NUMA's Ark

[乳头的早餐谷物]

Today, I was doing some planning for a top-secret project I'm working on (which totally isn't a new version of the #N IRC Quotes Repository) and got to thinking about how some of the things I was planning for that might work with NUMA. To be more specific, and yet tantalisingly vague, I realised that if successfully implemented, I had a system that would render sniping essentially irrelevant. Of course, that's a huge 'if', and for a site the size of NUMA I didn't imagine you could pull such a scheme off. But what if the site wasn't the size of NUMA? What if it was, shall we say, a bitesized NUMA?

My concept is indeed superficially similar to the old bitesizeNUMA, where a somewhat exclusive group of users choose maps from NUMA and submit them to a separate website. The similarities end there, however, because the underlying mechanic is entirely different. While the bitesizeNUMA concept works best with a like-minded group of users, my idea demands diversity.

Essentially, members would choose maps from NUMA and submit them to the new site, where other members will rate the map. Based on these ratings (and possibly other factors), the system will determine which members have similar taste in maps. Using this information, the ratings that users see will be tailored and weighted according to their tastes.

While poorly considered votes (including sniping) are not much of an issue with this system, since they only hurt the perpetrator who will get less accurate suggestions, it does require a fairly large number of votes per map so that the similarity-calculating system has enough data to go by. A large and fragmented website like NUMA seems a poor environment, hence doing it on a separate site.

It is highly likely I will have a go at making this. Questions or comments?

Edit: Idea now has a name by Skyline and is close to completed.

[Nexx]

This is neat, and should be a part of NUMA, or, since that's not going to happen, it should be as NUMA-oriented as possible.

First, from your description, it sounds like you're going to be using favorites in this system, but then you also mention sniping, which has to do with rates, so I'm just going to ask straight out: which is your measure for like/dislike? Because you could have users just submit maps they've faved, or you could have users submit maps and their rating of that map. The former will result in namely good maps being submitted to the site, while the latter will result in all kinds of maps being submitted to the site, as well as a more nuanced gauging system. You don't have to use the same rating system as NUMA - you could use, say, thumbs up/thumbs down/not specified, like Pandora does.

Second, the purpose of this site is to offer recommendations, no? So then the site shouldn't be serving as a repository of maps at all. Users shouldn't be submitting maps themselves, they should just be submitting references to the maps on NUMA. The site doesn't care. It doesn't look at the mapdata. It just says, "oh here's a thing you liked, so maybe you'll like this other thing here." No need to be divisive at all. (Whereas, for example, AltArc couldn't avoid divisiveness.)

[rocket_thumped]

I've championed a weighted voting system similar to this in the past. I hope it works out maestro.

[SlappyMcGee]

This is the most brilliant thing since Cotton, asshole. Approved.

[MattKestrel]

Fantastic idea. I'm all for it.

[Tunco]

I like this idea. I support it.

[Lenny]

It sounds amazing, but I think it would be an immensely complicated database, and generation of ratings for newer maps could be really complicated and possibly inaccurate.
Still, I'd be all for the idea of seeing it completed.

[Luminaflare]

This is the most brilliant thing since Cotton, asshole. Approved.

/me mentally moves comma one word to the left.

[乳头的早餐谷物]

I think my only worry with this system is that the "reviewers/choosers" that you pick will be a bunch of elite mappers who will only choose maps by their own secluded group of mappers. If this doesn't happen I will endorse the idea, bitesized NUMA was always a good idea (as was Top rated).

The more the merrier. To an extent. I'll vet applications to the site (initially, at least) but it's certainly not going to be elitist or highly exclusive.

First, from your description, it sounds like you're going to be using favorites in this system, but then you also mention sniping, which has to do with rates, so I'm just going to ask straight out: which is your measure for like/dislike? Because you could have users just submit maps they've faved, or you could have users submit maps and their rating of that map. The former will result in namely good maps being submitted to the site, while the latter will result in all kinds of maps being submitted to the site, as well as a more nuanced gauging system. You don't have to use the same rating system as NUMA - you could use, say, thumbs up/thumbs down/not specified, like Pandora does.

While I'd expect users to submit their favourite maps, they'll also submit a rating of the map, which is what will be used for the caluclatings. I figured I'd make rating system the same as NUMA's.

Second, the purpose of this site is to offer recommendations, no? So then the site shouldn't be serving as a repository of maps at all. Users shouldn't be submitting maps themselves, they should just be submitting references to the maps on NUMA. The site doesn't care. It doesn't look at the mapdata. It just says, "oh here's a thing you liked, so maybe you'll like this other thing here." No need to be divisive at all. (Whereas, for example, AltArc couldn't avoid divisiveness.)

Mmmhm, precisely. The submit page will just be a box where you input the map ID.

It sounds amazing, but I think it would be an immensely complicated database, and generation of ratings for newer maps could be really complicated and possibly inaccurate.

Eh, not really. The database does get pretty big when you're generating these similarities, but it's not at all complicated (the code is already running on the Quotes Repository with 1600 entries and 100 notable users).

[heatwave]

Wow, excellent idea. Would it be possible to host it as a nmaps.net subdomain? e.g. bitesized.nmaps.net?

[乳头的早餐谷物]

Wow, excellent idea. Would it be possible to host it as a nmaps.net subdomain? e.g. bitesized.nmaps.net?

Doubtful. At the moment, this is an ortsz thing and not an official NUMA thing.

[unoriginal name]

Sign me up, Oeneral Grtszbeckistan.

[otters~1]

Hot idea. Nice to see someone doing something.

[ENT474]

nice idea. go forth

[Sunset]

Sunset approves.

[Vyacheslav]

Would any of it be anonymous?

[otters]

What about an algorithm that finds maps with similar styles and adds them to your suggestion-er!

[Yoshimo]

Can you sign me up as a map-chooser when it's done? Or at least consider it? I want to get more involved in the community, and, I'm not really an elitist, so you don't have to worry about that.

[Rose]

I know this thread is a couple months old but is this idea still going anywhere?

[t̷s͢uk̕a͡t͜ư]

What about an algorithm that finds maps with similar styles and adds them to your suggestion-er!

^^ Reminds me of an article I read recently in TIME - http://www.time.com/time/magazine/artic ... 03,00.html ^^

If this was implemented you would be my hero.

Lol, I get paid to do this kind of thing for a living.
But, y'know, with text usually.

[SlappyMcGee]

I know this thread is a couple months old but is this idea still going anywhere?

It is, but like, top secret and shit.

[乳头的早餐谷物]

I was just reminded of this thing. It would still be pretty cool, though the current situation of a mostly-broken NUMA is problematic to say the least. I might as well offer a link to the beta version of two years ago: ARK. The security code is righteous. I can't really remember to what extent the site actually works, but you're welcome to try it.