Proposal: Allow play time for VNs with partials

Posted in

#1 by myou
2022-11-07 at 01:45
< report >I'm proposing a change to the current policy which is described in this post and conforms to the guidelines for length in d2.2.

I want to end the policy of not allowing play time/length ranges for VNs like Ciconia no Naku Koro ni, where a playable game exists (Phase 1) but it's not considered the "finished" game because further content is expected to be added.

So for example, Ciconia Phase 1, the only current "Phase" released so far, is 24 hours long on average. That puts Ciconia in the 10-30 range. What I'd like to do is have play time appear like this:
* Play time: Medium (10 - 30 hours, 33 uncounted votes)

But this is the counter-reasoning behind the current policy: "No, that's a bad idea. If you do that, and then Phase 2 is added in the future, this VN will probably be longer than 30 hours. Then 10-30 hours will be wrong. It should look like this instead:"
* Play time: Unknown (33 uncounted votes)

(For those not aware: the votes are called "uncounted" here, i.e. their average isn't displayed, because it's possible that any number of them could be cast either before and after subsequent content is added to the VN, and there's no mechanism to reduce the weight of votes that targeted outdated content lengths. In any case, this is a separate issue from play time ranges.)

My response to this counter-reasoning is... what's the problem with just changing the length interval from Medium to Long after Phase 2 comes out, then? Why not let people see how long the VN that they are actually considering playing today is?

This is the core of the issue: I object to the idea that VN length should refer to some [future] length of the VN when development is [eventually] completed (and thus be unknowable for in-development VNs). This idea is the foundation of the policy I reject here; it's what demands an implicit promise that a VN is "done" before we can have any clear displays of the play time. In other words, I propose these two changes to d2.2's section on length: 1. "finish all endings of the visual novel." -> "read through all routes of the visual novel." 2. "only trial/partial releases [etc]" -> "only trial releases. The exception is granted when the further development of the VN is abandoned, and the trial is all that has been released."

How can we know the "finished" length of a VN for certain, anyway, when a VN could easily receive a console adaptation which adds a route and increases the length even further? What is the problem with accepting that this piece of information can change? We already accept non-finalized working titles for VNs that are added to the DB, bad cover images that are replaced with higher quality ones later, and lately development status which inevitably must be changed from in development to finished after the release date.

The idea that the current length of an in-development VN is less knowable than a finished VN is incorrect. There is no objective way to measure the length of any VN unless you extract and analyze its scripts, or time a playthrough, but it isn't any harder or easier to do that if the VN is in development or finished. Whether testimony about the length of a VN is reliable or not at a given time can be judged with the same standards for both in-development and finished VNs. And true, any developers's claims about how long their VN will be in the future are extremely unreliable. But if we only care about the current length of the VN anyway, they don't matter.
#2 by cubky
2022-11-07 at 05:04
< report >As I wrote in the other thread - not only do I believe this should not be possible by policy, the field itself should be made impossible to change, since we have length votes now. If someone is interested in the length of an in-development entry, they might as well go click one more time to see the average.

What interests me for Ciconia no Naku Koro ni though - I see there are multiple length votes with reading speed. Once the entry moves to finished, will these be suddenly all counted, or is vndb smart enough to ignore them since they are tied to partial releases?
#3 by hansfranz77
2022-11-09 at 08:38
< report >The system we currently have is the best this is going to get. Show the playtime on early access / in dev vns optionally and when the thing gets finished open up voting to have folks vote on the final build again and show it on the page.

Every other way to do this will create a huge amount of work for mods and also spread wrong info around like nobodies business. So yeah 2.

#3 Looks like it's the latter. This Long Live the Princess recently came out of early access and i'm very sure there were some uncounted votes on it as it was still in dev. Now it has a single vote on it which seemed to have overwritten the partial votes that were on it before it went into the finished status.Last modified on 2022-11-09 at 08:44
#4 by beliar
2022-11-09 at 16:46
< report >
Looks like it's the latter.
No, it's the former. Uncounted votes remain uncounted after the Vn is finished. It would make no sense to count them, as they would still be based on a partial build, thus incorrect. Regarding the game you mentioned, the user must have simply recast his vote after the Vn was finished.
#5 by myou
2022-11-09 at 22:29
< report >
Every other way to do this will create a huge amount of work for mods and also spread wrong info around like nobodies business. So yeah 2.
Can you elaborate on why you think it would create a lot of work for mods or spread wrong info?

In the current situation, if anything, mods have more work: whenever someone adds a length interval like 10-30 hours for an in-development VN, a mod or another user has to intervene to remove that info. They have to specifically stop people from using the length field to share how long a VN is, which is what's normally intuitive to them.

It's up to the people to make edits and update info on this site. If this thread's proposal is accepted, we can finally allow people to estimate how long a VN currently is, so there's no need for mods to intervene. When someone plays a VN and finds that it's in the 10-30 hour range, they can add that info. Sure, maybe the length will increase steadily due to more updates. If later on, it becomes longer, another user can make another edit to say that it's now 30-50 hours. Either way, mods don't need to do anything.

About the idea of spreading wrong info: are you worried that people will assume that the length interval they see is incapable of changing, and thus their info will become "wrong" later? That shouldn't happen.

See how the development field is right above the length field? For reference, right now under the current policy, it looks like this for Ciconia:
Status: In development
Play time: Unknown (33 uncounted votes)

This is what it will be changed to look like, based on this thread's proposal:
Status: In development
Play time: Medium (10-30 hours, 33 uncounted votes)

Notice how "In development" is right next to the word "Medium"? Nobody will assume that the play time of a VN is completely unchanging and unfixable when it's so clear up-front that the VN is in development. Of course, its length may change later when it becomes longer. Just like how many other data points about a VN are changed or added over time. That's how VNDB normally works.

My first ever edit to VNDB was adding the kanji title of Umineko. Over the years, I watched as the length field of this game changed from short, to medium, to long, to very long. Because people wanted to know how long this VN currently was.

---

I'll reiterate the core of this proposal: I object to the idea that VN length should refer to some [future] length of the VN when development is [eventually] completed, which sustains the farce that length is "Unknown" (lol) for a 24-hour game like Ciconia. A VN's length should be defined as what the length of the VN actually is. Wikipedia also lists people's ages, despite the fact that their age will become "incorrect" a year later; we should let people share and see info that they care about.

The official guideline of never adding a length until a game is finished is out of touch. It was always out of touch with doujin creators who can't afford to develop a single finished release, and instead release partials that they continually build upon, for people to buy and play. In Japan, this typically happened every Comiket. But since the creation of Patreon, and later Japanese sub sites like Ci-en and Fanbox, many VN creators have moved to this model of partial releases. The current policy is biased against them, declaring the lengths of both a <2 hour VN and a >30 hour VN equally "unknown", and should be corrected.
#6 by hansfranz77
2022-11-10 at 10:10
< report >#4 I misspoke, the way you described it working was what I assumed it does work like. Though, that makes me wonder if the example I gave was reset somehow, which would not be an ideal case if that is true, yeah...

#3
Can you elaborate on why you think it would create a lot of work for mods or spread wrong info?
I try to do so the best I can.

In the current situation, if anything, mods have more work: whenever someone adds a length interval like 10-30 hours for an in-development VN, a mod or another user has to intervene to remove that info. They have to specifically stop people from using the length field to share how long a VN is, which is what's normally intuitive to them.

I mean, that is pretty much how the old system worked. You have estimates which have a huge span of time and these are put up by default and were based on script length of any given vn, not taking reading speed or skipped content into account.

Putting in a specific number that changes before the vn is complete gives the impression of that number still being correct, all the while the scope of the project could've ballooned with updates, doubling or maybe even more the length of the final game version. Let's call it 1.0 compared to version 0.5. In both cases, there needs to be a manual input done with removing the number. But estimates are easily recognizable as such, whereas a number like 6 hours for example needs to get checked manually after the fact. So in that case, doing a double check on the number is indeed creating one step on top of removing the estimate, like normally.


I do agree with you, it's not ideal in any case. If there was a way to fully automate or even tie these votes to build numbers, and have these show up on the page with their length votes. That might be ultimately the best way to solve this problem once and for all. But as far as I know and read here over the years, that doesn't seem to be possible. And of course there is the issue with the 6 hours example number being just left up on the game's page, because it does look like it could legit be true, so... if nobody votes on the final game after that point. It becomes misinformation if it is outdated and about half of the length of the real finished product by that point.

It's up to the people to make edits and update info on this site. If this thread's proposal is accepted, we can finally allow people to estimate how long a VN currently is, so there's no need for mods to intervene.

Which would mean we get the same edit wars we had in the past, with folks fighting over the estimate somebody else put up there and changing it back and forth...

When someone plays a VN and finds that it's in the 10-30 hour range, they can add that info. Sure, maybe the length will increase steadily due to more updates. If later on, it becomes longer, another user can make another edit to say that it's now 30-50 hours. Either way, mods don't need to do anything.


Think about it from the other side. There will be a huge amount of vns which have only a single maybe two votes and these will all be remaining at the wrong and outdated estimate number... if we are unlucky for years to come. Having these votes available but not visible like it's handled now seems to be the far better solution to prevent that scenario and edit wars from starting all over again, imo.

You can not just focus on popular stuff which will get updated very regularly, sure. But this would affect the entire database, and thus it needs to be considered for the rest of the vns in the db as well, which I'm pretty sure is the vast amount of data listed here.

I know these are a lot of assumptions, but I doubt any of these examples are completely out of the ordinary. The way I see it, there is currently no better solution to this issue than the one we do have, now. And getting back to a very similar system we had before, which seems to be what this proposal is in large part based upon, doesn't seem practical or like progress let alone the better solution to me.Last modified on 2022-11-10 at 10:49
#7 by myou
2022-11-11 at 01:51
< report >@#6
I mean, that is pretty much how the old system worked. You have estimates which have a huge span of time and these are put up by default and were based on script length of any given vn, not taking reading speed or skipped content into account.
It sounds like you're trying to condemn play time ranges as a whole by pointing out what the vote system does better. That isn't really on-topic, since play time ranges are an accepted feature of VNDB (see d2.2) which were broadly used for more than a decade, and are still used. But for the record:

Based on script length ... not taking reading speed or skipped content into account
That's not true. Script length isn't a sole determinant; play time ranges are estimated by looking at the ranges other VNs with a similar length, which thus accounts for reading speed. See d2.2: "To determine the length of a game, it's often better to ignore this time indication and instead compare it with other games you've played." Skipped content is also already factored into the guideline; d2.2 says it's the "time required to finish all endings" (which would just be changed to "read through all routes" based on this thread's proposal).

If you want to talk history: the "old system" (before play time vote averages were added) was suitable for all VNs. It was only replaced with vote averages for finished VNs because they're a better solution for finished VNs. And play time ranges are still the best solution for them if there are no votes. However, yorhel already recognizes that votes are not a suitable solution for in-development VNs, which is why they're uncounted. That's why we should rely on the best available and acceptable solution for showing the length for in-development VNs: play time ranges. That's what this thread's proposal is all about.

Which would mean we get the same edit wars we had in the past, with folks fighting over the estimate somebody else put up there and changing it back and forth...
Edit wars happen occasionally in every open database. Should we disable all descriptions just because G-Senjou was locked for 3 years due to an edit war there? Even when there were edit wars over play time in the past, yorhel never disabled the field outright. I can't see any support for your implication that play time ranges are unacceptable in every context.

Putting in a specific number that changes before the vn is complete gives the impression of that number still being correct, all the while the scope of the project could've ballooned with updates, doubling or maybe even more the length of the final game version. Let's call it 1.0 compared to version 0.5. In both cases, there needs to be a manual input done with removing the number. But estimates are easily recognizable as such, whereas a number like 6 hours for example needs to get checked manually after the fact. So in that case, doing a double check on the number is indeed creating one step on top of removing the estimate, like normally.
I'll repeat what I said in my last post. Nobody will assume that the play time of a VN is completely unchanging and unfixable when it's so clear up-front that the VN is in development. With all due respect, in your imagined scenario, the users of VNDB are so oblivious that if they're interested in a VN, they look at the VN's displayed length and cling to that data point and then flip out when they find out it's changed later, yet they somehow don't notice or care about possibly the most fundamental piece of info about the VN, that it is "in development" and thus will likely become longer in the future. In my view, it's far more likely they will be grateful that they were able to learn how long the VN is so far, at the current moment when they've expressed interested in it, and not panic if it becomes even longer later.
#8 by myou
2022-11-14 at 00:16
< report >A word about the poll at the top of this thread: it obviously wasn't designed to be scientific, and VNDB isn't a democracy where the vote counts in any concrete way, either... but all the same, there's a situation which I thought I should at least comment on, since people may assume the poll is a rough model of users' opinions.

After the first 48+ hours of the poll, there were 18 votes; the rate had slowed down to around 1 vote every 6 hours. But then somewhere inside a ~8 hour period, 10 people voted on a single option--with only 1 other vote in the same time period, for "no opinion". So it's likely that a certain amount of vote manipulation/brigading has affected the poll.

I think I've made all the points I wanted to make by now, and a full week has passed, so this is the last post I plan to make on the issue. Hopefully yorhel will take action on it at some point.

Reply

You must be logged in to reply to this thread.