So, what we're usually dealing with in retrieval is our documents. So we have different types of documents, which can be textural, they can be images, still or moving, they can contain audio, so like music or speech or other sounds, and they

can be mixed. There may be other types as well, but these are the main ones. Of course, we always have the question, which was already posed, what is a document, whether an animal in the zoo, which we're trying to describe, is a document or not, but we won't

go into that. How can we find what we want or what we need? Of course, these may be two completely different things, what we want and what we need. Of course, one way to look at it is we look at the text.

So we search full text, which is of course what we do every day when we search on Google. But there may be some problems with full text searching. If we take the example of looking for journal articles, we may have some problems.

First of all, and of course, much of what I'm going to discuss here has been mentioned here or there during the presentation during this week, so we're trying to bring these

things together. One of the problems is the terminology. So you have the problem of synonyms, so words that mean the same thing, like car and automobile mean the same thing, and we don't know which word the author of the text used.

Of course, the same thing, but in a different context, we have again the same concept, but in different languages. So in my language, car is called auto, and again, if we search for car, we won't find

documents that contain the word auto and vice versa, or you shouldn't assume that you will find it. There may be some systems that do have some mechanisms for that. And of course, the next thing is we may want to search for a particular characteristic.

For example, yesterday you had the two cases where you wanted to search for documents in a certain time period. So you may want to limit the search to a certain time period, or you may want to limit

the search to a certain journal, or you may know the so-called known item search. So you know of an article, you know perhaps the title and the author, and you may want to state that what you're looking for is an article with such and such title and

with this and this author. So you may want to have a way of expressing that, and of course, searching more precisely. And of course, this was only text. What about images and sound?

Of course, you may say if you have a speech, audio speech, you may transcribe that into text, and then again you're dealing with full text search. But you may also not have a, you may have music, you may have images, and of course

you can, you have systems that search by providing a picture or by providing sounds, you know, things like, I don't know, sound hound and so on, which work, but of course if you want to sing something, not everybody sings the same way.

You know, some people can't really sing and some people can't really draw, and you can't really search on that, and you have to know what you're searching for, because if you don't know what you're searching for, you're not going to get anything. So there are problems. So the solution is metadata. That is to say some structured information about the document.

And this is basically what Polona mentioned on Monday. Metadata can only involve having different boxes for different characteristics, so splitting information into different boxes, so different fields, let's call them, and so that's the

standard structure, so you have the elements. Then you may also have rules about how you're going to describe things, and then you may also have values, a list of values for which you will choose when describing things.

And of course you may also have the encoding format, so how things will be presented in the computer. Now, if we talk about controlled vocabularies, these are lists of values.

So we come to this point here, values. So you have lists that are prepared that contain possible values or approved values to be used for describing and also for searching.

That's the important part here. What you want to do is have the same terminology when describing things and when searching for things so that we can have a match when we compare the query to the description of the document.

So this may be things like codes for languages like EN for English or FR for French. This may be genre terms. And of course, one important thing when describing documents is the contents of the documents, so the topic of the document, the subject.

And that's probably one of the main problems in information science is how to describe the content of a document. Because all the other things, you know, the author, okay, you can write author's things in different ways, you know, last name first or first name first or whatever,

or there may be different spellings, different variations. But in terms of subjects, you first have to say, okay, what is this document about? And of course, no two persons may have the same idea of what a document is about.

And of course, some of that is done when done with automatic indexing. You get the...and the partial matching that you talked about yesterday, you're kind of trying to deal with this. So we have things like subject terms and classification notation basically.

So subject vocabularies are in their minimum, so at least a list of terms to be used, a list of values to be used, again, for describing and ideally also for searching.

They may also contain things like controls of synonyms. So you have different names for the same concept. They may also contain hierarchical relationships, like for example, a car is a vehicle, so you have the hierarchical relationship, and you may also have other relationships.

And in the extreme, you come to the ontologies where you have typed relationships. And these relationships, of course, enable us to explore the terminology of a domain

and the domain itself, the structure of the domain. So this is an example from iridesimal classification, which is one of the classification systems that is used. It's not the only one.

Where notation is, as you can see, is by numbers. It uses numbers. It doesn't use words, really. It may use some lettering, and it might use punctuation and so on, but it usually does not use, the classification systems usually do not use words. Of course, nobody can remember that by heart,

so you do have an explanation of what this means. And for example, if you look at this second line here, you have medicine, and within that you have specific diseases, and then disease of endocrine system and diabetes mellitus. So you have a hierarchy of how things are arranged

in this specific portion of knowledge. And of course, that is reflected in the notation here. Every number here, or a combination of numbers, means one of these constants.

Of course, you may also have things described with words. So this is taken from AgroVox, which is a vocabulary in the field of agriculture.

As you can see, there's a preferred term for a certain concept, which is forest products derivation. And then you have brother concept. In this case, you only have one brother concept, which is forestry operations. And then you have network concepts, copy, sync, and logging. So it's, again, you have the hierarchy.

And you also have something which is the equivalent of this term, at least I guess it's equivalent of this term in Turkish. Additionally, as you can see here, you also have URI. So we're going back to what you talked about

when you discussed semantic web. So identification is important. Identification of a certain concept, in this case. And of course, what we can do here is we look at the concept.

We get acquainted with, as a person, as a searcher, we get acquainted with the brother concepts, with the network concepts, so you get acquainted with the domain, the structure of the domain. And then you can use whichever concepts

you think are relevant for your query to query in a certain system. Of course, if the system uses this particular vocabulary. This is just another example. This is, you can't really see this,

what it says here, but it's St. Petersburg. And there's lots of information. Also, you have type of place and so on, so you have really quite a lot of information about the concept. You have the definition of the concept and so on.

Now we come to the mechanism called authority control, which is most often named authority control. So what it does is it clusters all names, all appellations, usually when we talk about

authority control, the first association is authority control of persons or companies, corporations. So it clusters all of the names that a person may use, or all the names that are relevant in a certain context that a person may use,

and usually there's one name that's preferred. So the one name that's to be used in the system for describing and also for searching.

So for example, if we have William Shakespeare, you can see that there are different ways of writing his name, there are different scripts, different languages, different way of describing the person. Of course, one thing, you may think of Shakespeare

as an author, but Shakespeare can also be a subject of a, for example, of a book. There may be a biography of Shakespeare, and then he's the subject, but it's still the same concept, it's still the same person. So of course, people and companies and so on,

and you may also have authority control of subjects, you may have authority control of words, so on. So they can use their names randomly, or their names can be used randomly on documents.

So for example, William Shakespeare always, if you look at historic records, you see that he signed his name differently in different times, but on the book, you usually see William Shakespeare, but if you go to China, you won't see William Shakespeare, you'll see the Chinese characters. So it's not that he named his,

or somebody named himself, named him, his parents didn't name him using Chinese characters, but in China, they use Chinese characters, so that's why that name is also used. Of course, names can change over time. Probably most noticeably, people who get married, and then they get either double-barrel surnames,

or they take the surname of their partner. Of course, some people also use different names in different circumstances. For example, Lewis Carroll used his pseudonym to write Alice in Wonderland, when he wrote about mathematics, he used his real name.

So, authority control is a mechanism that enables searching when done properly. It enables searching on any form of name, and getting words by, for example, an author, regardless of the form of the name, that's the point.

You bring together all the words that were, regardless of how they were, under what name they were described. So, organization of information has a long tradition,

of course, prior to internet and so on. And libraries were, for a long time, the main player, basically the only player. Of course, lately they kind of lost that place. But, of course, they're still organizing information,

and even in, the need for organization of information is even more important in the digital environment. Of course, traditional methods like cataloging,

so in-depth cataloging may be used, but it's not really feasible to catalog everything on the web, for example. So, sometimes simplified cataloging is used, most noticeably using things like doubling core,

for example, when describing documents in a digital library. And, of course, and this was also talked about yesterday, you may have automatic processing. So, you may have automatic indexing, automatic classification, and so on.

And, another thing that is done is, of course, similarity, which we see basically every day, when systems take information about user behavior, so about personal user behavior of the particular user,

or slash behavior of other users, so gathering wisdom from the crowd, so to speak. So, organization of information enables us to locate and use information,

and is, in my opinion, one of the central concepts of information science. Now, other people in this room may not agree about that, but I think it's what is the unifying concept

in information science. So, the role of information science is to develop mechanisms and tools, such as authority control, such as control vocabulary, such as doubling core, such as information retrieval systems, and so on,

to evaluate them. But, of course, also, and we shouldn't forget that, also study user needs and wants, and study user behavior. And, with this, I will finish my brief presentation here,

and I'll give the microphone to my colleague, Polona. Can everybody hear me? I will get away from the screen, don't worry. Until now, you mostly heard about the things

that we do, the things that we put in our collections, the things that we do with our things, you know, describing, talking about them, searching, and so on. My turn now is to tell you something

about the people who are involved in that. Usually, we call them users. Sometimes, we call them just individuals, just people, or groups, or something. But, anyway, this portion of lectures will be about what it says here.

Information needs and information behavior. Have you ever come across these terms? Both of them? Okay, how would you define an information need? I'm posing silly questions, I know, I do it all the time. What is an information need?

Something that is, wants to know, or to discover. Okay, yeah, yeah, and actually, since we are human beings, and since by nature, we are curious, we do it all the time, or we have these needs all the time. And when we try to follow them,

when we try to fulfill them, which we often do, it's this, there's the second term that I'm using here. We actually, information behave, if I may say so. So anyway, there is no doubt about the fact

that people need information. I guess we all agree about that. What could we do without it? What would we do if we didn't have a chance to get to the information when we need it? How many times a day do you not know something and try to get that, try to find that out?

Is it possible to count that? I mean, in your daily life, I don't mean the library, I mean it every day in the life. How many times? For example, what's the time? Is that an information need? Of course it is. But we actually sort of like distinguish

between the tiny and the more substantial information needs, and for the more substantial information needs, we of course divide systems such as libraries, such as databases, and so on. But we mainly need information to come to some solution,

to solve some problem, be it a small problem or a big problem. Usually problems are, or part of the problems are solvable with information. So information needs are motivators for, oh sorry about that, for motivators for information seeking, I place that cartoon there,

I don't know if it's visible in the middle, it's a girl doing a Sudoku, you know? Anyone doing a Sudoku? When you're doing that, you are, great. Yeah, it's a good mental exercise, you know? It will keep you young forever, so go do Sudoku.

But it's full of information needs. When you're solving such things, and others of course, you always, you constantly need information, and you constantly of course need to think to solve problems. So the problem with information needs, on the one hand, is that they are difficult to define,

because you have many different notions of information. If I ask you what is information, I'm sure that everybody or at least a few definitions will come up. They are not a basic human need, you know? We all from biology, we know what basic human needs are, and what they are for, they help us survive.

But information need is not such a need. It is inspired by other motives, for example, to eat. Do you ever have information needs related to your eating of needs? Yes, of course. When is dinner?

Where is the place I can eat? Which food do they serve? And so on. To have shelter, and so on. They are multidimensional. They are linked to other gratifications. They are linked, for example, to the notion of reducing uncertainty. For example, you have come to a lecture to a new professor, and you were chatting

with the colleague, and you didn't hear when the professor said, how many points do you need to score to pass the exam? You didn't hear that. How do you feel, and what do you do?

What would you do? You either ask professor, or if you're maybe a little bit shy, you go and ask, turn to a colleague and ask them, did they hear how many points to pass? And until you find out, your feelings are not as good as when you do find out.

This is uncertainty. Not knowing leaves you in a state of not a very pleasant, pleasant feeling. They are influenced by many factors. I mean, the solving of them, of the information needs. By your experience, by the situation you're in,

by the system that you are using. The system can also be people around you. It's an unstructured system, an ordered one, but it's still some sort of a system. The information needs are problematic also because they are very difficult to observe for a researcher such as myself. It's very difficult to know at this moment

what you don't know, what information needs you might have. How am I supposed to know that? Well, I have to ask you. And the moment I ask you, you begin describing your information needs, and it becomes information behavior. So basically what we do when we study information needs

is we investigate demands. We investigate what people ask. And as my colleague rightly said before me, the demands are not always what people need. For example, I have examples further on. In libraries they often get questions such as,

do you have any books on human beings? Well, yes we do, you know. But can you be any more precise? So what you demand is not always what you need. Not even always what you want, because sometimes you don't have the words,

you don't know how to ask, and so on. Anybody who has heard anything about a reference interview in libraries knows that a reference interview is aimed to solve such problems, to help people who have difficulties in asking questions.

So according to Wilson, as you see, researchers mostly study information behavior, not the needs themselves, because they are in the heads of individuals. I can't know what you don't know. How am I supposed to know? I have to ask you, or I have to wait for you to tell me. So the key to understanding information needs

often lies in understanding of problems and situations that cause information needs. For example, an experienced librarian who knows how research work is done, who has experience in searching, who has maybe done a PhD, is probably a very valuable collaborator to a researcher,

because he can somehow sympathize, or he knows certain situations that this researcher may find himself in. So it's important in order to be aware of the information needs to have contact with the users and to do some sort of user research.

It doesn't have to be formal research. It just has to be some sort of finding out what the users are faced with, what their usual day looks like, what their work is, and so on, what their situations are. So like I said, users don't always want,

don't always know what they really need. You know, needs are internally experienced. It's a necessity, and sometimes we even feel sort of like some moral obligation to help people with their needs, you know? We often hear that it's the mission of librarians,

of the information specialists to help people find what they want. So sort of we have a moral obligation, sort of like we are here to help people. Does this remind anyone of doctors, of the people who are here to help people get healthy

or stay healthy? And wants, on the other hand, are what we think we need. You know, it's often when we hear, let's say parents talk about children, you know, to say we cannot let them have everything that they want because it's not good for them,

or because it will harm them, because it's not the same, you know? What they want and what is good, or what they, what they will, what will really help them. So users sometimes even don't know that they need something. Have you ever heard of sleeping information needs, dormant?

I'll give you an example. Anyone here that likes chocolate? Okay, imagine riding a bus and standing next to two people who are talking among themselves, and they're talking about,

or one of them is talking about the fact that she just read in a newspaper that they found that chocolate causes cancer. Hmm. Would you, okay, you can decide not to care, yeah?

You can decide to doubt this information and to check it and everything, but would you want to know that? Did you know that you needed to know that before you heard it? There you are. There's your sleeping information need. You had a, let's say, an interest in that,

but you didn't know it. And the needs, in order to be fulfilled, need to be expressed. Unexpressed information needs cannot, we cannot do anything about them. But the actual needs can stay unexpressed if users see them as unrealistic,

if they don't feel that they can be fulfilled. For example, I'm not gonna ask this librarian because she's incompetent or because she doesn't know anything or because she doesn't like me. There are many reasons why users don't express their information needs.

Which brings us to information behavior. Individuals are motivated to find information to satisfy their information needs. And as Bates defined, the information behavior are the many ways in which human beings interact with information. In particular, the ways in which people seek

and utilize information. So the searching and seeking part is by far the largest here. But it's also a scientific discipline. It's a sub-discipline of LIS that engages in research conducted in order to understand the human relationship to information.

So anything that's researched to understand this relation between the users and the information is called research of information or can be considered research of information behavior. Wilson produced this model that you will probably see many times. Especially if you are going to deal more

in more detail with the area of information behavior. He tried to present graphically how it is with the information behavior as a whole and the information seeking behavior within the information behavior. It's that part that people try to find information in any source.

And within that you have a big bubble of the information search behavior. That's the searching in digital environment. And it's also a very large part. He of course left out some of the forms, some of the types of information behavior. For example, the information use which is also considered a big part of the area.

But he also defined some types of information behavior. The information behavior can also be unpurposeful, done without or with little purpose. Contrary to purposeful directed searching

which we express when we are actively looking for something. We actually also express other types such as browsing. Do you ever browse? How much purpose of finding something do you have when you browse?

It's probably very varied. Sometimes you even have a purpose and this is why the systems need to offer you this solution to browse. But sometimes you just do it to pass the time or to just see what's there or something like that. It's not only done in libraries or on the internet. It's also done in shops.

And this brings us to the third thing here that's sort of relevant or connected to this. It's serendipity. Have you ever heard of this term? Serendipity? Some have, some haven't. Serendipity means finding something without looking for it.

Remember the chocolate example. You had a need that you weren't aware of but when you got the information you realized that you were interested in this information. This is serendipity. And a lot of information needs revolve around this thing. Which brings us to the third group of notions or concepts relevant within the information behavior area.

It's relevance and pertinence and salience. The relevance and pertinence are quite closely connected. You probably came across the term relevance, right? Not only in the context of information systems but also in the context of information needs.

What is relevant information? What is relevant for me might not be relevant for you. Of course. And what's relevant for me might not be relevant for you but in general, what information

would you consider relevant? Useful, okay. Is with our needs. That fits your needs, yes, of course. Would anyone define it as information on the right topic or about the right things?

Because that's the usual notion of relevance or the more strict notion of relevance. Here again, you probably see the term aboutness. The aboutness or the topicality, how it's often called is also the thing connected to relevance. When you search in a database

and you find some, let's say hits, are they all relevant? You do a search and there are the hits. Are they relevant? Not all of them or yes? Because from the point of view of the information system they are because otherwise they wouldn't be found.

So you see there is a difference between the relevance as perceived by the system and between the relevance, or sometimes we call it pertinence as perceived by the user. As Belkin and Vickery have said, relevance is the answer to a question. And if the question is posed wrongly or inaccurately,

it may not be the answer, whereas the pertinence is the answer to the information. So you're very right. It's what the users are interested in to get the information that responds to their need, not to the question. For example, if I say,

when is lunch and you heard me wrongly and you will tell me it's at seven in the morning, the relevant information because it's an answer to my question but it doesn't answer my need. I need different information to be pertinent. So most users want pertinent information.

This brings me to the third one, salience. Salience is a different notion. It actually means what stands out in some way. What is vivid, noticeable, obvious, unusual and so on. And it affects our attention, it affects how we remember things

or if we remember them at all and how we recall them. But the problem is that what is salient, what stands out is not necessarily relevant or pertinent. Maybe it just stands out. And what is relevant or pertinent, if it's not salient, we can miss it.

It can go unnoticed. So it's also a very relevant notion. Other related concepts that also go within the context of information behavior is what Kay's called selective exposure, selective exposure meaning. Sometimes we find ourselves in a situation where we don't want information,

where there is too much information or that we cannot cope with it, that we're not capable of dealing with information. It's called information poverty. It's called knowledge gaps. But the other two, the selective exposure, the avoiding information and the information overload

are the types of behavior that we find when there is too much information. Has anyone ever found themselves in situation of too much information? Yes, because you all go to school and it's a very common situation, especially in educational and work related areas.

And another one here, not related to those is entertainment. Entertainment was for a long time considered not a serious area in terms of information provision, in terms of studying information behavior.

But in fact, research shows that there are not so many differences between serious information behavior, pertaining to education or work or something and educational information behavior. Another big area that I'm not gonna go into

in a very great detail is collaborative information behavior. And I'm sure that people that come after me will touch a little bit more on that. It's a relatively new area. It actually investigates or deals with situations where people work in groups,

where they collaborate to achieve higher quality goals, they work more easily, they divide labor and so on. They usually work to pursue a common goal that couldn't be reached if they worked as individuals. So it's usually connected with complex tasks,

extra cognitive load, and in these expertise of a group of multiple individuals, it's difficult, it's demanding and it's not necessarily successful. But it's in terms of behavior, a bit special. There are different types, it can just be coordinated,

it can be cooperative. So you need me to do my thing before you can do your thing and you can't do your thing until I do mine. It's not the same with coordinated, it's just that we do things together. And co-constructive is we do something together and we cannot work without each other.

I decided to mention some relevant theories within the information behavior. One is the TZIF law, TZIF principle of least effort. It's not the same TZIF law as the one in bibliometry, but it's the same surname.

Anyway, have you ever heard of the principle of least effort? We tend to do things with as little effort as possible to achieve a goal. Does that sound familiar to anyone? Of course it does, we're human. This is the way humans operate.

We all do that in a way. But within information behavior world, it means that searchers will minimize their effort in searching even if it means that they have to accept lower quality information or less information. People who search on Google are usually happy with,

what would you say, how many hits do they look at? The first page maybe? Oh, you are very optimistic, yes. If you look at the first page, then you are probably a librarian or studying LIS because it's very rarely to go over three hits.

And most often, we're talking about average individuals here. So yeah, yeah, the first page. If you go to a second page, then you're doomed. You're a librarian, no doubt about that. So that's the principle of least effort.

The problem is, I mean, the fact is that we have to count on that when we deal with information systems of any kind. Then is the use and gratification and game and fun theory, which means that people actively participate in exposure to various media because this is related to fulfillment of their needs. And they treat the input of fun

to satisfy their emotional needs. It's about this dual, the pleasure and pain. We look for pleasure and avoid pain, but also connect these two because it's actually the question of what is pleasant and is different for every individual. Now, this is being used by the mass media.

And the second part is the sense making and uncertainty reduction, which I already mentioned. It means that we either internally create information to understand the world around us, to predict things, to make sense, to feel well, to feel in control of our world.

And this is information to reduce uncertainty related to solving of problems. Also need to say just a little bit about who are we dealing with when working with users. In general, it's individuals and organizations

who have certain needs for information, education, culture, entertainment. And we need to say that typologies are many and they vary greatly. Some people or some typologies look at the people's age. They divide, for example, children and youngsters and elderly and so on.

Others look at context, demographics, experience, and so on. Other typologies look at individual characteristics that influence somebody's behavior. Such are the theories, for example, of information styles. Are you a collector? Are you a hunter? There are quite a few of these theories.

Others divide actual users from potential users or even from future users. What will our users be like in, I don't know, 30 years? Because if you consider the speed of development that has been until now,

what were the users like 30 years ago? So how can we predict our services and systems for the next, let's say, 20, 30 years? And of course, we have some notion of special needs. Here is one division by age. Children, teenagers, young adults, says Iffla.

Young adults are people who are close to finishing their secondary school but before getting employment, for example, or around that age. And they can be further divided, as you can see. Adults, which are actually a very heterogeneous group,

are all adults the same? No, of course, they are not. Some divisions even say that the people who are no longer working are still adults and consider them as one group. And some people talk about the elderly as a special group. But the age, per se, is not an absolute category.

Within the age, you can have division by roles. You can have somebody being a mother and a professor at the same time and behaves differently in different roles. There are different experiences, different contexts, and so on. And when we come to special needs,

I just need to show you very quickly the division, I follow here the Ifflas, sort of like system. We divide the users according to the special needs that they have, to the visually impaired, to deaf or hearing impaired, to persons with reading difficulties.

Have you ever heard the term print disabled? It's used heavily by Iffla to denote people who have any kind of difficulties with printed materials. Whether they cannot read it due to their poor sight or old age.

They cannot hold a book because it's too heavy, and so on, which is why you have those little shelves in libraries maybe, maybe you notice them, the things that you can pull out. They are there so that you can lay down a heavy book or a bunch of books. Many have persons with physical disabilities,

cognitively disabled. You have people in long-term care facilities. Why are they a special group? Because they can't come to the library. And the library needs to think about ways of getting to them. People in special circumstances, and of course,

a special circumstance, of course, are the multicultural communities. Why are they special needs groups in special circumstances? Because they need adaptations of services, of materials, of premises, and so on. We need to think about how to reach these people. Why do we need to know user types?

Because we become aware of the rights of all, because we all have human rights, we all have the right to information, we all have the right to education. It's just a matter of how we execute or how we manage to execute this right. We better understand the needs of our actual users, and we know who might our potential users be

and how to approach them. We perform our activities and services better, and we plan and design services and systems better in order to come closer to people. And how do we find these things out? We research, we have contacts with users,

and we educate and train ourselves, because it's about constant development of also the staff, the information professionals. Okay, here are some orientation questions, same as last time. I'm not gonna test you on these, it's just so that you can have a look and think about things.