What pages need work tool - design ideas

Hey tropers, new backend dev here looking for some guidance from the veterans in the community.

I want to create a tool that allows users to search through pages based on whats missing, what needs more work or pages we don't have but should. I have seen users asking "where do i even start?" when it comes to editing or adding new pages so a tool like this would solve those problems and allow current users to more easily decide what to work on.

My questions are, how should the tool function? If you could magically create such a tool, what would it sort by and why? (ex: maybe there's a checkbox on all pages to flag it for work w/ a reason and then it shows up in this search tool)

I realize there's a lot to "clean up" for current pages, but a big part of my focus here is also looking for ideas on ways to encourage more content creation, whether thats adding more content to the pages that exist or ways to streamline launching new ones.

Edited by kory on Jul 30th 2023 at 1:05:16 AM

Now monitoring Wishlist and Bugs

Wikipedia's SpecialPages have some good ideas for page finders.

• Wanted pages has a list of red links, sorted from most common to least common.
• Orphaned pages has a list of pages not linked anywhere.
• Short pages, a list of the shortest pages, sorted from shortest to longest.
• Long pages, the opposite of the above.

These would be most useful with a namespace filter, page type filter, or both.

This could all be combined in a single search function, with the following criteria:

• sort method (number of wicks, page length)
• filter (optional)
• direction (ascending/descending)

The filter includes page type, namespace, or index. This allows for a huge level of flexibility. Red Link would be automatically considered a page type, similar to redirect.

The sort method could be extended to include other things, like the number of edits, or number of links within the article itself.

---

Also, a combined "Related to" list for an entire group of pages, using the same filter above (page type, namespace, index).

This could have a lot of use cases. You could search for YMMV index wicks on work pages, Tropers and Sandbox links on the main wiki, or a complete list of wicks towards an index's pages instead of checking them all manually.

Pretty much what I'm thinking about.

TroperWall / WikiMagic Cleanup

Note that some of Wikipedia's special pages would have a very different purpose as applied to this wiki. For work pages, a list of the most linked-to red links would indeed suggest a "wanted" page, but for most other types of pages (or even work pages that were cut for violating The Content Policy) a heavily-linked-to red link is probably a sign of a cut page that hasn't been sufficiently thoroughly scrubbed from the wiki. Meanwhile, a list of longest pages, depending on how it's measured, could be very useful to the too-long page thread in terms of identifying pages that need to have subpages split off.

Also, I could see a sort-by-references function and (to a lesser extent) combined "related to" list having an impact on Trope Overdosed and Overdosed Tropes.

I do wonder whether a tool like this would be able to be fired at will, by anyone, without straining the servers. I think we've had similar tools in the past but they were either taken down or made (or were to begin with) mod- or admin-only.

Edited by MorganWick on May 16th 2023 at 5:12:04 AM

One problem I see here is that this tool would have to scan all pages for redlinks (this will take a while), so instead it'd need to be a be list that's updated when the wiki is browsed or edited (server-wise it's likely already doing that when deciding which wick color to show, or getting indexes), while for missed pages a background task can do the scanning once.

Also note Frequent Redlinks, Permanent Red Link Club and Content Policy

Edited by Amonimus on May 16th 2023 at 4:08:12 PM

TroperWall / WikiMagic Cleanup

If it strains the server too much to have anybody generate the lists at any time, they could work on a limited schedule, say, once a day. Filtering would have to work differently:

• The lists could either be split up by namespace or page type (similar to namespaces in our current Related pages)
• The filter could be public but just work though the scheduled list (most linked red links, shortest pages, etc.).
• The filter could be manually tweaked by mods or admins, so they could choose not to include Laconic pages for example.

I don't want the tool to be usable by mods or admins only so that the workload isn't pushed to them.

The Related pages already show wicks to red links, right? I assume the database would already have a list somewhere of links across the wiki, without scanning each page.

Edited by N1KF on May 16th 2023 at 3:32:36 AM

Hm, that's true. The Related can instantly give a list of pages that link to it, so they may be pre-compiled somewhere already.

TroperWall / WikiMagic Cleanup

Came from the latest comment on my wish:

Staring at the Itty Bitty Wiki Tools for ideas... And landing on "Strip Spoilers". For something along that vein:

If we can get tools that auto-scan pages and stuff, then can we can get things that auto-edit for us? Running through a page's Related wicks and do a namespace change, for example.

If server load is an issue, I'm willing to be locked out from editing for however long the macro takes. because unlikely to be possible, but Seriously, if that's possible, I'm willing to give up editing for like a day each macro?

Disambig Needed: Help with those issues! tvtropes.org/pmwiki/posts.php?discussion=13324299140A37493800&page=24#comment-576

An option like Show Spoilers that opens notes by default would be nice.

they/them pronouns. Look at my Neocities.

If we have tools that also edit, I'd rather have it in a whole form that shows the differences and prompts if the user is sure they want to save.

TroperWall / WikiMagic Cleanup

There's a wishlist item for that. Just gave it a bump.

If we want to target pages that need work on quality, Zero Context Examples and a few related issues (e.g. unsourced claims for unreleased works) are one of the most common issues on the site, and have a standard markup - bullets get commented out with %% markup, a ZCE/citation %% comment (hopefully) gets added to the top of the page and someone (hopefully) notices the issue and edits the page to fix it.

On work pages, much of the time that's not going to be the person who spots it - it's easy to spot the problem (e.g. "Bad Boss: Helen"), but you need to know the work to add the missing explanation and make it useful/interesting to a reader.

...and once it's commented out, the reading experience improves, but the people who might be able to fix it are unlikely to even notice there's a problem, unless they're already editing.

If it's possible to identify the ZCE markup in page source (%%*, %%** or %%*** at start of a line) I can think of two options:

• Use that to trigger a top-of-page message for readers.
  • For a work page, a straw man version would possibly be "Some examples on this page have been hidden as they lack information. If you know the work, you might be able to help us fix that" (not sure if we can workshop a version that also works for Main/Tropes with the same text).
  • This wouldn't disrupt a casual reader's journey, but would help to encourage edits/engagement to improve the page.

And/or -

• Generate a report view that shows which pages have this markup, and how many lines are commented out on each page. Possibly split by namespace.
  • Maybe add inbound stats and/or the number of incoming 'related' links for each page
  • That should help tropers prioritise pages that really need help - e.g. a page with 30 ZCE examples commented out needs attention far more urgently that a page with 2. And a page with 4,500 inbounds and 300 links should probably be more of a priority than a page with 10 inbounds and 5...

No idea how practical this is, but if the community's got visibility of which pages need work, it'll get a lot simpler to target them.

Edited by Mrph1 on Sep 7th 2023 at 8:57:16 PM

Bumping this thread since this will be the next project im going to be working on.

As far as ideas for new pages or new content a user could add, what would be a place to compare or get ideas from?

For example, i could pull a list of popular movies/shows/characters from IMDB and compare with what we have and show pages that they have that we don't?

Now monitoring Wishlist and Bugs

We've got the List of Shows That Need Summary page for older works users would like to see on the wiki, and Unreleased Works for upcoming things that we know will get attention when they get released (but, mostly, aren't ready for a page just yet)... but both are very manual.

If we could do something to make that easier to manage — and to get more attention for things on those lists — that has potential.

Edited by Mrph1 on Mar 1st 2024 at 11:28:26 AM

I assume IMDB fine with it? I know they have open API, but if we want IMDB, that integration may need discussion by the higher-ups, as I've got an impression the site would like to minimize connections with other sites. I don't mind either way. If approved, maybe work pages can have release info data pulled from IMDB as well.

I think a list of redlinks (Wanted Pages) filtered by namespace and sorted by occurrence would help in multiple ways:

• Motives creating news pages
• Finds invalid wicks to cleanup

but as I've mentioned previously, I don't know how the site tracks colors for wicks.

TroperWall / WikiMagic Cleanup

The IMDB idea came from site owner. TV Tropes also doesn’t need to “connect” to IMDB. We would run scripts to pull information and put it into tables that we then pull from and put on our site just like any information we currently have.

Now monitoring Wishlist and Bugs

If so, it's fine. Was thinking either work pages can have an automatic (i) icon that shows a modal pop-up with IMDB info, or a [[IMDB:Work_ID]] markup to add it manually.

TroperWall / WikiMagic Cleanup