Get effect size estimates too (use ANOVA if the lin-reg solver doesn't return)

Thoroughly discuss results and share with WP:VA

After discussion and comments, save model as 1st baseline

Task #3: Generate Mark I recommendation

Implement model in code (using my VA bot?)

Gather metrics for all articles

Generate & publish list of likely vital articles

Task #4: Integrate pageview data

It's often cited (along with interwikis) in proposals so it will be really interesting to see how strong a correlation it is:

Gather page-view data for all VA articles only

Use the Wikimedia Analytics API

Retrain and re-validate model; discuss results

Take baseline as model Mark II; generate & publish new recommendations

Task #5: Integrate page data from XTools

Gather other metrics from pages or XTools (will likely require a bot):

XTools-sourced Metrics
Metric	Frame	Expected dynamics	Breakout?	Other comments
Wikiproject priorities	Current	Stable	Tally by rank
Prose size	Current	Sticky		May be symmetric, follow a normal distribution?
Assessment	Current	Stable		Be careful, could be particularly circular
Watcher count	Current	Sticky		Redacted < 30, adjust down

Retrain and re-validate model; discuss results

Take baseline as model Mark III; generate & publish new recommendations

Task #6: Integrate page data from Wikimedia REST API

Gather other metrics from the REST API and scanning content (will definitely require a bot):

REST API and Text-based Metrics
Metric	Frame	Expected dynamics	Breakout?	Other comments
Citation density	Current	Stable		Seems promising, but details need some thought
Infobox presence	Current	Stable	Tally several with cap?
Media file density	Current	Stable	By file type?

Retrain and re-validate model; discuss results

Take baseline as model Mark IV; generate & publish new recommendations

Task #7: Automate recommendation sets

Should actually be pretty straight-forward, especially if the model is already coded.

Task #8: Collate historical list size data

This was a request on the VA talk pages, may be more insightful for Lv 4 and 5 subpages. This should probably get its own bot too. Obviously a pretty heavy lift so won't be implemented anytime soon

Grab more recent counts from edit-descriptions

Probably the simplest strategy going back as far as Cewbot documents the section count
- Obviously, won't be 100% accurate for all times (e.g. if Cewbot was down for maintenance)

Export data dump somewhere

This may make more sense as a table or page under WP:VA
The data should mostly (barring corrections) be append-only

Include moving-average calculations in data dump

(Wishlist) Data-mine actual page-versions prior to Cewbot

This could get tedious so probably won't implement anytime soon

VA bot plans

While it will probably intertwine with my work on the vitality estimator, I'd also like to whip up a more vanilla bot to further automate things at the VA lists.

To start, I think I'm just going to consume the json files gathered by Cewbot at Wikipedia:Vital articles/data. Eventually though, I'd like to help Kanashimi out some, and maybe my bot can handle some overlapping functions with Cewbot as a fallback. It could just audit by default, then actively edit only after it notices Cewbot has gone MIA for a few days.

Task #1: Create skeletal bot

Start proposal process for new bot

Create a bot account

Create skeletal bot (in Lua for kicks?) to perform actions

Can always fall back to Python if it's too much work

Perform some allowed test runs on sandbox to ensure I can read & edit

Task #2: Automate updates to VA5 table

Create a new quota subpage (as a single source of truth)

Add wikitable formatting to the bot (if needed)

Write up actual collating logic and test in sandbox

Quick improvement pass on wikitable layout

For example, supercategories should be genuine roll-up lines, not detachable (e.g. when sorting)

Start running on VA5 page

Update VA5 instructions to note table is automated

Rollout to VA4 page too

Task #3: Audit and sub-in for all counters

Gather list of all counters in VA project

Implement counting logic

Provide audit report (see database reports like Cewbot)

Check with Kanashimi and allow editing for miscounts older than 72 hrs

Task #4: Add supplemental list quality checks

Flag duplicates within a single level

Cewbot already does this too

Detect category crossovers between levels

For example, if Petroleum is in Chemistry at one level and Tech at another

Auto-resolve redirects

Cewbot may already do this

Flag other non-article types (lists, disambig, etc.)