[Contents]--[Abstract]--[1]--[2]--[3]--[4]--[5]--[6]--[7]--[8]--[References]
Recently MUDs were subject to several studies as is reviewed in section 6.1. A further study based on MUDs is described in section 6.3. This study is - to the author's knowledge - the only study to date researching navigation and spatial perception in textual virtual environments. The MUD the case study is based on is described in section 6.2.
There are several types of text-based virtual environment system that are used as network-based games. The systems commonly have an abbreviated name starting with an "M" to point out that they are multi-participant systems - for instance MUD, MOO, MUSH, MUSE and so forth. The term "MUD", which stands for "Multi User Dungeon" or "Multi User Dimension" is the most general term describing these systems. Using this term in the reminder of this thesis shall not restrict the generality.
As textual virtual environments do not use graphics the level of abstraction is much higher than in graphical virtual environments. The textual description of objects and locations provides unique possibilities to describe space but it does not allow the user to change viewpoint as easily as in the graphical systems. Textual virtual environments - like books - rely a lot on the imagination of the user -- one of the reasons why they are very popular for certain people and are rejected by others.
MUD systems often form a social community. MUD systems therefore are interesting research grounds to study formation of societies and social behavior. Most previous work on MUDs is in the field of social sciences. The trend towards seeing MUDs as research tools had its start with the work of Pavel Curtis for the LambdaMOO system at Xerox Parc. This work focused on social aspects of textual virtual environments. For instance in [CURT92] Curtis describes the different modes of communication in the LambdaMOO system. LambdaMOO is a Lisp based system run by Xerox.
Related research into sociological and psychological aspects of MUD-using (sometimes termed "mudding") has been done by Amy Bruckman at the MIT using the MediaMOO system (see for instance [BRUC92] and [BRUC94]). Bruckman performed interviews in the MUD and extensively studied the structure and internal relationships in a MUD society. Further work by Bruckman used the MUD as meeting place for a professional society. MediaMOO is a MUD that is accessible only for media professionals. In this system group gatherings are organized that function very much like real life get-togethers. There are gathering rooms, ball rooms, meeting rooms and so forth. People may meet in smaller or larger groups [BrRe93].
The same direction is followed in Curtis' more recent work, which is described in [CuNi93]. Curtis pushes the concepts of MediaMOO even further towards Media Spaces. In the AstroVR system - a system designed for astronomers to exchange ideas - it is possible to give lectures and to collaboratively use software tools or to share documents using the MUD. The other system described in [CuNi93] is the Jupiter system. Jupiter break the text-only barrier by providing multimedia channels for sound and video.
In the Jupiter system (see also [CuNi93]) audio and video channels can be used in the MUD. This requires special networking and processing facilities. In addition to the general communication facilities of MUDs the audio channel provides possibilities for a sort of preview into nearby rooms. When there is a meeting going on in one room the muffled sounds of that meeting still can be perceived in the next room. This fosters a deeper understanding of the spatial relationships of rooms in the MUD which even can be enforced using 3D sound.
Other research work is done to support collaborative information retrieval using MUD objects called "Gopher slates". These slates are described in the MUD as small portable computers. Since they are objects they can be given from one user to the other - including all the data that has been downloaded into the slate already. Placing the slate onto a conference table makes it accessible to the other people in the MUD room [MaOs93].
Most researched MUDs are used mainly as communication tools - as a sort of text based on-line conference rooms. An example for this is a MUD supporting communication in a system operators group [EVAR93]. Another system is used for discussions in the VR-SIG (Special Interest Group) of the British Computer Society. Meeting-MUDs normally are equipped with a recording facility. If a member of the groups enters a discussion later it is easy to rewind the ongoing discussion and to pick up the thread of the discussion easily. Protocols are sent by email to those people who are temporarily unreachable. For other research on MUDs see [TrDi94] and [BRUC94c].
Summary
This review shows that most work on MUDs has been done in the fields of social and psychological studies and on supporting collaboration. No previous work has been done to research spatial perception and navigation in MUDs and to use MUDs as spatial user interfaces. Only [ERIC93b] hints at the potential of these systems as social environments and as spatial user interfaces at the same time.
The Igor MUD is set up in a typical adventure game setting containing wizards, dragons and evil orcs. Such a setup includes magic as a normal thing. Typical transport system in Igor therefore are a flying dragon and magic portals. In another MUD based on a Star Trek setting the same magic features would be named shuttle and transporter beam. A setting that considers magic as normal is a good place to test "magic features."
Additional constraints are posed on the player by the maximum carrying capacity, fighting strength and the need for other treasures, like keys or certain tools to open doors or chests. MUD environments normally do not present all navigational information openly as this makes the "quest" more interesting and entertaining.
The following text shows how Igor's Newbie (novice player) Library explains adventuring:
** ADVENTURING **
On later years, it has become more and more complicated to adventure and
gather experience. This book was written to be a guide for adventurers.
As everyone knows, the meaning of life is to gain experience and raise
your level. This can be done here at the adventurers guild, but a small
fee is sometimes required if there is not enough experience. You may also
raise your stats here, but that always costs experience.
The best way to get experience is to exterminate bad monsters that stroll
around our city. To gain money bring items with some value to the shop and
sell them. Another factor you will notice is the alignment. Alignment is
a measure of how your characters moral status. Killing evil monsters makes
you good and killing good monsters will make you bad. Killing players
will make an Arch mad, and will ultimately render your character great harm.
The altogether top goal is to become wizard. All wizards will be assigned
a castle, which they can rebuild any way they like. Most areas in Igor are
made by wizards who by accepting wizardhood becomes immortal and leaves the
player level of the game. Moving on to the challenges only wizards know of.
Remember wizards are not normal players, they are creators of the game.
Support them with ideas, typo spelling errors and tell them bugs.
Wizards are not supposed to answer detailed questions about the game,
not meaning players can't chat and discuss with wizards.
Exampli Gratia: Do not ask of powers of swords, armours etc.
Do however use idea, typo, and bug. Further info of these in help.
Note: As this is a glorified version of real life, you better get drunk
if you are hurt. Then, you will heal faster. Never trust privacy. Wizards
have eyes everywhere, which they use to study usage of their castles etc.
The navigational task in MUDs is different from real-life navigation as MUDs are virtual environments allowing magic features in navigation. Examples are teleporting or areas that do not follow Euclidean geometry. These features can make it difficult to answer the question "where am I" after walking through a "shimmering magic portal".
The social hierarchy in the MUD
Solving a quest earns the player "experience points". These points can be spent to raise the player's "stats" - that is her skills in fighting, magic, carrying capacity and so forth. Treasures not needed to solve a quest can be sold in the shop to get MUD money. Money is needed to buy better weapons and armors and to buy food. Food is needed to heal wounds suffered during adventuring.
Depending on the experience points acquired players can raise their "level". New players start at level 1 - these players are called "newbies". The goal of every MUD player is to reach level 20 or above. At level 20 the player can decide if she wants to continue playing the game as an "avatar" or to change sides and become a "wizard". Wizards are not allowed to play the MUD - instead they learn to program the system. Their task is to enhance the system, to design new rooms, to troubleshoot old areas and to design new quests.
Higher level players may learn magic skills which range from simple attack spells to teleporting items or themselves. Only wizards are able to do general teleporting. This type of navigation is far more efficient and fast than walking the MUD. Most wizards therefore prefer teleporting over walking and soon stop navigating the MUD the normal way.
Wizards are responsible for administrating the MUD. In Igor several wizards are always appointed the duties of "law" and another group of wizards decides which new areas may be opened to the public. The "law" group is allowed to expel players from the game when they are severely misbehaving according to the social norms of the system. This strict hierarchy of players is typical for most MUDs. There are systems however, notably the MediaMOO system, a democracy [BRUC94b], which implemented different society structures.
Fig. 17: The center of Igor (the backbone).
Figure 17 shows the backbone as a set of rooms of the same size that are arranged in a grid. This arrangement is an abstraction used to draw most sketches in this thesis. MUD rooms may have different sizes and the connections between the rooms may have different lengths. The size and length is communicated to the player using the description of the room. If players really perceive rooms sizes as different using the descriptions is one of the questions researched in the study described in section 6.3.
You are in the local village church.
There is a huge pit in the centre and a door in the west
wall. Beside the door is a button. Ghosts can be revived
here. Dead people and people who have lost their
heart beats come to the church and pray.
There is an exit south to the village green.
There is a clock on the wall.
A newbie area is north of here.
Obvious exits: south, west and north.
North of the church the "newbie library" is located. This area contains documentation on how to use the MUD. Exiting the church to the south leads to a green meadow where it is possible to buy a newspaper to get information about important events in Igor. This place is one of the landing places of the flying dragon, a transport system. The meadow is connected to the rest of Igor by the main road which leads east and west from this place. This road is an axis most players start to explore.
Welcome to the I.M.S. Post Office! Here you should come to read your mails. Valid commands are: from List your mail messages. list List the available aliases. mail <addr> Send a mail to someone. node List the accessible nodes. read <folder> Read your mail messages. There is one obvious exit: north. You can get help with the command 'help mailread'. In the reader mode you can type '?'. To read mail in e.g. 'received' folder type 'read received'. Comments to postadmin.
You are in the local pub.
A vellum scroll hangs on the wall.
You can order drinks here.
First class beer : 12 coins.
Special of the house: 80 coins.
Firebreather : 275 coins.
A list of rules is attached to a Go board.
Obvious exit: west.
Go player is here.
This is a very important place to socialize and to get drunk. In MUDs players heal faster from their injuries when drinking or eating.
You are in a shop. You can buy or sell things here.
Do 'help shop' to learn the commands.
There is an opening to the north, and some shimmering
blue light in the doorway.
Obvious exits: south, west and north.
The shop has a short description but except for the church it is perhaps the most important place in the MUD. Every user who wants to do adventuring needs the shop to get equipment. Since users cannot keep equipment when they log out they have to sell all their possessions first or lose them - only the amount of MUD money a user owns is saved for the next login.
The guild is always a busy place - it happens very seldom that nobody is at the guild. The message board is located there, and people can raise their stats.
You have entered the Adventurer's Guild.
This is the main gathering point for adventurers all over Igor.
There are guild masters present who will help you gain your
new level or stats when you are ready. Use the command 'cost'
to know, and 'advance [level, str, con, int, dex]' when you do.
To see what perilous quests you must undertake, use 'list'.
There is a strange, shimmering blue light in the southern doorway.
From here the role-playing board is up and the LPmud board is east.
There is no longer a shimmering door here.
Obvious exits: east, north, south, up, west and down.
A graffiti plank, the Adventurer's Cat Board, a word box and a book on a chain are here.
The adventurer's guild is only one of several guilds in Igor. The other guilds are not open to the public - instead they represent exclusive "clubs" of players. Players often have to solve a quest before being admitted to such a guild. Most guild memberships provide players with special abilities.
This section describes the center of the Igor MUD, which is a village called Larstown. In the center the important infrastructure for the game is located. These locations are arranged in a structure known as the "backbone" to wizards as this structure is identical to all MUDs of the same type. The main facilities in Larstown are the church, the shop, the guild, the post-office and the pub. They are arranged along the main road which leads far beyond the center of Igor and is the major linear element in Igor's structure. The church is the most important landmark in the game as this place is visited by every player at least once in every session. The locations described above can all be reached using ordinary "walking" navigation.
This section shows the close surrounding of the center of Igor. In a sketch-map several areas "close" to the center are pointed out. Not all of these areas can be reached using "walking" but are connected to the backbone using magic features like teleporting or magic portals. A commonly used distance measure in MUDs is the number of rooms to travel.
Fig. 18: Areas near the center and their distance from the church.
Rooms in the MUD often are arranged in a regular grid by the designers of areas. However rooms seldom are described as having the same size and therefore they are perceived as having different sizes. When navigating the MUD layout problems according to this grid not necessarily are noticed.
The problems of the regular grid assumption can be shown easily using two parallel paths the left of which goes into a house and leads out of the house in the back to meet the other path again. Inside the house there are three rooms in sequence on the one path - on the other path there is only one room. Logically the three rooms inside the house should be smaller than the single room outside of the house - otherwise this structure is distorted. As the one room is in the "open air" it is not unusual to describe it as a forest path which is a description that gives no hint to its size. However if this room were described as a small cozy place next to the house and the three rooms in the house were described as "huge ballrooms" there were a problem. Assuming a regular grid with equally sized rooms and with no empty spaces in the grid there is even a problem with the direction the exits take from the rooms (see figure 19).
As this figure shows, there are three possibilities how to perceive this structure. The left sketch is based on the uniform room size assumption. Unless there is another distortion this sketch requires the connections between the rooms to be different to "fit". The problem can be solved as in the middle sketch when the single room on the right path is assumed to be bigger. Assuming both a uniform room size and a fixed grid leads to the rightmost sketch. As the sketch shows the exits that exit the left top room to the north have to lead a detour to enter the next room from the south.
Fig. 19: Perceived room-size and the regular grid.
Normally such a structure is not found in MUDs as the connecting room would need to have two exits to the south or to the north. This poses a problem on its own as there is only one "south" command. Still the principal problem of size perception in MUDs cannot be ignored.
The spatial problems in the perception of the MUD layout can be classified into four types of problems, two of which are directly influenced by the perception of the room sizes:
The following log file shows an example of either two missing or two overlapping rooms. It shows a walk from the pub to this small room. The player commands are shown after the ">" prompt and are given in the short form. That means that "w" means the command "west" or "go west".
You are in the local pub.
(...)
> w
A small town yard surrounded by houses.
To the west you can see a small break in the buildings.
North street starts here.
Obvious exits: north, south, west and east.
> s
A long road going east through the village. The road
narrows to a track to the west. There is an alley to
the north and the south.
You see a liana hanging down from the sky.
Obvious exits: north, south, up, west and east.
> e
A long road going through the village. There are stairs
going down. The road continues to the west.
To the north is the shop, and to the
south is the adventurers guild. The road runs towards
the shore to the east.
Obvious exits: east, north, south, west and down.
> n
You are in a shop. You can buy or sell things here.
(...)
> w
You are in a small and dusty storage room.
You can see the shop through the opening to the east.
Obvious exit: east.
Drawing a sketch based on this log does not work under the assumption of an evenly spaced grid and uniform room size. Figure 20 shows that there should be two additional rooms on the main road -- assuming a regular grid and uniform room size. Another way to see this situation is to assume that the small room next to the shop and the pub are overlapping.
Fig. 20: The walk from the pub to the small room beside the shop
Most MUD players see no spatial problem in this situation because segments of a road can be assumed to be much larger than a "small and dusty storage room".
Guran's Cave
Another example for spatial problems can be found in an area called Guran's Cave. The detailed sketch of this area (figure 21) again assumes that all rooms have the same size and height and that they are arranged in a regular grid. This area shows that MUDs are not flat -- it extends over 5 levels. These levels are connected using stairs and ladders.
The dotted lines on figure 21 show the spatial discrepancy in the arrangement of rooms. For instance there are two levels of stairs missing and one room in level-2 seems to be much larger then the others, or the rooms in the lowest level are smaller than the other rooms - otherwise the rooms would not fit the spatial layout. These problems are indicated by the question marks in figure 21.
Fig. 21: Guran's Cave.
The Golf Course
Another interesting area is the golf course. This area was designed according to the layout of a real golf course - therefore this area shows no spatial problems. The rooms are very differently sized and spaced and the shapes and sizes of rooms is communicated in the descriptions. For example there are two rooms described as "L-shaped fairway". What makes this area interesting is the fact that many players have difficulty navigating it although there are no spatial problems and although the room descriptions are very elaborate.
Figure 22 shows a map of the golf course. The boxes in this map are drawn according to the outline of the textual descriptions of the rooms. This way rooms with overly long descriptions or rooms using very long lines in the description become visible. These aspects of room descriptions are assumed to influence the perception of room size (see section 6.3.)
Fig. 22: The golf course of the Igor Country Club.
Compare the sketch of figure 22 to the hand-drawn working-sketch of parts of this area shown in figure 26 (in section 7.1.2.2.). Notice that diagonal exits are used only sparingly.
Summary
This section gives an overview over the MUD used for the case study described in the next section. The Igor MUD is a large adventure game MUD and runs on a machine in Sweden. The navigational task in a MUD is similar to navigation in real environments. The social hierarchy of MUDs was explained according to which there are two classes of MUD participants: players and wizards. Wizards are very experienced MUD players although they do not play the game any more. The layout of the Igor system was described. Its center is represented as a village in which the major infrastructure for the game is located. The most important landmarks in the game are the church, the shop and the adventurer's guild. Distances in MUDs is often defined as the number of MUD rooms to travel to reach a location. MUDs provide either normal navigation by "walking" or by using magic features. These features can lead to navigational problems and distortions in the spatial perception when the MUD is assumed to consist of a regular grid with rooms of uniform size.
Section 2.2.4. showed that it is possible to learn spatial relationships from descriptions. This leads to the assumption that MUD users can create the same sorts of spatial memory representations whether they read about an environment in the MUD or would see it in reality. Indeed MUD users manage navigating the MUD quite well. As was descibed in section 6.2.4. MUD space shows several "anomalies". These make it harder to learn a mental representation of the MUD environment.
The experiment described in this section tries to isolate possible problems in navigating textual virtual environments. It tries to find out how the properties of MUD space provide navigational features not existing in the real world and how to exploit them.
One of the first plans for this study was to design a research MUD containing specialized areas. These would be void of other users and fancy rooms descriptions but allow studying single navigational problems one at a time. A set of test subjects would navigate those areas and their movements would be logged. These plans were dropped for several reasons.
Instead Jolanda Tromp performed a study on spatial perception using questionnaires in the Igor MUD (see section 6.3.1.). A follow-up study was conducted by the author using interviews in the same MUD (see section 6.3.2.).
The questionnaire study consisted of two questionnaires of 5 and 13 questions. They could be found in a room called the "research center". This room was located in the church where it easily could be found.
"There was never any direct contact between the author of the questionnaires and the respondents. Motivation to answer the questionnaire was the attractive introduction and the promise of a copy of the results as a reward. Anonymity was not explicitly guaranteed, because electronic questionnaires can be designed to save the name of the respondent without the respondents knowledge. Also most MUD participants use a nickname in the MUD anyway, so "relative anonymity" is already a fact. For the same reason no gender questions were asked as it is not uncommon for MUD participants to use the opposite gender in the MUD." [TROM93]
Both questionnaires were in the MUD for about four days. In that time about 600 people entered the MUD. Both times around 70 completed questionnaires were received in that time. For more details on the implementation of the questionnaire study refer to [TROM94].
Here is a selection of the questions asked:
| I (1) | Do you know how to get from the church to the shop without looking at your screen? |
| I (2) | How do you learn your way around in a new area? |
| I (3) | What strategies do you use to recognize where you are and where to go? |
| I (4) | What do you do when you are totally lost? |
| I (5) | Do you avoid getting lost or do you get lost on purpose, and if yes, why? |
| II (1) | MUDs are derived from role-play games. Do people role-play on the mud a lot or a little or only sometimes? Why do you think this is so? |
| (...) | |
| II (4) | Can you give an example of a very large room and a very small room on this mud and explain what makes them large/small? |
| (...) | |
| II (7) | Most personal rooms of wizards are not connected to the mud, you can't walk there, only teleport. Do you imagine these rooms in a specific place and why? |
| II (8) | If you were put at a random spot in the mud in an area that you know, would you be able to draw a accurate map of that area? |
| II (9) | Does going to places consist of a series of connected movements or is it a complete image in your mind? |
| II (10 | Do magic exits or rooms that are missing if you would draw a map disturb your sense of direction? |
| II (11) | Do magic mazes (places where you for instance go n, s and end up in a different place than where you were before) make it harder to image what the area looks like? |
| II (12) | If you teleport to a place through a magic exit do you feel that the place you teleported to is far or near the place you teleported from? |
The results of the questionnaires are contained in the following table 3, reprinted from [TROM94]. For the questions marked with an asterisk (*) "the percentages represent the frequency count for each answer category where the contents of each response have been scored over multiple categories". [TROM94, p.12] The following abbreviations are used in the table: descr = description, dist. = distance, I = 1st questionnaire, II = 2nd questionnaire.
| I (1) | Yes: 84% | A bit: 9% | No: 6% |
| I (2)* | Memorise: 62% | Make maps: 24% | Ask help: 14% |
| I (3)* | Recognition: 66% | Maps: 20% | Ask help: 14% |
| I (4)* | Wander: 31% | Ask help: 33% | Quit: 17% |
| Teleport: 16% | Never get lost: 5% | ||
| I (5)* | Avoid: 56% | On purpose: 6% | Sometimes: 24% |
| It happens: 9% | Never get lost: 5% | ||
| II (1) | A lot: 31% | Half/half: 32% | Other: 4% |
| Not at all: 12% | Missing: 21% | ||
| II (4) | Description: 49% | Other: 50% | No difference: 1% |
| II (7) | Depends on descr.: 10% | Image: 68% | No image: 22% |
| II (8) | Yes: 76% | Roughly: 4% | No: 19% |
| II (9) | Connected: 47% | Image: 39% | Both: 14% |
| II (10) | Yes: 30% | Sometimes: 22% | No: 48% |
| II (11) | Yes: 67% | Sometimes: 17% | No: 17% |
| II (12) | Far: 45% | Near: 13% | Depends on descr.: 5% |
| None: 15% | Depends on real dist.: 23% |
Pro and Con of the questionnaire approach
The main advantage of the questionnaire approach is the amount of data that can be collected in a very short time. The effort of data collection itself is low however the programming of a questionnaire object for the MUD requires a lot of time -- this has to be done only once however. The main disadvantage of questionnaires is that unclear formulations and deviations from the questionnaire questions cannot be corrected as there is no contact between the person doing the questionnaire and the provider of the questionnaire. This is one of the reasons why opinion polls often use questionnaires that are filled in by an interviewer who may focus the conversation on the questions asked and can explain unclear formulations.
(Lag is the time difference between issuing a command and seeing its results when this difference is disturbingly large. Lag is normal in MUDs since typing of text takes some time. An answer to a question never is in real time. Because of lag communication in MUDs often is multi-threaded, whereas in real life it is almost impossible to conduct two separate conversations at the same time. In MUDs it is common (although a bit impolite) to have several communications running at the same time, especially when using different communication channels like talk, whisper, shout, tell, and special guild-communication lines. People in MUDs refer to lag somtimes in a similar way as people in real life talk about the weather. Expressions like "Bad lag today, isn't it?" or "A laggy day" are quite common. See for instance [YOUN94].)
Discussions in MUDs are often enhanced by non-verbal communication which again is described textually. Non-verbal communication of this kind is called "emoting". Examples of emoting are commands to "laugh" which result in a line of text like "Juggler is falling down laughing" seen by other players in the same MUD room. Another frequently used way to enhance communication in the MUD is the inclusion of "smiley-characters" in the text. Smileys -- when looked at from the side -- are representations of human faces like ;-) (winking) or :-( (looking sad).
A MUD interview is started by inviting a MUD player to a talk and getting her OK for logging the talk. Although there is no legal obstacle against logging in a MUD it is at least considered very rude to do it without permission. Several people declined being interviewed because of the logging.
During the interview the questions could be copied from a text-file into the MUD client program. Additional sentences, explanations and formulations which were frequently used during the interviews could be copied from this text file as well. This method helped to prevent misspelling because of quick-typing and allowed using similar explanations for similar problems during the interviews. In the client program (MUDDweller) it is possible to edit the line to "speak" before sending it away. This allowed minor adaptations of the preformulated questions according to the situation and to answers given to previous questions during the interview.
The interview partners
As MUD players stay anonymous in the MUD, except when they decide to talk about RL (real life), it is not possible to get information about the background of the interview partners. As MUD playing requires Internet access and a lot of time it is a safe assumption that most MUD participants are in the typical age of students. I know of one interview partner that he still goes to school. Concerning the background I know that one of the interviewees is geographer, another one studies laws and a third one took at least a few psychology classes. Most of the interview partners are involved in programming - at least in their function as wizards.
The first interview series
The first interview series was based on a set of 6 questions starting from which a free discussion on navigational aspects of the MUD environment was performed. These interviews were free-wheeling talks that could not be formally evaluated as they yielded too unstructured data. According to the terminology described in [PREE94] this type of interview is a "flexible interview" and is useful in a very early stage of a study. Flexible interviews "(...) generally have some set topics but no set sequence and the interviewer is free to follow the interviewees' replies and to find out about personal attitudes." [PREE94, p.628]
The base questions in the first interview series were:
This first series consisted of about 10-15 talks. One interview was a group interview with three players present in the same room and developed to a sort of "round-table discussion". This method could be a viable approach for further surveys in MUDs. The interviews in the first series lasted from a quarter of an hour up to one and a half hours depending on lag and on the other topics discussed.
One of the reasons for the very unstructured results from this first series is that the task of interviewing in a MUD is a skill, that has to be learnt. It is difficult to persuade people to give interviews and if is difficult to keep the interview focused.
I see the results of this first interview series mainly as a pilot series. The logs from these talks are too unfocused for a formal evaluation. The talks were very useful for designing a much more focused question catalogue for the second interview series however and I learnt a lot about how to approach and "handle" interview partners.
The second series of interviews
The second interview series consisted of two parts. The first part consisted of 21 questions for players. The second part consisted of 10 additional questions for wizards only. Again those questions were slightly modified to adapt to a previous answer or to a special situation.
This second interview series was done in a much more structured way and it was tried to ask the questions always in the same sequence, which was not the case in the first series. According to Preece's terminology the second interview series consisted of semi-structured or prompted interviews. In semi-structured interviews the "interviewer often has a set of questions that she can draw on to direct the interview if the interviewee either digresses or does not say very much. Prompted interviewing is used to draw out more information from the interviewee." [PREE94, p.629]. The second interview series was a hybrid of the semi-structured and the prompted interview type.
In this second interview series lengthy discussions of related topics arose from the base questions. These discussions cannot be evaluated numerically but are used for the interpretation of the results.
All interviews started with a short introduction that explained that interviewees could stop being interviewed at any time, that they did not have to answer questions and that the interview would be logged to allow evaluation. It was promised to keep all logs strictly confidential which is the reason why no interview logs are included in this thesis.
The confidentiality clause was introduced after the experiences of the first interview series. Several interview partners agreed to be interviewed only because of this clause.
21 question for players
10 questions for wizards
Similarly the explanation in question 5 was inserted after the first few interviews showed that this question was not understood or understood in a false way. The question thus becomes a bit leading. Still most interviewees did not perceive the hierarchical structure proposed in the example sentence. The leading effect therefore is neglectable.
The second series of interviews was concluded after 10 interviews. The interviewees were 8 wizards and 2 avatars. The duration of the interviews was between 1,5 hours and almost 3 hours. The average wizard interview took a bit more than 2 hours. During this interview series no unrelated topics were discussed.
Sometimes it took several sessions in the MUD to find an interview partner who was interested to give an interview and had enough time. I should be pointed out that many mudders enter the MUD from home using modems and that such a long talk would influence their telephone bills significantly.
The following table shows the results of this evaluation. The numbers in parentheses gives the number of interviewees that used the keyword in question. A result of (3-5) mean that 3 interviewees clearly used the keyword and that there were two additional cases in which the classification was not entirely clear from the log. A few of the keywords in the table are based on the discussions during the interviews and sometimes these are not directly related to the question. Several keywords that occurred only once are omitted from the table except if they express interesting point.
| Player questions | |
|---|---|
| 1 | experienced (6-7) experienced only in Igor (3) |
| 2 | know all Igor well (6) only older parts of Igor (4) |
| 3 | no estimation (3) small part of Igor (1) less than two thirds (2) more then two thirds (4) |
| 4 | Adventuring(3-4) Communicating (10) Coding (6) |
| 5 | no structure (1) sees structure (4-6) loosely connected areas (5-6) Some areas have no clear borders (1) |
| 6 | objects / monsters in the room (6) keywords in the description (2) formatting (1) something unusual (5) the way it is described (3) functionality (2) I remember only routes (1) centrality (1) if it help visualizing the room (2) activity in a room (1) |
| 7 | line-like elements do not help (2-5) they help (4) Example Main road (6) Example road in the plains (2) paths give a feeling for distance (1) paths provide directionality (1) different description styles can be a line-like element (1) |
| 8 | I dislike / hate it (2) I like it (6) It helps in general orienteering (9) It does not help when running (1) I sees room description as a "picture" and do not read it (1) |
| 9 | I notice nothing when running (2) I use short descriptions (3-4) I see nothing when running (2) I use short descriptions (3-4) I notice keywords when running (0) I notice only major changes (1) |
| 9a | no feeling of time (2) feeling of time (7) |
| 10 | prefer second method (7) prefer third method (2) prefer a mixture (1) method 2 and 3 make the area more interesting (3) want a "rich" environment (6-7) want a not too rich environment (2) there are no long documents in a MUD (1) |
| 11 | teleporting is disturbing (0) teleporting is helpful (7) they are hard to map (1) not hard to map (2) useful only if there is a reason why it is used (1) people have to learn how to use it (1) they need a clear description of what is going on (1) it is unpleasant when there is no way back (1) useful mainly to connect areas/realms (1) makes game more interesting (3) magic portals act as reference points (1) |
| 12 | length of description has influence (1) has no influence (1) rooms in vicinity does influence (1) does not influence (1) depends on number of objects (2) description (8) number of people (3-4) size of objects (1) detail in description (3) all rooms have equal size (0) |
| 13 | same distance as normal exit (1) it is different (6-7) leads to another level or mode (1) distance is the number of rooms (2) number of commands (1) distance depends on the description (2) several commands for exit are irritating (1) a well-known area appears to be closer (1) |
| 14 | after teleport a new map starts (2-3) a new area starts (4) I notice overlapping areas (6) Euclid valid only locally (3) the areas are not directly related (1) |
| 15 | the MUD is a plane (1-2) plane with overlapping areas (1) a space (1) a space with levels (5) a landscape (1) others (pancake, monopoly board,..) (3) |
| 16 | I collaborate often (3) sometimes (1-2) never (5) |
| 17 | give only precise directions (8) give only imprecise directions (2) imprecise directions in a MUD are possible (8) mixed descriptions are most useful (1) |
| 18 | often use transports (1) sometimes use (3) (almost) never use (5) I would get disoriented (1) use only when very far (1) I think I could forget path (4-5) I will never forget a path (3) transports are too slow (5) can be faster when laggy (1) I use them only when I do not know the way (2) I do not learn the normal route when using transport (1) transports have to few destinations (1) transports cost money (1) |
| 19 | improvements for transports are not necessary / possible (3) should go to more (important) places (4) more helpful when walking costs exhaustion points (1) should cover larger range (1) should be instantly available (no waiting) (1) should remember / learn places (2) |
| 20 | a guided tour is a good idea (5) is no good idea (3) it is a nice idea (1) should use regular exits (3) should use teleport (1) maps would help people much better (2) restricted access only for newbies (1) |
| Wizard-only questions | |
|---|---|
| W1 | I coded no area (1) I coded an area / am planning (5) < 50 rooms (2) 50-100 rooms (1) > 500 rooms (2) (large part) is not public (4-5) |
| W2 | interesting (6) few standard exits (1) funny (3) there should be a quest (3) easy to find (1) hard to find (1) mappable (2) "rich" area (3) hidden treasures (1) |
| W3 | I design on a map first (6) I use no map (1) |
| W4 | I use a grid (5-6) I use a different scheme (1) |
| W5 | avoid overlaps and holes (7) use several rooms for large places (2) sizes and distances are in the description (3) |
| W6 | exits are connections (3) I use rooms that represent connections ("glue") (4) |
| W7 | no rooms with several positions (5) I do use them (2) implementing such rooms is a problem (2) |
| W8 | use non-standard exits when it fits (2) to confuse players (1-2) to make area more interesting (2) to separate "areas" (3) for one-way exits (1) enactment is necessary (2) |
| W9 | design no transports (2) design transports (4-5) use them as non-standard exits (1) use them as shortcut (2) provide them as a teleport (1) only in the main area (1) there should be several lines or tours (1) |
| W10 | Guidelines make sense (1) no need for guidelines (3) are useful only as advice (1) evaluation team is better (3) could prevent unnecessary work (1) |
After this first evaluation it was tried to informally correlate questions in this interview series. The questions do not support this type of evaluation well however and also the low number of interviews makes this approach somewhat questionable. Five groups of questions were tried to correlate:
Another drawback is the time needed to do an interview. First an interview partner has to be found and then a time where both the interviewer and the interviewee have enough time and where the network-lag and game-lag is not too high has to be found.
Therefore the number of interview that can be performed in a reasonable time is much lower than the number of questionnaires that can be expected to be filled in. Interview partners are likely to be from approximately the same time-zone, whereas questionnaires reaches MUD participants independently from any time-zones.
The exchange of questions and answers can sometimes be severely influenced by lag. Because of that the interviews may be only one of several discussions performed at the same time by the interviewee. This could lead to a less concentrated interview. The degree of detail of an interview and the possibility to clarify unclear points makes the interview a very useful instrument for most surveys in text-based virtual environments.