"BadWiki", or rather "BadLearnDB"

[Laraso]

EDIT: I meant to put this in CYC. If a mod feels this is an inappropriate subforum, please move it there.
_________________________________________________________________

This is something not many people want to talk about, but this is something that I feel is very important and needs to be discussed. Lately I have been improving my skills and am now able to achieve wins fairly often, and I find myself needing to look up more information than I ever have before as I come across things in the game that I have not previously experienced. I have four options for obtaining this information: looking it up on the wiki, looking it up in the learndb, asking in IRC, or asking here in the Tavern. The IRC and Tavern should not be relied upon as a database of information as it relies on having people online who know what you're looking for, having those people read your question, and having them bother to answer it. A fixed database of information that can be accessed at any time and always gives the same answers is ideal and is what most players need. So really that only leaves the learndb and the wiki. People often make fun of the wiki, calling it names such as "BadWiki", and much of the learndb consists of attempts at witty humor designed to scrutinize the wiki:

badwiki[1/1]: An old wiki (with lots of out of date and incorrect info; enter at your own risk): http://crawl.chaosforge.org/

famous last words[1/28]: according to the wiki

In fact, a majority of the learndb is comprised of unhelpful attempts at witty humor, and a disturbingly large portion of it is simply direct quotes from messages sent in the IRC that often have no actual relevance to the topic at hand. For example, take the entries on Okawaru:

okawaru[1/5]: Heroism (*) temporarily gives +5 to all non-magic skills except Invocations and Evocations, although Fighting skill from it gives no extra HP. Finesse (*****) halves attack delay but does not stack with haste or berserk and is blocked by stasis. Okawaru also gifts weapons and armour, and ammo if you have a ranged skill >= 8.
okawaru[2/5]: <G-Flex> I picture oka as this old guy going through his attic and just discarding old junk as gifts
okawaru[3/5]: < valrus> I AM OKAWARU, GOD OF COMBAT AND FIVE DOLLARS FOR THE OSCILLATING FAN IS MY FINAL OFFER
okawaru[4/5]: You finish putting on the -5 robe of Misfortune {god gift, -Cast Contam+ *Tele EV-4 Str-2 Dex-2 Int-2 Stlth-- Curse}.
okawaru[5/5]: Gives almost no piety for easy fights, none at all for trivial ones.

Okawaru is a major god and as you can see, only two of those entries actually have any relevant info pertaining to him and the info that is provided is very barebones-basic. The rest is just completely unhelpful, useless dribble to anyone who is actually looking for pertinent information about the god.

What does the learndb have to say about Zin?

zin[1/1]: Zin gives mutation resistance (growing to immunity at max piety), {recite} (enchantment-type effects on enemies), {vitalisation} (protection from debuffs), {imprison} (tomb other), {sanctuary} (creates a temporary attack-free zone), and passive lifesaving. At ****** you can get all your mutations cured (only once).

Yup, that's all of it. Wow, two whole sentences! That sure is a lot of information! The wiki could never beat that...

... oh wait. It actually does, and it completely dwarfs the learndb's information by comparison. There really is no contest here, the wiki gives much more in-depth information about Zin than the learndb ever could and personally if anyone was actually looking for information about Zin I would direct them to the wiki, not to the learndb. Do note that Zin is not an exception, and that a majority of entries on the learndb share this lack of information.

tedric put it best: --

Also, if you want consistently reliable information re: anything Crawl, use the learndb, not the wiki: https://loom.shalott.org/learndb.html

I do use learndb (via knowledge bots) but often the information there is too simplified to answer my questions fully. In many situations the bots are only good for half an answer, or for reminding myself of something I'm already familiar with. The penance entry, for instance, doesn't give the chart for correlating the message on the ^ screen to the numerical range of penance you have to work off -- which was the only piece of information that gave me a solid sense of how mad my god was. The wiki has that sort of stuff. So I check both, and defer to learndb whenever there's a discrepancy.

-- ... this is a problem.

While many people criticize the wiki for having bad/inaccurate information, the learndb can be characterized by it's lack of information and the information that it does have is very hard to find if you don't know the exact term you need to be looking up. When searching the learndb, I routinely come up short on the information I'm looking for and I have to resort to asking in the IRC or Tavern anyways. A wiki is a much more streamlined and user-friendly way of providing a database of information about the game and is much easier to search through. A wiki is capable of doing things the learndb literally cannot do even if we wanted it to such as providing visual images and formatting information into easily-readable formats (tables, bold text, italics, bullet lists, etc. This is a prime example: http://crawl.chaosforge.org/List_of_spells_by_school -- the learndb will never be able to replicate that kind of functionality). Yet it seems that the people who are most knowledgeable about the game and also the most capable of 'fixing' the wiki are the same ones who protest against the wiki and advise people not to use it.

This is a also a problem.

So how do we fix this problem? The most obvious answer is to stop pretending like having a wiki is a bad idea and put effort into providing it with accurate and up-to-date information. I've suggested this multiple times in IRC over the years since I've started playing this game in 2009, and every single time I've typically been met with these same excuses:

"We don't need a wiki, we already have the learndb"
"Updating the wiki is too much work, we already have to worry about managing the learndb"
"The wiki is so bad that I don't think anyone actually wants to bother to fix it. It's broken beyond repair~~"

That first reason is untrue, and the second and third reasons equate to pure laziness. However, I have recently been getting a new and alarmingly legitimate reason for not fixing the wiki:

"Editing permissions and account creation have been locked down and we couldn't fix it even if we wanted to."

... This is a serious problem. I tried to log into my account that I registered on the wiki many years ago, and sure enough -- my account has been deleted.

It is for these reasons that I would like to propose a radical idea that is (sadly) most likely going to be completely ignored by the people who I want and need to take it seriously. We need to either contact the administrator of the wiki and ask to restore the previous editing permissions and get to work on fixing it up to provide the accurate information that everyone desires, or create a new wiki and build it with accurate information from the ground-up.

Please consider this.

[and into]

Well the size, interlinked functionality (which thus necessarily produces multiple sites of similar info being presented), and the many details contained on the wiki all actually contribute to making it very hard to keep up to date. So yes it would be nice if the wiki were better but improving it is a non-trivial task. I have an account and have done some small things to improve it on a smattering of pages, but a really thorough revision would be quite an undertaking.

The learndb has less detail sometimes but on the other hand the chance of something actively misleading someone and bringing them to harm is much, much lower with the info on the learndb. So I basically think of the learndb as taking more of a "first do no harm" approach which is not at all an unreasonable approach to take.

Also there's many problems that the wiki has beyond just being out of date or inaccurate, but I'll just limit my comments to the question of the accuracy of factual information on the wiki.

[duvessa]

I never liked telling people the learndb is good (because yes it sucks). That said, I think that maintaining a wiki for a game as actively developed and obscure as Crawl is unrealistic. The DoomRL and ToME4 wikis are similarly useless, and indeed I can't think of a single example of a good wiki or wiki-like entity for a game meeting those two criteria.

[wolpertinger]

What does the learndb have to say about Zin?

zin[1/1]: Zin gives mutation resistance (growing to immunity at max piety), {recite} (enchantment-type effects on enemies), {vitalisation} (protection from debuffs), {imprison} (tomb other), {sanctuary} (creates a temporary attack-free zone), and passive lifesaving. At ****** you can get all your mutations cured (only once).

Yup, that's all of it. Wow, two whole sentences! That sure is a lot of information! The wiki could never beat that...

Actually that's a pretty good example of a learnDB - You gotta remember that the brackets around {recite} for example is the equivalent of a link - it means if you ??recite, you'll get more information on recite. Let's take a look at ??recite.

recite[1/4]: Causes various effects to chaotic, unclean, evil or heretical creatures in view. Four-turn action, with strength depending on piety and invocations. Effects scale from things like confusion and fear on humanoids, to smiting and instakilling chaotic creatures, to permanently blinding or paralysing heretic priests.
recite[2/4]: Abominations (chaotic, differently insane): damage, Ablutions (impure): damage, Apostates (intelligent): fear/confusion, god worshippers: blind/anti-caster, Anathema (unholy): fear/confusion/damage
recite[3/4]: You can do anything during the recitation that doesn't require your voice. Things that stop recite: spit poison, breathe * abilities, zin's abilities (including donate gold), t command, eating, drinking (potion or fountain), reading a scroll, casting a spell, or being silenced, paralysed, confused, slept, petrified (completely) or berserked.

That's actually pretty good. It's even possible to link charts. It's not to say that a wiki is a bad idea, but the learnDB is capable of being plenty informative, especially if you aren't overly interested in large mathematical formulas on exactly how something works, and just want the jist of something.

[rchandra]

One of the good features of the learndb is the public nature of it. It can't get false information without being seen, and if somebody reads false information there's a decent chance of somebody correcting it. Jokes generally signify the end of useful information, so Okawaru[5] should be moved up, and I will do so now.

Incidentally the Zin quote above tells you what to look at if you want more details - the ability names. Most or all of the information section of the wiki article is there, and much of the wiki article's strategy section is useless and misleading blathering.

[Laraso]

Actually that's a pretty good example of a learnDB - You gotta remember that the brackets around {recite} for example is the equivalent of a link - it means if you ??recite, you'll get more information on recite. That's actually pretty good. It's even possible to link charts.

Having to go through entry after entry and being redirected to other entries which redirect you to other entries until you finally get to what you're looking for is not ideal, and is very far from being user-friendly. It is not similar to links at all, because you don't have to click any links to get to that information on the wiki, it can all be found on the Zin page. When I ??zin, I want information about Zin, I don't want to be told "Oh yeah, Zin lets you recite. That's all we're going to tell you here, If you want to actually know what recite is and/or does, you're looking in the wrong place!".

If I wanted to look up everything about Zin, I'd have to jump through these hoops:

  Code:

??zin
??recite
??recite[2
??recite[3
??recite[4
??vitalization
??imprison
??sanctuary


...and even then, I'm probably not going to get all of the information I'm looking for. Those entries all have very basic definitions, and are still far less helpful and informative than the wiki is.

For example:

imprison[1/1]: Zin ability. Puts a tomb of silver walls around a targeted creature. Duration is increased with invocations skill.

Imprison - Temporarily creates a silver wall prison, similar to a Tomb card, around an enemy and push away adjacent monsters. After the imprisonment is over the monster gets Recited. Imprison can only be used on monsters which are susceptible to Recite, and strong monsters may break themselves out. Costs 5 MP, 125-250 Food, and 4-6 Piety.

And as I said earlier, you need to know the exact term you need to look up or else you're not going to have an easy time finding what you want to know in the learndb. With a wiki, all relevant information is easy to find and is presented to you all at once instead of in very short individually-regurgitated messages. Information that you may not have known but are interested in knowing is readily displayed in relevant topics on the wiki, however in the learndb you have to know what you don't know to find those things out.

the learnDB is capable of being plenty informative, especially if you aren't overly interested in large mathematical formulas on exactly how something works, and just want the jist of something.

Actually, I tend to find it's the other way around. The learndb often has large math formulas without any easy-to-understand explanations, however the wiki attempts to simplify these by providing superior information formatting and including examples which try to make it easier to understand. For example:

launcher damage[1/4]: Average damage = (B/2 + L/8 + A/16 + min((str - 10)*(2B+A)*C, (L+1)/4))*(1+skill/D)*(brand multiplier) + slaying/4 + A/2 + (L/2 if using xbow). Here A is min(3, skill/3), L is the launcher enchantment, and B, C, and D are constants depending on what type of launcher you are using (see [2]).

Compared to the wiki's version: http://crawl.chaosforge.org/Ranged_combat#Damage

The wiki is far easier to understand (not sure if it's accurate though).

One of the good features of the learndb is the public nature of it. It can't get false information without being seen, and if somebody reads false information there's a decent chance of somebody correcting it.

That's kind of like the entire concept behind wikis. It's editable by the public, and anyone who sees false information can correct it. If you try to add false information to the wiki, the information is attributed to you and logged in a history, and the history also allows you to compare and contrast previous versions of articles and restore them if necessary.

Most or all of the information section of the wiki article is there

That's completely untrue, the learndb only has the bare-essentials. In addition to many other things, the ??zin entry doesn't even tell you what kind of conduct the god has, while the wiki goes in-depth into explaining what Zin does and does not like. Someone new to the game looking up information about Zin in the learndb wouldn't find anything explaining that Zin strongly dislikes it when you polymorph monsters, and while some may claim that the inaccurate information on the wiki can mislead players and create harm, the lack of information on the learndb is capable of doing the same thing. In fact, the learndb doesn't even bother to mention how Zin uses gold, which is a pretty core mechanic to Zin and is certainly critical information. The learndb doesn't even come close to the wiki, even with it's numerous sporadically interspersed entries.

==========

the wiki article's strategy section is useless and misleading blathering.

*AHEM*...

okawaru[3/5]: < valrus> I AM OKAWARU, GOD OF COMBAT AND FIVE DOLLARS FOR THE OSCILLATING FAN IS MY FINAL OFFER

ashenzari[4/5]: God of bondage.

elyvilon[4/4]: <crate> no you fucking moccasin the stairs are right there dont walk away from them <crate> no why <crate> arghh

kobold[2/2]: While the true nature of the Kobold remains a mystery, some ancient texts have described it as "a sort of brown moth".

great sword[2/2]: the sword world continues to overlook in my opinion the efficiency and power and i even say majesty of this great sword

wiglaf[3/4]: sombrero priest

wiglaf[4/4]: < murphy_slaw> he's the rootenest tootenest priest of okawaru weeeeest of the Pecos

sigmund[2/4]: sigmund haters: whoever buffed grinder

snake pit[3/3]: sorear longs for the good old days when all three stairs went into the vault, virtually every time.

the shining one[7/7]: <DracoOmega> TSO also disapproves of lying on your tax returns

death yak[4/4]: <mikee_> remember death yaks? those death yaks are like death yaks but in the lair

summoning[3/3]: <Poncheis> but isnt summoning an offensive school <elliptic> most players are offended by summoning yes

giant spiked club[3/3]: a - the +3,+5 giant spiked club "Agagh" (weapon) {holy, Dex+4 Int+2} <wheals> "agagh" sounds like a noise someone makes when they get hit with a giant spiked club

lernaean hydra[3/3]: <Sizzell> araganzar (L18 DrMo) killed the pulsating lump shaped Lernaean hydra. (Swamp:5)

blue death[1/1]: Churchmouse the Blademaster (L27 HaWn), worshipper of Okawaru, blasted by a smoke demon (divine providence) (summoned by a deep elf sorcerer (summoned by a Blue Death (summoned by a deep elf demonologist
(summoned by a Blue Death (summoned by a deep elf demonologist))))) in Zig:21 on 2011-08-02, with 649571 points after 98719 turns and 13:08:27.

throwing net[3/4]: <minmay> "the giant spore struggles against the net"

teleport[1/1]: The best strategy.

redback[2/2]: http://upload.wikimedia.org/wikipedia/c ... ck_8MP.jpg

stealth[6/6]: http://i.imgur.com/PMlmu1H.gif

troll[3/3]: Sometimes inhabits ##crawl and advises people on the most efficient way to splat.

zot ~ THE WIZARD'S CASTLE[1/5]: MANY CYCLES AGO, IN THE KINGDOM OF N'DIC, THE GNOMIC WIZARD ZOT FORGED HIS GREAT *ORB OF POWER*. HE SOON VANISHED, LEAVING BEHIND HIS VAST SUBTERRANEAN CASTLE FILLED WITH ESURIENT MONSTERS, FABULOUS TREASURES, AND THE INCREDIBLE *ORB OF ZOT*.

...

... and it goes on.

[crate]

if you find a learndb entry that is actually unhelpful you are free to delete it

outside of entries that are for personal use (no one is going to randomly query "craterc") I would actually prefer if the learndb were kept to actual gameplay information

edit: the real problem is actually this sentence:

I find myself needing to look up more information than I ever have before as I come across things in the game that I have not previously experienced.

Either 1) you do not actually need that information
or
2) the game itself should present it to you

[Laraso]

(I accidentally thanked when trying to click reply)

edit: the real problem is actually this sentence:

I find myself needing to look up more information than I ever have before as I come across things in the game that I have not previously experienced.

Either 1) you do not actually need that information
or
2) the game itself should present it to you

I'm sorry, let me rephrase that.

"I find myself wanting to look up more information than I ever have before"

[tedric]

I would love to have a source of info that is more reliable and more complete than any of the options available now, though I suspect it's an eternal pipe dream.

But I don't see what's wrong with including humorous stuff in learndb. It spices up an otherwise dry and formula-filled resource -- and a few of those "jokes" are actually quite illuminating. It's just as efficient to make a sarcastic comment about crappy Oka gifts as to explain that you shouldn't count on getting good items, and the joke makes for better reading because it actually captures the frustration you should be prepared for as an Okawarite.

Also I have run across this entry several times before:

stealth[6/6]: http://i.imgur.com/PMlmu1H.gif

and it never fails to crack me up. There's no harm done, unless the humor is overshadowing/overwhelming/replacing/being mistaken for real info.

So how does one edit learndb, anyway?

[Tedronai]

I think learndb and wiki have different target audiences. People who use the learndb are usually familiar with the mechanics of the game and just want a quick reminder of some very particular case. Asking those people to work on including more detailed explanations will yield no result.

If you want a better source of information you have but one option: Look for for people who have a similar interest in having those information available and provide them yourself (in form of wiki / learndb or whatever you feel suits best as a format). Just be aware that keeping those information up-to-date will be an insane effort (as others already pointed out). At least this will be the case for trunk-builds regarding the frequency of commits. Then again, you will not find many supporters for stable-builds online since the online community is vastly dominated by trunk players.

"I find myself wanting to look up more information than I ever have before"

Right now, if you really want to find out how something actually works you have to search the git log for the last time the mechanic has been changed and look up the relevant parts in the code.

So how does one edit learndb, anyway?

https://loom.shalott.org/learndb.html#learndb

[Kate]

If you want to set up a wiki with open registration then maybe just do that! Not sure who these "people who need to take you seriously" are who you are expecting to do it. As far as I know the hosts of the existing wiki don't have time to upgrade the wiki software to fix the spam issues that caused registration there to be closed, that could be wrong though and I'm sure you could contact the Chaosforge administrators about it if you liked. Certainly there's lots of unhelpful stuff in the LearnDB, I don't think arguing about that is likely to be productive in terms of making the wiki better though.

So how does one edit learndb, anyway?

It's editable in ##crawl with a bunch of different commands (see ??learndb).

(Edit: Tedronai beat me to it and linked the learndb help and generally said what I meant but better).

[tedric]

So how does one edit learndb, anyway?

https://loom.shalott.org/learndb.html#learndb

I should have known there would be a learndb entry on how to edit learndb entries

[Magipi]

By the way, how can you use the learndb now? The searchable version is outdated, and the html version has no search function. You can use the browser's search, of course, but is there a way to search for titles only?

[Tedronai]

By the way, how can you use the learndb now? The searchable version is outdated, and the html version has no search function. You can use the browser's search, of course, but is there a way to search for titles only?

In the ##crawl channel on freenode IRC you can ask Sequell about learndb entries by prefacing your query with "??" (or "s??" for full text search, not sure about this one though - cannot try right now).

[billob]

Don't ask what the wiki can do for you, ask what you can do for the wiki? Personally I found it very helpful a lot of times, and if it misled me, I am still blissfully ignorant of that fact. Maybe it helps that I am still playing 0.13?

Random thought, it would be awesome if lists like "monsters with natural holyness", "spells god X doesn't like", "spells Vehument inreases the range for" and whatnot could be spat out by the game itself, in JSON or something, and then be parsed by whatever wants it for whatever purposes. An API to request facts about the game if you will. Perhaps that could also be used by an ingame Crawlpedia, Sid Meier style? ^^ I'm not saying the docs could be automated, but some parts of them probably could be, especially list-y things. Of course that would probably quite the initial effort, but assuming this game will stay around and keep changing for long time, it might actually amortize some day. What sayest thou?

[Psiweapon]

Also there's many problems that the wiki has beyond just being out of date or inaccurate, but I'll just limit my comments to the question of the accuracy of factual information on the wiki.

I beg you: Please elaborate.

I'm tired of gratuitous wiki bashing, I use both the wiki and monsterinfo, and both disappoint me occasionally. Different flavors of disappointment, yes, but calling the wiki "badwiki" doesn't make monsterinfo any better.

If there is some sort of beef between wiki guys and devs or goodplayers or IRC regulars, I'd really like to know.

[Kismet]

I like the learndb jokes, please don't delete them.

[Laraso]

If you want to set up a wiki with open registration then maybe just do that! Not sure who these "people who need to take you seriously" are who you are expecting to do it.

Because I don't want to make a new wiki if there is nobody else who is interested in maintaining one (I thought that's how it was, but clearly that's not the case!), and if it's possible to restore the previous editing permission policy that used to be on the wiki and get to work on updating the vast amount of knowledge the wiki already has, then I'd much rather prefer to do that instead. I'm fairly certain that the administrator of the wiki isn't going to suddenly change everything just because one person asks, so if that is going to happen at all then I would need help. The people who are knowledgeable about the game and capable of helping contribute information to a wiki are the people I want to take me seriously.

Also, please note that I'm not against fun nor do I hate jokes, but when anyone tries to say that the wiki has useless information in it they are overlooking the fact that the learndb also has its fair share of useless information.

EDIT: I'd like someone to explain to me how crawl development somehow makes maintaining a wiki disproportionately more difficult than maintaining the learndb. You're either editing the articles on the wiki or editing the information in the learndb, so what's the difference?

[and into]

Also there's many problems that the wiki has beyond just being out of date or inaccurate, but I'll just limit my comments to the question of the accuracy of factual information on the wiki.

I beg you: Please elaborate.

I'm tired of gratuitous wiki bashing, I use both the wiki and monsterinfo, and both disappoint me occasionally. Different flavors of disappointment, yes, but calling the wiki "badwiki" doesn't make monsterinfo any better.

If there is some sort of beef between wiki guys and devs or goodplayers or IRC regulars, I'd really like to know.

Well I have editing ability at the wiki—unless account was closed very recently—and I try to improve it; I'm not interested in "bashing it" but I do and will continue to caution people about it and take the info with heaping serving of salt, until the average quality of the advice and information has improved a lot from where it is now. Admittedly the same can absolutely be said about advice on Tavern, of course.

Anyway, the other problems with the wiki include a certain verbosity that while colorful can hide the important information or unwittingly mislead; "more writing = more *useful* information" is not always true and is in fact often not true on the wiki. A lot of the pages are info-dumps without any perspective about what is important for game play, for new players or for veterans, which can be misleading even when it is not incorrect. Sometimes there is the opposite problem: information is true but underspecified and given way too much weight; the racial equipment benefits of Demonspawn used to be guilty of that, for instance, until I edited it. A lot of the advice and strategy offered is bad or problematic in various ways.

[wheals]

I'd like someone to explain to me how crawl development somehow makes maintaining a wiki disproportionately more difficult than maintaining the learndb. You're either editing the articles on the wiki or editing the information in the learndb, so what's the difference?

Generally it's that, as someone said above, people are generally querying the learndb in public (though it's possible to look at it online at https://loom.shalott.org/learndb.html or in a PM window with Sequell). If someone there knows that the information is wrong, they can generally edit it right away, and the fact that every commit is announced on ##crawl also makes this a bit easier. I'm not convinced as duvessa is that a wiki is ultimately doomed to be bad, but there are problems with the difficulty in actually fixing things, and it doesn't have some of the advantages of the learndb.

[duvessa]

EDIT: I'd like someone to explain to me how crawl development somehow makes maintaining a wiki disproportionately more difficult than maintaining the learndb. You're either editing the articles on the wiki or editing the information in the learndb, so what's the difference?

I don't think anybody has ever actually claimed that a wiki is inherently harder to maintain than the learndb. There just happen to be far more competent people editing the learndb than the wiki and this is unlikely to change. (In my first post in this thread, "wiki-like entity" includes the learndb.)

[KoboldLord]

I dug up and old thread on his same topic:

viewtopic.php?f=5&t=4338&hilit=wiki

In principle, it would be nice to have a wiki that works properly, and there's no reason that having a good wiki is a thing that cannot happen, but there's some bad blood in previous interactions between certain admins there and people who actually know what they're talking about. Not all the admins, mind you; there have been several people since then who have tried to rehabilitate the Crawl wiki by that method and maybe someday they'll succeed.

It really cripples a potential contributor's enthusiasm for the project when an attempt to fix obviously wrong information is answered with stonewalling and a reversion war, based on the premise that no contributions can be bad contributions and people who want to look things up on the wiki can distinguish the good information from the bad on their own.

The thread I dug up is an old one, though, so maybe it's time to give it another try. Good luck to you if you do that.

[dpeg]

I would like to say that as far as I know, developers are not really involved with either learndb or wiki. Both are player-driven undertakings, as it should be.

[duvessa]

If you do want to fix the wiki, my advice is actually to get rid of the whole thing and start over. Very few of the pages are actually good as-is aside from the automatically generated ones (which are easy to generate again, that's the point of them) and many should really be removed altogether (hi, character guides). Rewriting them won't really take much longer than fixing the incredible mess, and it will be much harder to miss something that way.

I would like to say that as far as I know, developers are not really involved with either learndb or wiki. Both are player-driven undertakings, as it should be.

Thats incorrect, elliptic deletes plenty of learndb entries

[wheals]

Thats incorrect, elliptic deletes plenty of learndb entries

No reason to insult MarvinPA like that by leaving him out

[Arkhan]

Oh I was under the impression that "badwiki" was actually a joke and the wiki was meant to be the primary source of information, with the learndb for experienced players that just want a quick touch-up on something they knew but forgot. I know that I used entirely the wiki to learn the finer aspects of the game, and I still have a bookmark on the weapon speed page. And I know it updates at least on some sort of frequent basis because they're keeping up with super recent changes like to the Long Blade names (claymore, etc.) and if you look around you'll find entries on stuff that isn't even in stable yet like Dith, so I always just assumed that the wiki was the intended place to go.

[damiac]

The wiki is hugely helpful and informative, and the information is typically up to date with whatever version it claims to be accurate for. Yeah, there's a little issue with advice and strategy, but the learndb's just as bad(actually worse, because the learnDB also sucks to use)

It sounds like the main problem with the wiki is the fact that not many people can fix the things that are wrong. If the wiki's under lockdown by whoever owns it, I guess there's nothing to be done about that, but I'd think that's the thing that needs fixing.

[dpeg]

Could the wiki automatically include the learndb entries (not just the link) in a box, for various topics? That way, the permanently updated learndb is always included in the wiki, and wiki readers get a different (if really terse) outlook on the other side.

[reaver]

It really cripples a potential contributor's enthusiasm for the project when an attempt to fix obviously wrong information is answered with stonewalling and a reversion war, based on the premise that no contributions can be bad contributions and people who want to look things up on the wiki can distinguish the good information from the bad on their own.

This. The problem is that Crawl Wiki has developed in its own insular space, separate from the core Crawl community. Early in Crawl wiki's life is was targeted by spammers, which resulted in the "temporary" solution of forcing everybody to email the admins to get an account, which has now been going on for the majority of the Wiki's lifespan. This division has further developed into the aforementioned "all contributions are good" culture, where it's easy for ignorant fools to add nonsense but very hard for experts to remove it. As a result, nobody in the core Crawl community respects Crawl wiki, leaving them to languish into their own delusions. Crawl admins have done things like delete the account of active developers let Elynae's species difficulties entries be lengthened in contraction of their original purpose or allow the Mummy Necromancer guide to stay up for months (before eventually admitting it was so bad it probably made the game harder for whoever followed it.)

If I was given the task of making a good Crawl wiki, here's what I would do:

Torch the current version to crispy ash. (or at least archive if so future historians can laugh at it)

Create a new policy: Every fact must be line referenced to the source code.

Make a few preliminary pages for the gods and such.

Create an automated system for commits to be place on a "unaccounted for commits" page.

Whenever a commit comes in which changes the source code associate with a wiki page, connect it to the commit.

Contributors come through and "resolve" commits by checking all the wiki pages whose associated source code has changed, updating the article. This work is peer reviewed, then sent to the Trunk page.

Pages of current version are semi-protected.

Since that's all a lot of work, I'll just note that the Tavern user MoogleDan is a semi-active admin and you can PM him any corrections without getting an account.

[Arkhan]

Ya know, various other games have wikis. Those games have thousands of players (more than crawl) that, conceivably, are trying to edit information. What are they doing different that makes it look so easy?

[duvessa]

Having thousands of players maintaining them.

[mikee]

This topic seems to come up every month or so, and I find myself participating in a lot of these (you might want to search past topics if you really want to know what people have to say about the wiki or the learndb). Mostly the discussion doesn't go anywhere (I mean, do you really want to know what I find so silly about that zin page that you linked? I doubt it).

I think this will make some people angry at me, but I've long wanted to ask: can we remove the link to the wiki from the CDO main page?

[Magipi]

(...) do you really want to know what I find so silly about that zin page that you linked? (...)

Yes

[Patashu]

Ya know, various other games have wikis. Those games have thousands of players (more than crawl) that, conceivably, are trying to edit information. What are they doing different that makes it look so easy?

Those games aren't being constantly updated to add monsters, remove monsters, add spells, remove spells, tweak core gameplay features and change the optimal way to approach the game.

There's also the problem that in Crawl, you play for hours to get a win, and if you die all that progress is 'lost' (you still maintain the practice and knowledge from it, of course). So any information on a wiki or similar resource is expected to be good advice, optimal play, etc. unlike in other games where it's not as big of a deal if you mislead by accident or by omission.

[reaver]

(I mean, do you really want to know what I find so silly about that zin page that you linked? I doubt it).

Could you tell me this? I'm curious is it this error you're talking about. Is it the emphasis on choosing books for recite when that was long removed, the absurdity of using Vitalization to test potions, or something else I missed because I fear if I stay on Crawl Wiki pages too long I might get infected.

I think this will make some people angry at me, but I've long wanted to ask: can we remove the link to the wiki from the CDO main page?

That actually almost happened at one point. I would try bringing it up on ##crawl-dev.

[Kate]

I like the fact that there's no mention at all of the thing that Sanctuary actually does, personally. That's, uh, sort of a really big deal.

Other completely untrue things at a glance: imprison reciting to monsters/requiring a reciteable target (true in 0.7!), lifesaving costing piety, sanctuary granting rMsl (what?). This is a whole bunch of stuff that is trivially testable without being some kind of a Crawl expert or even looking at the source in most cases. Like, this page is really pretty bad. Maybe not such a great choice of example!

And as others have pointed out, when obviously-wrong things like that show up in small chunks in ##crawl as somebody queries them, it is fairly likely that somebody will spot the error and fix it (this could potentially be happening for the Zin page now! But most of the time when somebody reads a wiki page, nobody else sees it).

Anyway yes, I would advocate that anyone interested in there being a more useful wiki should consider starting a new wiki or at the very least contacting Chaosforge staff to find out how likely it is for registration there to be re-opened.

[ontoclasm]

Also, their chosen example for having two recite books available, and having to choose between them, is a skeleton.

Only Ablutions can be read to a skeleton.

[Kate]

That's a 0.14 change, though, so there's one example of a case where the page is in fact accurate for the version it claims to be accurate for.