Holla_Waldfee part 1
===

### General Questions ###

investigator: Okay, now recording is in progress and we can continue with the next page. Are you on next page? 

Holla_Waldfee: Right? Just click continue. Okay. I'm on session one questionnaire, general questions regarding usage of digital libraries. 


##### GQ1 #####


investigator: Okay. Perfect. So, the first question would be which tasks do you usually use a digital library for?

Please tick all answers, which apply and complete your own tasks. Please give a short oral example of the task that you are ticking. So, for example, you could say "I'm ticking person search because I like to keep track of myself". So yeah. What tasks do you usually do with a little library? 

Holla_Waldfee: Okay. So just to clarify, but this doesn't include, for example, in case of the first answer person search, if I Google someone and then get directed to digital libraries that doesn't count as usage of digital libraries, right. Or would that be using digital libraries as well? 

investigator: Depends on your own definition, if you think you are using a digital library by doing so, then yeah.

Holla_Waldfee: Okay. Then let's start.

investigator: Okay. 

Holla_Waldfee: Well, in that case, I would say, yeah, that "person search" definitely. Because, well, especially if I look for people where I don't know exactly in what scientific field they contribute I wouldn't know what digital library to search for that person. So, I usually go the route of Googling that person and then checking what did libraries Google offers me. So, for an exploratory kind of search thing, I would definitely use them. Yes, "paper search", obviously. That's the main reason I use digital libraries especially those that are more well focused on a special scientific field and then offer detailed possibilities to filter the results.

That's pretty useful for like some, getting an overview. I have used digital libraries for venue search. But I do that rarely, so I don't usually do that. I do use it to get BibTeX data though, or that I used to do that a lot. Now I have some browser plugins, but still do it too but automatically and in some cases.

"Get full text papers", do I do that? That's a tricky question. I'm not sure because I usually get them from this table download. So, yeah, but it's tricky, but I would still say yes, because if I need a PDF quickly and I know the title of a paper, I just look for the paper and then get the PDF file.

And I do use it to study relations study venue, author relationships. Yes. If that includes citations. Yes. Otherwise, I tick the box and say well, in the optional further answers I would add call it like citation chains or something where I look for references in one paper, then check like the relevant papers from these references for other relevant literature and so on.

The citation ... can't write and speak at the same time. Okay. Next question? 

investigator: Okay. Perfect. Is this everything, or is there any other tasks you also do with digital libraries that has not been mentioned here?

Holla_Waldfee: Well, more abstract tasks, I guess. Like for example checking for keywords or phrases that I would include in future publications. That I'm planning. Just to check whether there has been research in that field or how much research has been done or what keywords related research usually applies to research that's close to the research I'm planning on doing. 

investigator: So, something like keyword search?

Holla_Waldfee: Yeah. I mean, that's not in the list of options. 

investigator: Yes!

Holla_Waldfee: So, yeah. So, what else? No, I guess that's it. Yeah. 


##### GQ2 #####


investigator: Okay. Perfect. Then I think we can continue to the next one. 

Which system or digital library do you usually use to solve these tasks? Please tick all answers, which apply and name others, which also apply, but are not given here. And please give a short oral description, why you like or why you use the system? 

Holla_Waldfee: Okay. first note on how I answered this I understand it as if I use one of these in some way to fulfil one of the aforementioned tasks, I'll check it, okay? 

investigator: Yes. 

Holla_Waldfee: So, I have used ACM digital library quite a few times, although I rarely Go directly to like the ACM digital library, but get directed there most of the time. Same with Bibsonomy. web of science, yeah, rarely, but I guess I occasionally use it. Not usually. Clarivate... That doesn't bring any bells for me. Dunno. So, I don't know. I check it because I have used web of science. Rarely. Ah, dblp all the time, although that's probably rather due to my field of science than... scopus. I used to use this the most actually when I was in communication sciences. That was my go-to digital library.

Google scholar. I actually never used Google scholar because I actually prefer semantics scholar and I built my own crawler for semantic scholar that actually improves on the user interface quite a lot. So, I use that. Springer link is difficult because many publications are on Springer link, but I get referred to it. So, I don't know if this counts as usage?

investigator: No, also by the search for ...

Holla_Waldfee: I don't use the search function as spring. Okay. So, if that's the...?

investigator: No, no, no, no. I was asking if you were searching for papers and then the result page led you to springer link. 

Holla_Waldfee: Yeah, for example. Yeah, exactly. Stuff like that.

Yeah. For example, if I get redirected there from, I don't know, semantics scholar or something. Same with... well research gate is... I've never gone to research gate and used the search functions there, but I get redirected to research gate all the time for PDFs or so on. So, I'm quite familiar with research gate as well.

And well, usual search engine. No, actually, I never used it for scientific research. I think I've never Googled something where I wanted to find literature or information, except for maybe information on authors or like conferences or something like that. But I wouldn't even check that here. 

investigator: Okay. you skipped dimensions. Do you have anything to say about that or...? 

Holla_Waldfee: Never heard of it. 

investigator: Okay, perfect. So, do you have any other digital library or system to include here? 

Holla_Waldfee: Nothing. Nothing I use professionally. I'd say no. 


### TASK 1 ###


investigator: Okay. Then I think we can continue to the next page. 

Holla_Waldfee: Okay. 

investigator: And we can start with the first task consider the following task. Find two experts on a topic of your liking. Here example topics could be for example, "domain specific query languages" or "hashing functions". But they should be from the broader area of computer and information science. So, you can choose whatever topic you like. And you do not really have to do this task, but we are going to talk about how you would do this task.


##### TASK 1.1 #####


So, my first question is what is your chosen topic? 

Holla_Waldfee: Okay. That's a good question. I feel really uninspired right now. So, I'd just say, I dunno, what could be a good topic? That's it's difficult! And it doesn't have to be related to my research, right?

investigator: No. 

Holla_Waldfee: I don't know.

investigator: But it can be.

Holla_Waldfee: Let's say "image recognition". 

investigator: Okay, perfect.

Holla_Waldfee: Not related to my research, but just like something I'm interested in right now, so.


##### TASK 1.2 #####


investigator: Okay. How familiar are you at this topic?

Holla_Waldfee: Mm, in that case I'd give it a two out of 10 on a professional scale and five out of 10 on a common knowledge scale. 


##### TASK 1.3 #####


investigator: Okay, nice. And how would you define an expert? 

Holla_Waldfee: Someone who is regularly cited in other people's definitions on that topic or in other people's introductionary sections of publications, for example. That's usually where I'd start on finding like the people with the most expertise in topics.

investigator: Okay.

Holla_Waldfee: In other words, if other expert deem that person an expert, that's how I'd say that someone is definitely an expert. 


##### TASK 1.4 #####


investigator: Okay. That's nice. Because now with question four, you are asked to describe how you would solve this task. How would you find two experts on a topic of "image recognition"? 

Holla_Waldfee: I think I already gave away a little of that.

So, I'd search for publications in that field probably by keyword search and then probably first check for publications that are referenced a lot. That have a lot of citations. Then I'd also look for like recent publications and see whether like the authors of the most cited publications so still published today because that's another criterium for me. If someone used to be an expert in a computer science field a few years ago, that doesn't mean anything necessarily for today. So, in a scientific field, I don't know much about as in this case I'd be careful.

So, I'd probably start with a keyword search, then narrow it down to like the last few years and then check for authors, check, whether there's like clear first authorships and or authorships that I can identify multiple times. And then I'd also check in, for example, abstracts and related works, if these papers are quoted in a way that I think that the expert ...

well, I'm looking for the person with the most expertise first. Right? So, or like the people with the most expertise. So, this would probably be my way to identify people that are well, the best-known expert in that field.

investigator: Okay. where would you do the keyword search that you begin? 

Holla_Waldfee: Mm. Yeah. Well, if we're talking about computer science, I probably start with, well, no, I'm not, not gonna start, I'll I think I'll exclusively check dblp. Because well, in contrast to, I don't know, Google scholar, semantic scholar, dblp has much less literature, but also much higher quality of literature. So, I think all of it is peer reviewed or most of it's peer reviewed. I don't have any weird publications that are not weird publications. So, I think I have the least noise to deal with in the search results. So especially if I don't have much expertise in the field to skip over results quickly myself, by my own judgment.

I would really trust dblp in providing me with good results here. 

investigator: Yeah. Okay. 

Holla_Waldfee: And relevant literature. 

investigator: Where would you check the references of papers that result your keyword search? 

Holla_Waldfee: Oh, well, do you mean like...

investigator: In the number of citations? 

Holla_Waldfee: Okay. I think that's returned by the dblp API, isn't it? I'm not quite sure. At least I got it from, I have a dblp crawler that returns me that. So, I think that's provided by the dblp. 

investigator: Okay. So, you would not use the dblp webpage, but you would use the API. They provide to check this information?

Holla_Waldfee: I would use dblb, but I'm not quite sure, actually, if it is... let me check. I have to check that quickly because I know. No, it actually doesn't show the number of citations here. That's weird. That's actually where I'm... I have to say I'm in an awkward situation because I've searched for lots of literature for a recent literature study, but that was all done in a self-built crawler which returns to me the number of citations.

But I think that, well, I spoke very confident that I definitely search for it on dblp. But now that I'm here and just have done a little quick exploratory search, I would probably not know whether... well, the number of publications here is manageable, right? So, I think I still go for it.

If I don't find what I'm looking for, I look for a venue where I can find the number of citations. If I don't have the feeling that what I'm finding is the most relevant literature field. 

investigator: Okay. up next, you said that you would want to check authors. Would you check authors in dblp or would you switch to another system or...?

Holla_Waldfee: Actually, I would probably still check dblp and then maybe... well, I previously said I wouldn't use regular search engines, but if I'm especially interested in a person and want to check what that person has done. And for example, want to check the personal websites or an Institute's website or something.

I would probably still use Google for that, just to see where that person well, has put up, like, for example, a private website. You won't find that on dblp, right? So, if I'm really interested in a person, but there wasn't the original scenario, I think I'd use the search engine for that. 

investigator: And you mentioned that you would check authors' abstracts. How do you go from finding an author that might be interesting to checking their abstracts? Do you go to the personal homepage?

Holla_Waldfee: The other way around!? 

investigator: Oh!

Holla_Waldfee: I go to papers that I find interesting. 

investigator: Okay. 

Holla_Waldfee: Then check well, not abstracts. I think I meant to say introductions. So, I think that introductions of papers that exactly fit the scientific field I'm after. A good source to find really relevant milestone literature, so to say in the given field. So, I'd probably start with a good paper. A paper that I find that it hits the nail on the head, then check the introduction, check what four or five references we have in the introduction that kind of open and define the topic. and then go there and see what I like best. 

investigator: Okay. would you get a full text of the papers that you need to get the introductions from dblp or would you get them somewhere else? 

Holla_Waldfee: Yeah, probably dblp and, or but dblp or just like the my Zotero importer plugin, because it's just a simple mouse click and then I have it all in Zotero in a structured way. That's probably my go to way. Because I actually, yeah, that's what I would do probably because I wouldn't... I would first get like four to 10 papers that I'd like to check and then check them in Zotero because then I can well better see what's relevant and what's not. 

investigator: Okay. And now that you have some candidates for experts from those introductions of papers what's your decision criterion for deciding whom to pick? 

Holla_Waldfee: Depends on why I need the expert. If I just want to, I don't know. I rarely actually search for experts because I need an expert. I mostly search for experts because I'm interested in the literature they publish. So, yeah, I wouldn't choose one expert, but an expert, depending on how much literature I want to read.

And if I want to interview an expert, I'd probably take into consideration who has published papers I have questions about, or was really interested in, especially, or, and that's probably the argument that's weighing more in total is like who's accessible. Who seems to be a person that I'd like to talk to and who lives nearby or who has like, I don't know. Who do I have connections to, for example. If there are already connections that I can leverage to get in contact, I'd probably go that route. 

investigator: Okay, nice. Do you want to add anything else to this question and task, or should we continue with the next one? 

Holla_Waldfee: No, I think we're fine. 


### TASK 2 ###


investigator: Okay. Then you can click continue and you should see task two.

Holla_Waldfee: Yes. 

investigator: Consider the following task, find relevant papers from a topic of your liking, which appear after 2017. Example topics could be "paper recommendation" or "author disambiguation", but should be from a broader area of computer and information science. 


##### TASK 2.1 #####


You can also come up with your own topic or you can also reuse "image recognition" as your topic.

Holla_Waldfee: Yeah, let's go something close to image recognition, but I think I've already answered quite a lot of these questions, but I think so if I pick a slightly different topic, let's go with texts to "image generation" and "artificial intelligence systems", something like that. I dunno. 


##### TASK 2.2 #####


investigator: Okay. Nice. How familiar are you with this topic?

Holla_Waldfee: Just as little as with a previous topic. 

investigator: Okay. 

Holla_Waldfee: So not, not very familiar. 


##### TASK 2.3 #####


investigator: Okay, perfect. And how would you define relevancy? 

Holla_Waldfee: That's a simple, but good question, I think.

If experts... I know that's a lazy definition, but if people that know something about the topic that have more expertise in that topic, say that this is relevant, I would trust them. Because yeah, trust the experts, if you don't know anything about the topic. 


##### TASK 2.4 #####


investigator: Okay, perfect. And then the last question would be, how would you solve the task. How would you find relevant papers from your "text to image" topic which appeared after 2017? I think they all appeared after 2017 or most of them. 

Holla_Waldfee: Yeah, probably.

investigator: Yeah. How would you go on solving this task? 

Holla_Waldfee: Mm, well, that probably depends on why I want to find literature on that topic.

If I want to conduct research on that topic myself I would probably Leave it with a really open search query just put in like text to image generation or something. And if I want to understand like parts of the process, like, for example what role a language model has in... for example, I just talked to a colleague about exactly that. And we talked about like two systems and the newer system, the only difference, as far as we know, is the kind of language model that's used. And I'd really like to find out what the difference is actually in the produced output in terms of what the used language model is.

I know that's really in that topic domain, but anyways, so I'd probably add more keywords to the query. The more I know my information need and the more precise I can formulate, what I'd like to know. I use that in the search query. Because I don't wanna read abstracts of 500 papers just to find the five I'm interested in.

investigator: Okay. Where do you put the query? The keyword query you starting with?

Holla_Waldfee: I just simply do a keyword search in the chosen digital library, for example, dblp. 

investigator: Do you have a specific one with which you would start? 

Holla_Waldfee: To be honest if it's computer science, I always start with dblp because, well, in my experience, I have to deal with the least irrelevant publications in there.

Of course. I miss some, but I will find, let's say 95% of the most relevant literature and only 5% of the irrelevant literature there. So yeah. 

investigator: Okay. And how do you find more keywords, which you want to add to the query? 

Holla_Waldfee: Usually iteratively, by reading abstracts, introductions, conclusions of papers. So, yeah, A really exploratory manner usually. 

investigator: Okay. Where do you get the full text of papers to do so? 

Holla_Waldfee: Well usually if I'm in my university's network, I get access to most of the computer science literature, and well, one out of, I'd say 200 publications I don't have access to. And if I don't have access to it, it's not relevant.

investigator: okay, nice. now that you are adding keywords to your query and checking papers, how do you decide on which paper is relevant?

Holla_Waldfee: Okay. So aside from the year of publication, of course, and aside from whether they are not authors of the paper are already familiar to me from previous searches or other research or whatever. I first of course read the titles and sometimes I can already tell by the title that it's not relevant.

For example, if there's like... really often in computer science methods, you get like "detection in medical pictures" or something, because that's, I think in, especially in like image processing, that's well, there's lots of tasks in medicine. So, if I cannot filter that out, I probably check for alerting keywords that signal irrelevance to me, rather than keywords, that signal relevance to me in the titles and if there's no "no-no" signs in the titles I check the abstracts really briefly mostly like the first few sentences and the last sentences. Yeah.

And if I still think that the paper could be interested, I put it to my Zotero search folder I create for most of the literature searches. 

investigator: Okay. That's great. is this all you do, or do you want to add anything else to this question?

Holla_Waldfee: Well probably a thing, but that's not connected to digital libraries, but of course I always talk to colleagues with more expertise. If I know that, for example, a colleague of mine has already worked with a certain system or has already published in that scientific field or whatever. I always try to ask my colleagues for advice, cause that can save quite a lot of time in the long run. Yeah, I think... do I have anything else? I think, no, I think that's probably more or less it. 


### Thank you ###


investigator: Okay then, thank you so much. And I will stop the recording.

