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

The discussion of Image Pickin' reminds me: what sort of ideas are out there for how to handle TRS functions?

TRS is, obviously, much more complex than IP in terms of the kinds of decisions that need to be voted on. I don't see that simply attaching votes to things is going to help there. I do envision integrating article-specific discussion topics between the wiki and the forums, so that a user searching for information won't have to go someplace else to get it.

How this will work, exactly, is still up in the air, but I'd like there to be an "article discussions" section of the forums where all active threads may be viewed, ideally categorized by article type. As now, "repair" threads should be tagged with specific keywords that will identify the issue. The main difference will be that those threads are also directly visible (and can be posted in) from the article itself. (General discussion topics, such as are currently in our Media subforums, would be kept separate but should also be cross-visible.)

The database overhaul will make most aspects of article maintenance much simpler:

• Renames can be accomplished simply by modifying the active title of the article (inactive titles will serve as searchable redirects). Since wicks and example links will be stored using the article's internal ID, all the renderer has to do is look up the current title when it displays those pages. Obviously, any wicks that no longer make sense given the article's new name would need to be edited.
• Example cleaning will be much simpler given that all examples will automatically crosswick. Deleting, editing, or moving any example on the trope page will automatically have the corresponding effect on the associated work page, and vice versa.
• We envision a "batch cleaning" mode where you can select multiple examples and perform an action, such as moving or deleting, on all of them at once.

edited 23rd Nov '15 10:58:57 PM by Fighteer

Question: The new setup will do away with wrongly categorized works on trope pages, right? Like a work opened under namespace Series will always appear under that category and not Web Original as it is possible now.

As long as the work is categorized correctly, any examples from it will be categorized automatically on trope pages. So the answer to your question is yes.

It'll also do away with those awkward subpages (soft or hard split) that contain examples from two completely different series of the same name, simply because subpages don't have the medium namespace the main work pages have.

I don't think the problem is not trusting tropers to make good decisions about a page image, it's not trusting that system to result in good decisions about a page image. The current one involves a lot of nuance about what "good decisions about a page image" means, whereas the system described in #62 seems like it would devolve into a work source popularity contest. It's not true that "we already vote on images in Image Pickin' topics", at least not completely... only some of them get voted on (and that is always after discussion). More than half of IP threads are resolved with no crowner. Now and then the forum even decides that no image at all is preferable to any of the suggestions.

The image equivalent of a Zero Context Example is Just A Face And A Caption. Whereas effort is being expended to make ZCE entries impossible, that system would make JAFAAC images more common. (Page images posted without IP discussion very often have that problem, but images picked by IP very rarely have that problem.)

I appreciate the desire to integrate important decision-making that happens in the workshop forums into the wiki proper, but I think there are valuable elements of the IP forum that would be lost that way. Maybe introduce an approval step of some kind, before an image is added to the roster of images that can be voted on? Mandatory discussion somehow? Something completely different?

edited 25th Nov '15 1:57:58 PM by rodneyAnonymous

I think that can be solved by allowing IP to overrule voted images if there are problems with them.

Currently anyone can add an image without discussion or any form of consensus. How is a voting system worse than that?

Overturning a bad image that was added by a vote is much harder to do than to overturn the same bad image when it was added by a singular troper.

Why?

It's not hard, technically. A moderator will be able to place a check mark in a box and lock an image to a page. The difficulty is social: convincing the troper body to accept an image that is different from one that demonstrably received significant voting support.

Specifically, Septimus and others are saying that we can only trust people with Image Pickin' experience to cast educated votes on images. This is profoundly anti-democratic, and yet we have seen evidence that Quality by Popular Vote is a terrible idea when it comes to many wiki decisions.

My concern is that we have isolated the decision-making portion of the site down to a tiny space that is populated by a few dozen active tropers, out of a body of tens or hundreds of thousands. We've said many times that we want to reach out to the general body of users to get them more involved in the decision-making, and exposing the process directly on the wiki would accomplish that, but we seem paranoid about those same users drowning out the established consensus by sheer volume.

We want to have our cake and eat it too: complain that not enough people participate whilst raising immense barriers to participation. Democracy can be kind of painful sometimes.

edited 25th Nov '15 4:27:48 PM by Fighteer

I don't think that's what anyone is saying. There are usually a lot more votes in an IP crowner than there are commenters in the thread it's attached to, and nobody has a problem with that. (Quite the opposite. It's explicitly preferable to have more total votes.) It is not anti-democratic to have debate and discussion before opening a topic to a vote, nor to have requirements for what appears on the ballot, nor (and this is not the current system, but I think it's analogous) to have the public elect educated experts to vote on something rather than having the public vote on it directly.

edited 26th Nov '15 12:05:50 AM by rodneyAnonymous

It seems that the alternative to having integrated voting in an article's image gallery is to let users simply pick which one gets to be the front, with the next user to come along being able to do the same — basically what we have now. I'm not convinced that's a superior alternative. Why not, at the very least, have a formal voting system for images rather than the ad-hoc system we have now that involves mangling crowners to include them? If there has to be a consensus to open an image vote, then we can figure something out tech-wise to make that happen.

edited 25th Nov '15 4:36:46 PM by Fighteer

I am a fan of democracy. That's a big part of what draws me to the wiki structure in general. More people voting and fewer barriers to vote are great. What I disagree with is that specific implementation: I think several aspects of the current system result in higher-quality page images on average, and a simple popular vote might not replicate those aspects. (For example, I don't think popular vote alone would have picked the page images for the Nightmare Face; Exit, Pursued by a Bear; or One-Winged Angel articles.)

edited 26th Nov '15 12:32:21 AM by rodneyAnonymous

This may be the wrong place, but will the forums change that much with the new redesign?

We expect tighter forum-wiki integration — for example, the media thread for each work being visible on the work's wiki page. The formatting and functionality will be improving with the current beta overhaul and its later iterations; there's nothing specifically on the table for the 2.0 overhaul.

We've had issues in the past with Quality by Popular Vote, precisely because of Fan Myopia. See the mess that forced the administrative fiat with Nightmare Fuel. Or some renames like The Mario or The Libby. I also remember complaints about such patterns occurring in IP, too.

That is for things with anonymous no comment voting, anyway, like crowners are. The Image Pickin' system is sort of a commented open voting system as-is and it is much better at avoiding that specific issue. A closer integration with the wiki may improve the input of opinions, too - I see that the threads are often short of them.

As far as the forums go, I am fairly sure that an overhaul of them is not part of the "database" change. I've been pushing for closer integration between forum threads and wiki pages, though, and others have made such proposals as well.

Instead of saying how an idea's bad, why not figure out how to better the ideas instead? After all, there's no such thing as a bad idea. Simply an idea that was failed to be worked out. Besides, it's a bad design practice to outright reject any idea one comes across without seeing if it could be improved upon.

The problem with "Quality by Popular Vote" can be solved simply by adding some sort of features to it catered to moderators, like giving the mods the ability to lock pictures in place (so that they don't change to problematic ones), give them the ability to override the popular votes to a more suitable one, allowing them to remove problematic images from pages, etc.

edited 26th Nov '15 7:25:08 AM by KarjamP

That would require moderators to handle all potential problem image choices. Nah, sticking to the current method of discussion and occasional vote is much better. A closer integration with the wiki side is a must, though - I frequently come across discussion page queries and suggestions about images.

We're supposed to be brainstorming ways to improve the wiki, not dismiss possible innervations like as this idea since you feel the current way things work doesn't need to change.

The best ideas usually start off crazy. That is, until they're been refined into something that works greatly and perfectly. This is true with anything that involves designing, be it a video game, a character set to appear in a show and even software like the one powering up TV Tropes.

To be honest, Septimus Heap, I don't really think you want to be convinced that this is a better system as you seem to prefer the old one. If I attempt a counter argument, you'll try to make another excuse as to why we should stick with the older system. This type of stubbornness isn't productive. In fact, it will lead you nowhere.

Karjam, I am not under any obligation to not dispute a change that I think will degrade the quality of the wiki. Part of discussions about a change is to discuss whether a change - or aspects thereof - will improve the wiki or not. And in fact I am not convinced that voting on images by default is an improvement and will thus argue against it.

You're unconvinced it's better than the current system. Therefore, you're going to keep on arguing about it. That's what you said, right?

If an idea has problems, the better solution is to brainstorm ways to fix the problems, not to immediately dismiss it as being unsuitable. Stubbornness arguing is rarely productive as that can lead to nowhere fast. Especially if the proponents of the idea are equally as stubborn.

edited 26th Nov '15 10:56:53 AM by KarjamP

Sometimes the solution to the problems with an idea is to not implement said idea.

The things that need to happen, for me, are the following:

• Closed IP and TRS topics need to be archived on the pertinent discussion pages (which will be split between various elements) where one can see them.
• Non-ephemeral forum threads need to be connected to discussion pages. For example, the quotes and trope description threads need to be linked on the trope and quote discussions by default. Similarly, work pages should be linked from the work discussion pages.
• One might consider have IP and TRS topics be transcluded on the pertinent discussion pages (which will need a way to deal with the ban systems, however).
• One might consider allow IP and TRS threads be started directly from the discussion pages, without going to the forum.
• One might consider have some discussion pages by default to be links to forum threads, rather than discussions in their own right. Essentially, replace the discussion buttons with links to the pertinent forum threads.

I suspect that most complaints about important decisions happening "behind the scenes" are related to decisions made in TRS.

You're basically proving my point by reasserting the better solution is to not implement the new idea. You're not arguing for the sake of figuring things out. You're arguing to with the intent to convince others of your views. Neither of us are right, and neither of us are wrong. If I was also arguing with the intend to convince you we're right, neither of us would convince one another as neither of us are willing to admit wrong.

To get back on (proper) topic, why not merge our ideas together? We have "popular vote" on by default, and if we find Quality by Popular Vote is in effect for the current image, we open an "Image Pickin" using the system Septimus Heap proposed, override the image chosen by "popular votes" with the one chosen by "Image Pickin" and disable the "popular votes" feature for a specific page.

edited 26th Nov '15 1:00:33 PM by KarjamP

I am fully in favor of the idea of merging discussion pages and forum threads in some fashion.

It does seem to me that any decisions not made by popular vote would involve some kind of fiat by someone. Since the current structure doesn't really have any troper "ranks", that would logically put it in the hands of the mods. The alternative...well, I have an idea of my own.

I'd propose to copy the idea behind the P5; different councils of tropers, each assigned to enforce some aspect of wiki policy. One group for IP, one group for each major category of misuse that would bring something to TRS, and so on.

Assuming some aspects of democracy are to be preserved, some of the people on these councils could be elected by regular tropers (subject to mod approval, of course). Major changes could have the members of all the different councils get to vote. That way, issues and changes would be voted on...but the number of votes on each issue would be limited.

The P5 has worked pretty well, so I don't see a downside to trying to replicate that for other aspects of the wiki.