WAD tagging system?

This is just an idea at this point, I don't know how it would be organised yet but I wanted to get some opinions. Would there be interest in a WAD tagging system? I mean like in movie databases they have tags like "mother son relationship", "car chase", "surprise ending", etc. Maybe we could have something similar but for doom WADs. Then you could use it to search the archive for all castle maps or something.

Sounds all right in theory, but in practice it'll just be one more thing for people to spam/troll/abuse like the rating system already is.

Well, most of the tags would be completely neutral ("castle", "hell", "sandbox", "circular architecture", "mandatory secrets", "low ammo") so there would be less incentive to abuse the system I believe. It doesn't have to be free to edit for everyone anyway, a team of 10 experienced players could already do a lot of work by going through their WAD folder and making quick notes.

Memfis said:

a team of 10 experienced players could already do a lot of work by going through their WAD folder and making quick notes.

Again, in practice, it's problematic to find just one person to write a /newstuff review of a given wad, not to say a team who would check every new wad in depth.

Tags are not a bad idea, but it would be troublesome to:

• make them, and
  
• verify them for each individual wad.

Steam tags.

That's reason enough to really restrict this idea. Just a bunch of BS that has to get filtered constantly. That would have to be strictly registered users at best.

I think this would be great (I wanted to do the same thing by myself). I figured the easiest way might be to include the tags at the bottom of a wad's text file like #castle #hell #sandbox, etc. That way people could use the /idgames search whole textfile tool to search for tags (and only certain people could upload the new txt files with tags).

Should we start by making a list of tags? I'll definitely volunteer to tag some wads.

Heh, tags from ToD is definitely something I'll be looking after.

It's a pretty good idea on paper, but it'd probably need strict oversight. Not just trolls mistagging wads, but terms like #sandbox can actually be hard to define clearly and soon you'll have even trusted editors disagreeing about borderline cases. Just pitching a wild idea, but maybe have the tags colour-coded depending on how many editors agree with it?

Steam tags.

Yep. The system works.

That's reason enough to really restrict this idea.

...ah.

Maybe I'm just weird, but I don't flip out when a game I like is tagged "walking simulator", or when a Barbie title pops up in the "survival horror" category. I chuckle and move on.

Why people always strive to be offended and to enforce censorship in every system is beyond me. Even silly tags come with information, a reflection of public perception around particular games, a potential indicator of controversy, etc..

For those who don't know, Windows 8 allows you to add tags and ratings (currently only to image files) so you can search/sort by them. If Microsoft has tags, can the rest of the world be far behind? :)

Seriously, this is a good idea. I've done searches for 'castle' and 'Episode 4'. Being able to search for 'new textures' would make a lot of people happy. Ty has already started doing something along these lines by identifying "trapwads" and "terrywads" with additional text files. Tags is a logical next step.

I agree with Memfis that most of the tags should be neutral and only be used to describe things inside the wad, and not be used for more subjective things like "quality" (that's what the ratings/comments are mostly used for).

dew, that's true that even the trusted taggers would disagree about tags. I think a tagger's name should always be included with their tags, so maybe people would get used to their tagging. Colour-coding the tags would be cool, but stuff like that would have to be coded in by Bloodshedder (if he even likes the tagging idea).

Demonologist, almost all my tagging would be done on doom2 wads from 1994-2001 because I've played virtually all of the vanilla/boom ones :) going through those wads again should jog my memory about them and quickly tag them

I like last.fm's approach with tagging music genres, where a band can be tagged any number of different genres by the listeners, and the genre is described by the list of tags in the order of most frequent. That really helped me identify what particular subgenres of punk music I was into long ago.

However some of the REALLY underground bands with very few listeners often got silly tags like "crybaby-core" and "bluegrass" when they were obviously not.

One of the differences between wads and music that might effect this system negatively, is that wads are getting increasingly disposable, where one may listen to a particular band for months, and come back to it frequently, where in Doom, a map may get played once and then never again. So I can expect to see only really popular and replayable mapsets like your classic megawads getting accurately tagged, while many other single levels, regardless of quality, will be tagged as "1994" "terry" or "shit" or what have you.

I think this could have worked well had this system been implemented decades ago, but to go back and retroactively tag wads youve played before might negatively effect the results that turn up when searching them at this point.

I really approve of the idea of a tagging system for doom maps/mapsets. It would make it easier to find things to satisfy niche tastes like CrateRoom or Grindy-Puzzles for people like myself. Otherwise, my next best approach to maximizing my chances finding the better doom maps for me to play is to look at other people's favorites, or browse a library like idgames and sort by rating (most of which are just texture packs and stuff). This way only takes the least common denominator of map preferences though.


I think any sort of tagging system would be an improvement to the search of any mapset library. It's just a matter of what type of tagging system to implement. I think I agree with 40oz that the last.fm system is pretty good for music, but also think it works for doom maps as well. It involves the least micromanaging (micromanaging attempts usually are implemented, in practice, by one-size-fits-all solutions and that is against the mindset of tagging systems, which is to let niche (edit: and unaccounted for) map styles have a way to be identified and searched for).

Genre semantics are dynamically defined by the average/general usage of each genre-word that the culture uses; people's collective individual taggings should give an efficient sense of the genres or styles of that map. You will get some mapsets tagged with misspellings and redundant tags (like a castle map tagged with #catsle, #castlemap as well) but these don't really surmount to anything bad overall. You will also get a few #sucks and #badmapdonotplay tags on, lets say a specific hypothetical #crateroom map, but overall the system works and people searching for a #crateroom map will still have an increased chance of finding that particular crateroom map relative to having no tag system at all.

40oz said:

I think this could have worked well had this system been implemented decades ago, but to go back and retroactively tag wads youve played before might negatively effect the results that turn up when searching them at this point.

I don't see it. Using your music analogy, it's like saying I can't tag Pink Floyd retroactively, because my opinion isn't spontaneous enough anymore.

i approve. it should be optional by the author. is the map a boss map? slaughter map? puzzle map?


it makes looking for certain kinds of maps that much easier. jokewad?

i don't think it would be troublesome to make them; it's just an array of strings that the uploader can type. search is done by looking for string. what would be nice is to provide a recommended standard like slaughter, boss, puzzle etc.


looking for inappropriate tags is no more difficult than looking for terry wads.

imo, stuff like #sucks #donotplay etc. shouldn't be used at all, that's for the ratings and comment box. Only objective facts (pretty much) about the wad should be used for the tags. And there should be a definitive list of tags that people pick from, to make searching much easier.

At least that's how I was planning on doing it before Memfis started this thread.

imo, something like this needs to happen eventually. The archive is huge and old. When people are looking for something, they have to rely on other people to suggest something, and there are so many wads out there that are never suggested.

This is something that has to be done carefully and not by random people.

Some ideas to start with (suggest your own):

environment: hell, city, nature, base, caves, mines, mountains, ocean, castle, marble, village, maze, spaceship, space base, church, warehouse, apartments, prison, ruins, desert, void, snow\ice, sewers.

elements (use when they are a significant enough part of the map): real world objects, pyramid, 3d bridges, tortured bodies, stonehenge, lava, nukage, water, sewage, hidden entrance, parallel dimensions, central hub, arena, multiple floor lift, pentagram, barrels, obligatory secrets, multiple cyberdemons, numerous arch-viles.

characteristics: nonlinear, very short, short, long, areas connected by teleports, circular architecture, rectangular architecture, low ammo, sandbox, completely linear, striking height variation, high lighting contrast, overall dark, color themed areas.

Memfis said:

Some ideas to start with

How would they apply to multi level wads? Most of them hardly.

I think multi level wads need separate tags for each level obviously.

dew said:

I don't see it. Using your music analogy, it's like saying I can't tag Pink Floyd retroactively, because my opinion isn't spontaneous enough anymore.

By the time Pink Floyd came out, music production quality was already very good. A better comparison of the range of mapping between 1994 to current would be like the opinions of young kids who listen to pop dance music, post-hardcore, dubstep (whatever kids listen to these days, I don't know) being introduced to the music of the 1920's, which is really grainy and fuzzy and unlistenable to their standards. The appreciation for that kinda stuff isn't universal in this decade because most of us weren't around at the time it was growing into something remarkable.

40oz, that's why the tags should only be for objective facts about the map and not whether someone likes it or not. Maps from any year can have #mountains and #techbase, etc.

Memfis, nice list so far. I have some to add later.

This is random analysis barf I'm just throwing out

I guess to continue with our comparison to music related tag systems, we can look at Pandora Radio, which uses the alternate approach to tagging rather than dynamic-anarchy user-influenced. Pandora does things in the style that TimeOfDeath favors; it has a professional panel of music analysts that come up with a finite list of tags (they call them elements) that are applied to each song. It does a good job. The only reason I would advocate dynamic-anarchy style tagging is because it helps to link the micro-niches of map genres together. It makes it so elements of style that at least one person cares about don't fall through the cracks and fail to be distinguished. If someone or a group of people figure out some newly considered map characteristic that is suddenly meaningful to them that wasn't considered before, user-tagging-anarchy may have a relative benefit in this regard.

On the other hand, it means there are infinite variety of tags, and if you want to do something related to tag search service, you would have to come up with a heuristic to find things and sort things like it. You will always have, for example, a non-city map labeled as a city map by at least one person in the large pool of tags that it might accumulate.

In regards to music and the kids these days and their awful rap music and dubstep, I will argue that if granularity of a song effects someone's satisfaction from it, then grainy-sounding or old-feeling is a valuable characteristic that needs a tag (or the reverse, make tags for non-grainy or non-old-sounding). These things are characteristics of the song itself, and not the intrinsically related to the release date. If someone finds it valuable to search for maps or separate by old-ness feel, then tagging for oldness-feel serves it's purpose. It is an "objective characteristics", in other words. I agree that sucks is not one of these characteristics too though.

Now, when weighing anarchy tagging to preset professional-based tagging, I guess it really comes down to those niche, "underground", maps or songs, whichever we are talking to. Even though popularly tagged mapsets will converge to a fitting genre as number-of-tags increases, these less popular maps are more susceptible to being labeled wrongly in user-based-tagging (I'm not sure if last.fm lets you write your own custom tags, or if you pick from a preset selection I guess that would change my thoughts on it). On the other hand, if these maps are niche because they are weird or different, then user-defined tags might work better on it. On the third hand, it might be unpopular (unpopular refers to frequency played, not necessarily it's perceived quality) because it sucks and doesn't get spread around, thus deserving an objective breakdown of it's style to shield its tags from random bashing. On the fourth and final hand, maybe #sucks tags do have objective merit - they identify if a map, at one point, had any surge of hatred or animousity thrown at it for some crazy reason. If you are a curious Doom Historian, maybe the #sucks has value to you and has benefit as a possible tag option users can write-in arbitrarily. Also, people searching for niche map styles that know they like things that other's generally don't like will probably ignore the #sucks troll-tags that get thrown around.

When comparing Pandora Radio to Last.fm, they both work, they both take a small degree of effort to implement. The same would probably apply to Doom maps too. I think any sort of tagging is awesome, that's all I have to say. I got way too cheesy with this analysis and don't actually have a real opinion at this point.

Memfis said:

I think multi level wads need separate tags for each level obviously.

In that case tod's idea of just adding #tags to map descriptions wouldn't work.

Well, all you have to do is MAP01: #tag #tag #tag, MAP02: #tag #tag #tag :)

If the preferred system is to make addendums to existing text files, that could be really useful for someone like me who has the /idgames database saved on an external hard drive (as opposed to a thing exclusive to the website). Even windows XP can search for exact words or phrases in text documents. I'd just need to update my files frequently.

some tags I can think of:

outdoors, cave, bbs, (wads made specifically to advertise a bbs server) myhouse, factory, sewers, 1994, classic, details, crates, gothic, Egypt, Aztec, Plutonia, wolf3d, quake, hexen, heretic, duke3d, blood, rott, strife, torches, vines, flesh, brown, ashwall, startan, rock, fireblu, red, gray, tan, green, blue, gstone, metal2, marble, wood, tekwall, bricks, community project, single author, collab, team, puzzles, traps, tricky, tricks (vanilla mapping tricks), hard, easy, innovation (for projects like 10 sectors, 100lines, monochrome map project) 1024, speedmap, ctf, duel, ffa, old school dm, altdeath, zdoom, gzdoom, eternity, edge, vanilla, limitp removing, boom, resource, experimental, example, test, prefab, acs, decorate, dehacked, exe, crowd control, suspenseful, joke, preview, demo, beta, unfinished, abandoned, speedrun (designed for speed running), tyson, bfg9000, megawad, short episode, weapon mod, new sounds, E1, E2, E3, E4,

all pretty neutral I think, but I can imagine tags like 1994, unfinished, or joke being used in a derogatory sense. Maybe more positive blanket terms like 90s, and funny.

i think a tag system / database would be pretty invaluable for people looking for more stuff to play. kind of reminds me of Geniac's Doom Wad Played List but more functional. idk of anything to suggest, since a lot of the terms i use like "adventure" and "exploratory" are rooted more in the way i feel about the stuff i play, whereas a tag like Plutonia is usually straightforward.

tagging by author would be another p good parameter

Phml said:

Maybe I'm just weird, but I don't flip out when a game I like is tagged "walking simulator", or when a Barbie title pops up in the "survival horror" category. I chuckle and move on.

Why people always strive to be offended and to enforce censorship in every system is beyond me. Even silly tags come with information, a reflection of public perception around particular games, a potential indicator of controversy, etc..

There's also plenty of tags such as "Would you kindly?" which tend to spoil the game for new comers and beat-off-meme-tags like "The cake is a lie" that really added nothing to the conversation. Looking through it now, it has gotten better, but before there was any sort of lookout for that sort of thing, it sucked.

In other words, Steam went to paragraph two and actually fixed some stuff. Go figure.