Repairing the Trope Repair Shop

As it currently stands, launching a new trope is vastly more convenient/accessible/pleasant than bringing an existing trope to the repair shop. This incentivizes people to create Ridiculously Similar Tropes rather than cleaning up old ones.

Right now this is on my mind because there's Step Servant draft in TLP. It's Adopt-a-Servant except for stepchildren and bastards. When I suggested expanding the scope of Adopt-a-Servant instead, no one disagreed with the concept — but no one thought it was very feasible. I completely understand — I've brought stuff to TRS before and it was exhausting and took forever, and I dread doing it again.

I think it's pretty widely agreed that fixing up old tropes is better for the site as a whole long-term than creating new Ridiculously Similar Tropes. I know this is a hard-to-manage logistical problem, but is there anything that could be done to change the structure of how TRS works so that people won't dread and avoid it so much?

Edited by Eievie on Oct 18th 2023 at 6:28:45 AM

This may be a conversation to have, but it does not belong in Projects. I'll move it to Wiki Talk.

~Eievie — just so you don't think it got deleted or something.

Like I said in TLP, it is something to be discussed, Being the sponsor for Step Servant, I don't want to see my time spent in vain, so I would like to know ahead of time how it will go.

Oh, we have not made it for bastards yet; it was an idea to move those to this trope were it launched. I'm currently looking at the proposed trope as being for anyone who is a relative that is made into nothing more than a servant/slave for the rest of the "Family".

Edited by Bedford on Oct 18th 2023 at 9:58:54 AM

I only mentioned Step Servant as an example of how this dynamic works. I do not mean to discuss the specifics of this instance here — this is not the place. I want to talk about what could be done to mitigate the circumstances that created it and instances like it.

I did not at all intend this as in individual callout post, and if it came across that way I'm very sorry. I just wanted one example for demonstration to show it's not hypothetical.

Edited by Eievie on Oct 18th 2023 at 7:17:57 AM

I think the problem stems from a couple different sources.

• First, and hardest to change IMO, is the interest in TRS. We can do all the restructuring we want, but if we have the same small number of tropers willing to come to TRS and put in the work to fix tropes on the regular, it's still going to be a slog. Recent events have not helped as we've lost some regulars. We can only do so much with so little manpower.
• Number of TRS slots v. number of tropes on the queue. Many, many tropes need attention, and they are being brought to queue faster than we can clear it. The TRS only allows 50 open slots at a time and, with two projects with 5 digit wick counts currently open, we technically only have 48 slots that can be used for the foreseeable future. I am not advocating for more slots. I think the limit we have now is fine and even now we have open projects just wasting away. This also ties into manpower above. More manpower = faster wick cleaning = less wait time on the queue. One thing I think can help that I saw mentioned like a couple months back was to have a "fast-track" thread that would deal with any tropes considered starving, as they take very little time to discuss and clean.
• Badge of Honor mentality. Let's face it, some people get really excited about being able to declare they created a brand new trope. There isn't really that much of a BoH mentality in TRS, and cleaning a trope is a less obvious achievement than making a brand new page. Unless you put it in your signature or on your troper wall, no one but the other TRS participants will know how much work you put in to fixing a trope.
• There's also just the fact that it can be quite easy to pass a trope through the TLP with the way things currently work. I'm hoping the upcoming changes will help fix this, but currently, if a draft gets a large number of hats, it can be easy to steamroll through questions and concerns. True, things get sent back for having unresolved issues, but I'd rather they not launch in the first place until everything is squared away.

I've suggested this before in the meta thread, but one idea would be to have a month where no new TRS threads are opened. The sole objective of TRS for this month-long period is to clean projects currently open. Maybe we could identify a project and have everyone focus on it, focusing our current manpower on one project instead of having everyone split up over various projects, cleaning and closing a thread faster, and then focusing on the next one, etc.

Cleanup is a pretty thankless job, and can often take up a substantial majority of time a thread spends in TRS. Creating or changing a trope is fun; changing each link one at a time is not. There have been proposals to add a tool to mass-change all links to renamed tropes, but that only works if there's no misuse or ZCEs at all, and given current TRS standards, that leaves only underuse that isn't severe enough to warrant a cut as a situation where that would happen, and even that fairly specific situation requires a lot less work than most threads. (You could also have pages that have been expanded to fit misuse, but it's probably rare that that would involve a rename, and assumes every example fits the expanded definition.) That means you have to not only go through each individual contextualized example, you have to figure out what to change it to, which is significantly more difficult, especially if you might not have a lot of brainpower at the moment or didn't participate in the thread and don't have a good grasp of each category of usage, and especially if it doesn't fit any of those categories, which is more likely with larger tropes where things can fall through the cracks of a usage check. It almost takes the same skills as a usage check, but for every single link the trope has, which can be thousands of them. Not every thread is like this - at least before the recent unpleasantness, most threads wrapped up within a month or so - but it can be hard to predict when they are. And if you just change things on autopilot as though you did have an automatic link changer, you just create more work for later.

I don't know if there's much that can be done about that. Making links to disambigs green helped, but the thread I tried to help with cleanup on months ago, the first link above, is still open.

I mean... yeah, the big thing it always comes back to is the lack of willing manpower. And not just willing manpower, but people who will do that work correctly. See, we've actually done really well at getting more people to visit TRS threads in the opening phase to suggest ideas and vote on crowners. The issue is that activity drops significantly when cleaning starts, and right now I think pretty much every thread is stuck in that phase or is struggling to come up with a good solution. And while we can occasionally get people to help out with a bulletin, this is a double-edged sword because it may instead result in people "cleaning" by just renaming every trope without checking to see if it actually fits. This has happened multiple times before, and is the result of people going for quantity over quality... which means that paradoxically, sometimes slower periods results in better overall work since the people who are tackling those cleanups are less likely to just fly through it.

That said, I'm not opposed to the idea of a "thread closing month" or something. The issue I have is that it's already required for threads to close to get new ones in, and people still aren't doing that. These issues used to be even more severe before the cap was reintroduced, as not only could people just fill out the thread with any project they wanted to tackle, they would often be open spontaneously while projects still in the TRS could sometimes take years to resolve. TRS has actually gotten a lot more efficient in recent years — but that doesn't mean it's a speedy process, because a lot of times going fast only leaves a worse mess. (This is why I'm absolutely opposed to the idea of any sort of automation or script — I don't trust it to not just either delete everything or change everything over without any regard for context or content.)

The two logistical things that first come to mind are:

• Add TRS to the "Important Links" list. TLP is already there, and I think being visible so often puts it in people's minds more and creates activity.
• The structure of the TLP page where launches which are active and there is persistent interest in float to the top of the list, whereas ones people aren't so interested in sink to the bottom. In contrast, TRS is more like a legal docket where everything is weighted the same regardless of interest or activity. Maybe it could be structured more like TLP, but instead of "launching" when it's done and ready and has consensus, it goes to the "to implement" list or something.

Edited by Eievie on Oct 19th 2023 at 9:45:28 AM

The question I have is whether trope expansions necessarily have to go through TRS. Perhaps TLP could have a side-function to launch expansions rather than creating too similar sister/sub-tropes as mentioned with "Step Servant".

That's a good point. TRS deals with several different categories of repairs, and they all go through the same channel. The one channel is primarily concerned with wicks, which aren't relevant evidence for all kinds of repairs.

For example, I remember trying to merge some tropes and being asked for lists of wicks, and it was really frustrating. Whether or not people are currently using a trope link for certain kinds of examples doesn't really say much about the what is or isn't the same trope on a conceptual level. For that, it would be more useful to talk about the idea and get a consensus from as many people as possible — much more like TLP.

At least at first glance, splitting TRS into "trope is being misused" and "trope concept needs reexamining" sounds like a great idea.

They do, because you're changing the definition of a trope. TLP is not an effective place to gain consensus for this sort of thing because the nature of the area means that feedback is limited to hats and bombs — and while updates are coming eventually, it still won't be a great place to go to do anything that impacts the way a pre-existing trope is used.

We see how someone even just adding a stray line onto a description can really mess up how people view and use a trope. That's why TRS is needed for these changes, even if they seem small.

I'm not suggesting an exact replica of the TLP setup, just something that takes some elements of the TLP setup. Clearly this is its own thing and should be tailored accordingly.

You ninja'd me - I was actually responding to the post above.

I am 100% behind anything that makes the TRS less frustrating and time consuming. I think one of the best lines in this thread is "The one channel is primarily concerned with wicks, which aren't relevant evidence for all kinds of repairs."

There are a number of tropes I've wanted to repair. Started the wick check in a sandbox...only to quickly realize that it doesn't actually prove the issue I see. A great example I tried recently was Characterization Marches On (Trope Talk thread here)... only realize that the issue I was trying to prove (many examples are simply of standard Character Development) was impossible to do unless you're familiar with the work to know if the development is actually shown and justified within the work. I can tell by examples from works I am familiar with that it's being misused...but the wick check is supposed to be random, and I don't know the works.

But overall, yes, it is far, far easier to launch a trope than to repair one and it's kind of a problem.

I mean, the importance of wicks makes a lot of sense to me personally for most of these efforts:

• Duplication? The case would be a million times stronger if you can prove the usage is identical. This can be hard to run a wick check for, but that doesn't mean it's impossible — I've done it many times.
• Expansion needed? Prove that the scope of how the trope is used is broader than how it should be.
• Narrowing needed? Prove that the scope is broader than how anyone is actually using it.

Etc.

I genuinely don't understand what people mean when they say wicks aren't useful for these kinds of things — they very much are, if you know how to form the wick check.

I dunno, the example Beer Baron brings up is genuinely problematic. You can have an example that's perfectly contextualized for CMO but misses the elements that make it Character Development, and it's impossible to know if it's being used as a PCE for Character Development, if it's being used correctly but the troper missed the Character Development, or if it's being used correctly but the troper doesn't think the Character Development was actually Character Development. The only way to know for sure if they're using the trope correctly is if they include something like "this is never explained", which could be repetitive, verges on Captain Obvious, and could come off as whiny and complaining. I'm not sure making it YMMV is a good solution, and I'm not sure making a cleanup thread would work very well.

I'm not sure exactly when we started mandating usage checks for every non-Not Thriving thread (for a long time my activity on the wiki was rather off-and-on), but I feel like it might not be that long ago. Not that it wasn't warranted - I can easily imagine a thread started on "vibes" that goes on for a while only to find the problem doesn't actually exist outside of, like, one troper (or in one case, no tropers). But there definitely can be problems that a usage check isn't well suited to pick up, and there definitely are threads where usage checks can feel like mostly-irrelevant spackle. (Granted, a lot of times that's because it's a Not Tropeworthy thread where the OP doesn't seem to be trying very hard to see if there are any tropeworthy patterns, so it just comes off as saying "yup, sure is a thing that exists" and reaffirming that it isn't tropeworthy.)

To be clear, I'm not arguing against mandated usage checks, only suggesting that there can be cases where they may not be very good at picking up the problem, or where they're only useful if you pay close enough attention to do them right, which can raise the barrier to entry or success for new or casual tropers. (Besides Not Tropeworthy, this can also apply to Duplicate Trope threads where the OP may miss fine distinctions between the tropes and have usage categories so broad that they can't catch the distinctions and where a verdict of duplication seems inevitable.)

That's a specific circumstance though, not a case against using wick checks for other similar issues. And if such a check can't demonstrate an issue, how can we make a case for the Trope actually being misused outside of those examples Beer already knows about? If it's actually that hard to tell, well... then what if it's not actually the big deal it's believed to be?

This is why I believe checks have become basically mandatory over the years. They prove that the problem does definitely exist. And even if some cases are more difficult to check, in those cases we currently lack hard evidence that the issue exists in a large enough scale to warrant TRS. This is also why I personally think it's backwards if someone makes an OP before the check is done- they're assuming they know that the problem exists and aren't letting the data be the deciding factor.

Edited by WarJay77 on Oct 20th 2023 at 9:15:54 AM

I don't think anyone is claiming that we shouldn't do wick checks, they are helpful in most cases, but there are cases not captured by them like I just described. Perhaps rare, but wicks shouldn't be a "be all, end all" type of evidence either.

Another one I started to nose around recently was Lethal Harmless Powers based on the thread here. Again, a raw wick check doesn't help with the problem the trope seems to have, which is "what does 'harmless' mean"? Even within the thread responses, there is uncertainty. I could count it up based on how I believe it should be definied...but is that actually right?

Anyway, just my $0.02. Like I said when I saw this thread, I'm all for evaluations of how the process currently works, as some problematic tropes don't neatly fit the current system for being fixed.

I know nobody is advocating the end of wick checks or anything. I guess I'm just concerned that if a wick check can't provide evidence for one reason or another, well... what evidence would be suitable? What could prove the problem is real beyond a small forum discussion? I'd agree more that they shouldn't be the end-all-be-all if anyone could provide an actual alternative route for these harder tropes.

As for the TRS itself, I'm really not sure what we can do. If people don't want to do the work there's nothing we can do and that is what I believe the main issue is. Putting the threads together may be harder up front because work is needed to make a case- but it doesn't matter if the entire process becomes streamlined if people don't want to clean up later.

Another factor that's slowing down TRS are the cleanup efforts with 4-digit (and one 5-digit!) wick counts. Those naturally take a while.

There's something we have to do regarding those. A lot of those stagnant efforts involve green links, which are easy to spot, but not many know what the green text signifies. Perhaps the admins can think of something, even something as simple as a new edit tip.

Edit: One thing I can think of that can reduce the backlog strain would be to offload the quick efforts into a dedicated thread.

Edited by Berrenta on Oct 20th 2023 at 7:08:27 AM

I am not sure that trying to split quick efforts (how and when do we define quick efforts?) from non-quick ones will contribute to get the non-quick ones resolved. I suspect the opposite is more likely to happen.

Come to that, maybe we need to ask ourselves how many of the things on Sandbox.TRS Queue need an actual repair effort. The Super Weight thing does not seem to be worth of any action, for example.

It's proposing a move to JFF and my impression is those sorts of moves, under current rules, require TRS.

At one point we discussed the idea of limiting threads involving pages with high link counts, and I discovered more than a few threads that didn't have super-high link counts but still took a long time. And a lot of times we won't know what solution a thread will cone to until we actually have it. So any sort of segregation of high-effort threads from low-effort ones, with the possible exception of Not Thriving, is hard to imagine how it would work.

Edited by MorganWick on Oct 21st 2023 at 4:48:25 AM

Personally I think 50 threads is far too big a limit since most of the time 95% of those threads are dormant. As of me writing this the bottommost thread on the page had its last post back in June. I think reducing the limit from 50 to 10 would help in the long-term. I know it's not a very effective short-term solution, but fewer threads mean quicker wick cleaning efforts.

How much evidence is there, really, that the limit actually speeds cleanup any? I mean, it seems like what you're proposing would, at best, lead to us going through threads at about the same rate, but it would feel more productive because each thread would have to wrap up five times faster to maintain that rate.