TV Tropes 2.0: Database level redesign (Not in active development yet)

That sounds interesting. By having a council of tropers assigned to each area of repair, we might actually be able to keep them focused on dealing with the topic at hand. See I think that the reason why the trope repair shop repairs are never implemented is because no one feels obligation to do so. I always thought that if you give someone a responsibility, they would feel more obligation to fulfill it. I know I would.

Huh. I wonder how one might appoint these people.

One of the things I've asked for is a function to do mass edits spanning several elements simultaneously. That function would create an interface where one can edit various elements or specify actions for them (e.g delist, move to other trope/work), a few hundred elements at a time, and then the system implements the changes. Could make trope cleanup much easier.

@Fighteer: So, in future there'll be just one discussion thread per wiki page? - But aren't there pages with wiki discussions which have more than one?

@Septimus: A form with several hundred elements? That sounds scarily big...

No, there won't be one discussion thread per page. After all, there won't be "pages" anymore - just discussions per element.

As for the hundred entries thing, I do sporadically process a few hundred cut requests at once. It does lag noticeably even at smaller amounts.

Not if you've done database work with posts/entries in the millions. Although that's for automatic or sweeping changes, with or without some algorithm for variable output. But even if you're doing it manually and checking each entry, it's a bit tedious above a hundred. Depends on the amount of work for each entry, naturally.

@FELH 2: No, there won't be just one thread per wiki page, but there may be one "main" thread that's linked to the applicable media forum that will serve as the general discussion, plus additional element-specific threads, such as discussion about a particular example or a particular image.

It is my hope that most conversation will use the general thread.

How many threads should be connected to a given discussion page and how the connections are established are two of the biggest hurdles of any forum/wiki integration.

Hey all. As a long-time reader, I've become quite fascinated with the prospect of the database-driven 2.0 site, so thanks for this thread! It should be quite informative in the development period to come.

As such, I have a few questions/suggestions/nuggets of food for thought:

• How sophisticated will the filtering options be? I mean, it would mostly be used for "show me examples that involve Legolas," but can it do "show me only subversions of Giant Robot Hands Save Lives from Mecha Anime in The '80s"? That kind of thing would be pretty cool. Presumably, examples could be tagged as Subverted or Defied or whatever, which would better explain the relationship between the trope and the work.
• If works can be subdivided into seasons and individual episodes, would this take the place of Recap pages? I presume then that you can add an example as happening in a specific episode, and it will show up if you look at the page for the entire show (but sorted by trope as opposed to episode). My concern here is that if something happens practically Once an Episode, you might end up with an example with a crapton of sub-bullets (although your solution of collapsing those sub-examples appears to be a good one). If we have tags for subversions and such, those could take precedent in the sorting mechanism (i.e., it happens every episode, but this time it was unusual), which is usually how such examples are listed.
• The character tags strike me as important here. Presumably, a character would be tied to a work or franchise, but can be used in different works. My thinking here, though, is that if a character is used in a derivative work (like a Fan Fic or an Abridged Series), you don't necessarily want that character's examples from those works showing up in the parent work. I mean, imagine reading the Harry Potter character entry only to find a completely nonsensical example because it came from Thirty Hs or something like that. Would it then be possible for a derivative work to "borrow" a character from another work? It would be explicit on the derived work that the character is from somewhere else (and you could link to that character's parent work), and that derivation would not appear on the main character page.
• Will there be a way to define the character's relationship to the trope? For example, if I wanted to see Armor Piercing Slap, what if I wanted just the list of characters doing the slapping rather than the ones who get slapped? It's like a subject and predicate in a sentence, only with a trope instead of a verb. This could also be done with works or sub-works (e.g. Call-Back linking from one episode to another). Then again, that would be relatively hard for Tropers to get the hang of. Just curious as to whether it could work.
• Again with characters — how do we handle different versions of the same character? I mean, you could treat them as separate characters, but with such tropes as The Nth Doctor, it might cause issues. Would you tag such examples with an overarching "The Doctor" super-character, with specific examples pointing to specific incarnations?
• Are negative tags possible? That is, going off the Doctor Who example, let's say that a certain trope applies to all versions of the Doctor except one (say, the War Doctor). Would you then be able to tag the example as applying to "the Doctor, but not the War Doctor," and it would thus show up in all the individual Doctors' character pages except the War Doctor's?
• I know we're not IMDB, but I wonder if it would be possible to add certain data points to works just for reference's sake? I mean, if you wanted to see a list of every film a certain actor was in, IMDB would be better for that purpose, but only we have the tropes . We might also include things like country of origin or release date for indexing purposes. That would allow you to sort works chronologically, automatically create indexes like French Literature or The '80s, and allow you to group examples more easily in a creator page.

To get back on track with prior discussion — I support the idea of discussion pages and forum threads linking back to the article. But I'm not a big fan of straight-up anonymous voting for the reasons previously stated. I believe many Tropers come to such votes with agendas and such which are harmful to the wiki (such as Fan Myopia), and if there's a verbal discussion, they're more pressed to put their thoughts out in the open, which might expose the Fan Dumb ideas for what they are. But I did suggest a while back that individual Tropers' accounts could allow them to specify works they like or are familiar with, and the database could try to display for them images from those works assuming they'll be familiar with them. (And it would help avoid the occasional problem of a given work's fans replacing perfectly good images with images from their favorite work.)

Sorry for the word vomit, but my curiosity is getting the better of me. Your work on this is much appreciated — even if it's just getting started!

That's a lot of stuff. Let me take a crack at your questions...

• Meta-trope tags on examples is something we are considering, as long as they don't get too overused. There's no reason why complex filters couldn't be made available, but it certainly would not be a primary design goal. It's more of a down the line enhancement.
• It is intended that the work hierarchy replace the current Recap pages. The problem of lots of examples in the parent lists will definitely require some thought; I envision two ways of handling it.
  • Collapse the list so it's expanded on request.
  • Permit a "parent" example to override child examples... if The Hero is on the main page for the work, the main page will only show that example.
• That's not a bad idea. I have in mind that examples on a character's main landing page will always be sorted by what work they appear in, much like a work article and its various subworks. Examples from derivative works can be identified as such so they can be filtered out automatically if you don't want to see them.
• I don't see an easy way to make that kind of "actor-actee" relationship to be expressed in the example editor. It's not impossible but it is difficult.
• With "super-characters", I imagine we could use the same hierarchy we do for groups of characters. "The Doctors" would be considered a group, of which each individual incarnation is a member.
• I would prefer not to use negative tags. It creates too much potential for problems. I'd rather we fall back to the "super-example" idea I mentioned earlier.
• We intend to have meta-tags on works, such as date of release, medium, genre, and setting, as well as creator associations (production company, developer, director, actor, voice actor, etc.). We aren't IMDB, but there seems to be no reason why we couldn't copy some of their ideas. Certain meta-tags will be used specifically for disambiguation, in place of the current method of creating a disambig page and appending the year or creator to the title.

edited 28th Nov '15 5:06:08 PM by Fighteer

If it's a derivative work (as opposed to a Spin-Off or other official work), wouldn't we treat the characters as completely separate from the canon ones?

Negative tags is something that should probably only be handled while sorting and filtering. I've seen it in databases, and it tends to cause more problems than it solves. It's sort of like a database version of spaghetti code. Or maybe more like goto commands.

Works will be identified as derivatives, so we can design the views any way we like. By default, derivative versions of a character won't be shown when you go to that character's landing page, but there will be links to view them, assuming that the relationships are set up properly.

The way I see it is that tropes ubiquitous to a work won't need example entries for every component of that work, unless they are exceptional in some way. Aang is The Hero of Avatar: The Last Airbender; the example for The Hero should attach to the work as a whole, not to each individual episode.

Examples should be attached to specific seasons or episodes, as opposed to the main work, only when they are exceptional or specific to that subset.

Question: Is there still a point in continuing to cross-wick? I don't want to waste time on duplicating information that will potentially be discarded due to a restructuring of the site at a later stage.

This is a long way off, if at all. Don't stop doing things you should be doing just because we might be overhauling the place some day.

The more pages are according to form, the smoother it will go to transition to the new system.

edited 29th Nov '15 3:05:10 PM by AnotherDuck

"It is intended that the work hierarchy replace the current Recap pages. The problem of lots of examples in the parent lists will definitely require some thought; I envision two ways of handling it.

• Collapse the list so it's expanded on request.
• Permit a "parent" example to override child examples... if The Hero is on the main page for the work, the main page will only show that example."

Why not both? One parent example for the overall work ("Captain Bob is The Hero of Wagon Train to the Stars because [context].") that's visible by default, followed by an expandable note saying something like "There are 2 season-specific examples and 37 episode-specific examples: click to view" that would lead to a list of examples like "Lt. Alice is The Hero of her Day in the Limelight episode (s2e10) in which she [context]."

edited 30th Nov '15 7:52:24 AM by HighCrate

A combination is certainly feasible; it just depends on the amount of programming needed to pull it off. I'm a whiz at databases, not at web development.

Skimmed through the "new" IP discussion, so I might have missed some stuff.

I like the current system because the tropers that currently frequent IP tend to be intelligent and know what images work best, and I would prefer to keep it that way. Going to sound like a bit of an ass, but the more people we have contributing the more we get who don't know what's good and it will affect overall image quality. If we were to change it up at all, I would rather have a "council" as somebody brought up comprised of a body larger than the P5 (11 tropers minimum) with the ability to ask for outside opinion like the P5 used to do sometimes. Caption discussion can probably be more lax since it's not as big of a deal.

That is anti-democratic :)

I just think choosing images requires a bit of finesse. I would prefer no change to the system at all, if it's possible, but I would not be against a council system.

It does require finesse, which can be learned by anyone who is interested.

I think the most important and valuable aspect of the current Image Pickin' system is that page image changes are almost always discussed, usually before they are made. I embrace change, but I think any new system should maintain (or improve!) that step. I like the idea of mostly leaving the IP forum how it is, and letting IP decisions overrule the semi-automated vote selection. That would indeed be similar to the current system, but I'm not sure it really meets the "wiki/workshop integration" goal, and I suspect we have an opportunity for more dramatic improvement.

edited 3rd Dec '15 2:08:26 AM by rodneyAnonymous

P5 worked because most people either didn't want to confront the issue too much or just wanted to keep all the creepy stuff on the wiki at all costs, so the people whose opinion we actually valued were fine with delegating it to a panel of experts. That wouldn't be the case with everyday functions like IP or TRS.

It sounds like we want to broaden Image Pickin' discussions to a wider body of tropers without turning it into a Quality by Popular Vote contest where our image standards end up completely abandoned. Sticking the discussions to the page itself would be one step in that direction, the question is what else would need to change under the new system.

A few thoughts of mine on that issue, posted on the previous page:

Some other considerations: How do we arrange Image Links for the 2.0 design? Keeping in mind that most Image Links things cannot be used as images because a) they are not hosted on TV Tropes and b) they were "good images but there is a better one for the page".

Could use two versions of them: uploaded images and external links. The latter would be excluded from wiki-specific functions, like if we'd have that voting system that automatically replaces images on trope pages. Could still have the option to vote on them, but they wouldn't show up. I don't think it'd be too complicated.

I was more wondering on the database object structure. Would it work like an example list, with each link its own element? Or would it be a free form edit page?