Member Discussions

Since my last post here, I went ahead and built a relatively fast, scaleable web
telnet client using flash and javascript (Source code available on private
request if anyone's interested). I wrapped that in a facebook app and spent
about $10 on advertising, and now the facebook app is getting around 50
new account creations a day, pretty much all of them complete newbies to
the mud concept.

I've noticed a couple of things in observing these newbies -- I'm working on
my own solutions to the issues, but if anyone else has noticed/solved these
issues, I'd be curious to hear your thoughts :)

- Newbies have no idea how to interact with the command parser. If verbs or
nouns are case sensitive, they will likely be frustrated by that a few times.
They'll probably try "wield hatchet and wear boots" or "wield hatchet, then
wear boots" -- anyone who isn't at least vaguely familiar with the command
line from their OS of choice will probably have trouble figuring it out, and
anyone who isn't fairly familiar with a unix-ish command line will probably
have significant trouble with an LPMud. Additionally, they might struggle with
some of the nouns -- experienced mudders tend to use 1-word ids, while a
lot of the newbies try to use the full query_short() (in LP terms) to refer to an
object.

- Newbies have less patience these days. There was a time when playing a
MUD was enough of a novelty that the game didn't really need to reward the
player and make them feel competent, accomplished, etc. for at least a few
hours, but now they want to feel like they know at least something useful
after a minute or two. I've experimented with spawning the newbies in the
mud's traditional newbie school, as well as spawning them in the general
player area with an object that gives contextual hints and guides them
through their first few kills, and the latter approach seems to lead to much
longer average session length. YMMV.

That said, I think there's a TON of players (and some good money, if you
swing that way) available to the first mud that figures out how to dramatically
improve on the traditional mud learning curve, and I'm working like a maniac
to make sure that happens sometime soon.

If anyone else is doing work related to social network apps or java-free web
based clients or anything in that vein, let me know -- I'd love to share code
or ideas.

Cheers,

Moceanu

I havent really investigated the Facebook app idea, how easy was it to integrate a telnet client into it? Any pitfalls or advice you have for those of us with little or no experience creating a generic like telnet client and the wrapper for it?

Id be very interested in doing something like that. Perhaps some of your work could be released to the community to help others get connected to the new social networks and promote MUDs.

I think its a really powerful tool to use. Some of those Apps (Scrabulous, for instance) generate a ton of cash (It generates 25,000 a month in AD revenue).

Just a thought.

Thanks for any insight! Sounds good, great job!

-Syn

Command parser - Now that you have your own client, change your UI and get rid of most of the command-line interaction. Click to move, equip, and do other common actions. I can go into excruciating detail if you want.

Less patience - IMHO, you have 5 minutes to demonstrate why your game/experience is worthwhile. If you spend 4 of those minutes creating a character, selecting attributes, and whatnot, you've already lost. In the first 5 minutes, you should have the players do/experience something that they can't do/experience with WoW/grahhics.

1, I dont think the UI is there for the entire game, it was just a means to allow him to create the wrapper for facebook.

As an alternative to that, perhaps add a How-to section to your facebook app that gives some common tips on parser interaction, use some metaphors or examples that would be more up to date and common for average users.

2, I would agree that you have a short period of time to get someone interested. I would highly disagree that character creation is a waste of time, and that shortening that experience is to the benefit of every game.

I would give the example of City of Heroes. People can spend HOURS on character creation there. If anything, have some prebuilt easy packages (Like GodWars 2) and then a more complex character creation for 'vets and people who want to explore that aspect more.

Further, you should create a newbie area. Make it engaging and fun. The reason the traditional one is boring, is because its boring. Make one that isn't really a 'NEWBIE ZONE', one that is compelling, interesting, and starts off your story and theirs. That will go much further then the old boring newbie 'school' thought, and the new one of just drop them into the fray and say have fun.

-Syn

I'm not a coder or know anything fancy about computers, but let me take the practical point of view.

A couple of things:
1. If the newbies have very little patience, then just take out the newbie school and toss them into practice grounds. They want action right away, unlike aetolia which has, like, a 5 hour tutorial. So, just tell them the basics: kill <target>, get <item>, wield/hold/equip <item/armor/weapon>, help <topic>, and chat <message>. They'll figure it out from there.

2. If they don't know how commands work, give them small hints like 'You can only do 1 command at a time' and 'You can't use certain commands'. And if the social network is buggy, like the public chat bothers some, make sure to put in ignores commands. Also, be sure that they aren't bugging you for help like 'How do I use this or that' or 'How do I play?'. (If they ask the 2nd, tell them to go find out themselves)

3. And if there are more problems concerning the commands, just force them through the tutorial. That way, they aren't trying any stupid things like you said.

Hope this helps!

-Faar, Master of Annoyance

I'm interested to know if any of these facebook newbies are actually sticking around and becoming regulars. Was your $10 worth it?

For those wanting to try this with an alternative method: Set up a java telnet client that looks pretty (maybe a dressed up MudMagic client if Kyndig allows dressing up, otherwise the stock one for ease), and set a Google AdWords account to send traffic straight to your telnet client.

When signing up with AdWords you get a $50 credit anyway. After they take out administration it drops down to $40 or so, but if you can get say 20 - 50 cent clicks that easily turns into 100 right there. I don't imagine the market you're looking for is to especially crowded so the price might be as low as 10 cents a click which would be a good payout on the money.


--
MM

I believe $10 bought me only 20 clicks, but somehow we're getting 50 account
creations a day thanks to the viral nature of facebook apps. So yeah, $10 to
jump-start the process was definitely worth it ;-)

Syn,

Building a telnet client w/o java was actually fairly easy from a technical
perspective - just make sure that you're building on top of good libraries. I
used Ruby on Rails, Juggernaut, and DRb. And since it was all based on Ruby
on Rails, wrapping it in a facebook was a fairly easy step using the RFacebook
plugin. So bottom line, it's not terribly difficult as long as you build on top of
some good libs.

That said, deployment can be a bitch if you don't have superuser access to an
http server with some spare horsepower -- a traditional php web host simply
won't do, because you need socket access and probably need to run a
custom daemon or two to maintain connections. My setup right now is using
three processes and about 70mb RSS on my solaris shell.

After a day or two of hacking on this project, I abandoned all pretense of
writing organized, modular code and just started hacking 'til I wound up with
the featureset I wanted, so right now it's definitely not ready for public
release, but if this proves useful, I'll try to release both the telnet app and the
facebook app for use on other muds.

I've been experimenting with an LPC object that sits in a player's inventory
and provides a JSON-based api for interaction with the custom client. Right
now, we're just using it to provide hp/fp meters on the side of the screen and
a "score" popup window with tabs to show skills, finances, quests, etc.
outside of the terminal window, but I think the long term possibilities of
special client features could be pretty cool. I'll keep you posted.

Cheers,

Moceanu

I'd also note that Python as a language has telnet support built into one of the libs and is a common combination with LAMP dev. Building a Python telnet client is as simple as importing the telnet library and piping input and output from the standard commands.

http://docs.python.org/lib/module-telnetlib.html

Check out the Objects and Examples for more information. Assuming you get the web interface down the telnet side of the client should be extremely easy to integrate. Not sure if Python meshes with the Facebook APIs though. Someone more knowledgable or interested in researching can probably comment there.


--
MM

Thats really interesting. I have a VPS myself so that shouldnt really be an issue, the back end that is.

Sounds like a really nifty idea. I think that this could really help get MUDs known again in a grass roots sort of way. Nice work.

-Syn

Python has http://code.google.com/p/pyfacebook/, which seems to be a pretty
decent lib for facebook.

About python's telnet lib (and ruby's, too): those libs are ptty much designed for
small scale botting -- they implement the basics of the telnet protocol, but if
you want to get things like color information and echoon/echooff interpreted
correctly, you'll have to do quite a bit of work on top of the stock telnet lib.

Right now, my application is a proxy between the user's browser and the actual
mud server. The mud sends output to my app, and the app cleans it and
converts it (converting ANSI color sequences to html span tags, etc) before
sending it on to the user. The proxy is in there because parsing and handling
colors and other escape sequence behaviors in javascript just seemed like a Bad
Idea(tm).

Synorel wrote: "2, I would agree that you have a short period of time to get someone interested. I would highly disagree that character creation is a waste of time, and that shortening that experience is to the benefit of every game.

I would give the example of City of Heroes. People can spend HOURS on character creation there. If anything, have some prebuilt easy packages (Like GodWars 2) and then a more complex character creation for 'vets and people who want to explore that aspect more."


Two somewhat contradictory thoughts here:

1) If your MUD character creation is as fun as designing costumes on COH then go for it. (Most/all MUD creation systems that I've tried have been signficantly un-fun.)

2) The first 5 (or 15?) minutes of gameplay should give players a good idea of what gameplay is about. If you have extensive character creation, even though it's really fun, I'd still recommend some represenative gameplay in the earler experience since (in the long run) the game is not about creating characters/costumes.

I dont see that it's contradictory. I also wouldnt call your character creation process an attachment of the initial GAMEPLAY aspect.

As, your creating, and if that is fun bodes well for the initial game play.

-Syn

I'd be willing to venture that some level of complex creation will create an initial investment in your game which if it proves to be fun enough will draw the players back again later. The more investment a person puts into a given activity, the more likely they are to return to this.

On the other side of things, a game that requires no investment and just throws them into the midst of playing is no different than the average flash game. Completely forgettable within minutes.

This isn't to say that it should take more than 5 - 8 minutes to access the basic content of your game, it's just that the creation process and the player->character investment in which it entails should not be overlooked.


--
MM

This sounds like a brilliant idea - really bridging the gap between MUDs and
modern/younger internet users. I would have never thought to integrate it with
a social networking site like that - I'm really very impressed.

Would you mind posting a link to the Facebook app so that we can try it out?


Thanks!

We've actually been toying with creating a "newbie" race to be the default of all
new players, so we could allow them to play for a few hours before they have to
go through the proper character creation process. A little backwards, perhaps,
but I think it might strike the right balance for now. YMMV.

We've actually been toying with creating a "newbie" race to be the default of all new players, so we could allow them to play for a few hours before they have to go through the proper character creation process.

I've known circleMUDs to do this in the past and then later on have class and race selection done via scripted in-game mobs that you went to. The story if I remember correctly on one of them was that you were born as an incorporeal entity and only later did you actually have a corporeal form so that in essence you were more or less a bodiless soul until you chose a race.

Aye, I've seen that too. Is it only in a select few, or was it added in as a stock sort of thing? Memory fails me.

- Newbies have no idea how to interact with the command parser. If verbs or
nouns are case sensitive, they will likely be frustrated by that a few times.
They'll probably try "wield hatchet and wear boots" or "wield hatchet, then
wear boots"

Is it possible for you to elaborate a little bit on this? Was there a high occurrence of "do this then this" or "do this and this" or are those just examples of people not understanding how to use the commands?

Is it possible for you to elaborate a little bit on this? Was there a
high occurrence of "do this then this" or "do this and this" or are those just
examples of people not understanding how to use the commands?

It really seemed to vary -- the two syntaxes I saw most commonly were
"<verb> <noun> and <verb> <noun>" and "<verb> <noun>, then <verb>
<noun>". There certainly wasnt enough consistency there to actually
implement it.

The newbies also seemed to try actions that experienced players probably
wouldn't try -- for instance, convention on my mud is that any noun in a
long desc should be examinable, but many nouns in NPC-envirospam aren't,
and the newbies tended to try to interact with nouns from envirospam before
interacting with nouns in the long descs. I think that might be indicative of
the fact that newbies tend to focus in on the bottom line of the buffer
because of the sheer information overload, while experienced players can
more effectively process a whole screen full of information. I dunno, though.

Cheers,

Moceanu

Interesting thing I ran across today -- a page about Game Neverending, a defunct browser game created by Ludicorp before they scrapped it and went on to create Flickr.

Personally, I think a well-done browser-based MUD could be quite successful -- look at all the people who play those mindlessly repetitive "build a kingdom" games where you punch in numbers and ooh, look I have 3 stables! It's a lot of work, but it could also be pretty cool.

My introduction to mudding was actually through a web-based MUD called Rock: Conflict of Interests, back in 1997 when you'd be lucky to have 56k. Click an item to pick it up, click [equip] to wield it, no command parser needed -- although you can still type in commands for more complicated things and chatting.

The nice thing is that the browser gives you everything that the multimedia MUD clients have been trying to do for years, except the players don't have to download anything. You could also do it in Flash.

MUDs + Web2.0 (as much as everyone hates that phrase). MUD2.0 perhaps.

I recognize that it would be a significant effort to parse a large amount of various newbie command inputs, but if the number of dropouts is significant you might really want to think about making that smarter command parser. I'm a little biased because I'm a linguist but I think there might be some worthwhile effort to be spent in developing a really open-minded command parser.

If you want to talk about it conceptually feel free to shoot me an e-mail at mark.mrwizard@gmail.com or on aim at evilestmark

I think inform 7 has a really flexible command parser. They also built a programming language that is supposed to be very close to natural language.

Also, help messages could automatically be displayed if a player types a command that doesn't work.

Well, since this has been revived...

You can have more flexible command parsers, and really, Bartle's MUD (written in what, the 80s?) has a more flexible parser than 99% of the muds on TMC. However I don't think it's necessarily a more flexible parser that's needed but better ways of teaching that parser to new players. I've sat down and watched people play interactive fiction for the first time (in I7 in fact) and flexible parser or no, that parser isn't picking up more than 10% of the commands that person is typing into the game. However once they learn the conventions of the grammar it goes a lot better.