Knuspermüsli
===

### General Questions ###

investigator: Okay, recording in progress. 

Knuspermüsli: Got it. 


##### GQ1 #####


investigator: Now you should see general questions regarding usage of the libraries. And we can continue with the first question. Which tasks do you usually use a digital library for? Please tick all answers, which apply and complete your own tasks. Please give short examples of tasks you are ticking.

For example, you could say, "I tick person search because I like to keep track of myself." So, what are the tasks you usually use a digital library for? 

Knuspermüsli: So, I definitely use person search, not only to find experts and so on, but also more importantly doing my work where I need to identify individuals with the individual research.

So, searching by persons is very important. Searching by papers is important as well. Since I want to inspect who written the paper and so on and I wanted to have a look, which is very special to my case. Look at the first page and metadata is contained on the first page of a paper. But sometimes of course, I also want to read a paper. 

So, venue search, it's actually something I also do very much, but this is something that usually most digital libraries don't support, but something like, IEEE Explore and in such case, it is important for my work also to find a venue. I do not really look for BibTeX data. I do occasionally want to read a paper, so to get a full text and or to inspect just the first page for the metadata.

And I would not say that I use a digital library to study co-authorship relations, or venue-authorship relations, this is more or less something that I do using different tools than digital library. 

investigator: Okay. Is there any other tasks that you want to include here as well? 

Knuspermüsli: Yeah, let's think about it.

So, it's maybe not obvious from the other points is that I also do searches by identifiers so like ORCID or DOI. Like ISBN or something like this, since this is also an important point of identifying publications. Yeah. Paper search I clever assumes more like a title search or something like this.

investigator: Okay. But with these identifiers, you also do kind of a specific paper search? 

Knuspermüsli: It depends but I might also be looking for a person if the library supports this. So, this is a little bit more crosscutting, maybe than the questions above, but yeah. But usually, it's kind of a targeted lookup for them what I'm searching, not as much as an explorative lookup. 

investigator: Okay. Anything else you want to include here?

Knuspermüsli: I don't think so. 


##### GQ2 #####


investigator: Okay. Then we can continue to the next question. And this question asks you which systems or digital libraries do you usually use to solve these tasks? Please tick all answers, which apply and name others, which also apply, but on a given here, and please give a short oral description why you like or why you use a specific system? 

Knuspermüsli: So most definitely the ACM digital library, since it's such a predominant publisher in computer science, and there's a lot of publications to be found there. I should add that I usually use it only to use the published copies of ACM to look up there.

They have also this expanded listing of documents published elsewhere. I don't use the search to look at those expanded items in the ACM digital library. I've never fully understood Bibsonomy so I don't use it. I don't use the web of science since it's essentially not relevant for computer science. Same, a bit also holds for dimensions, which is nice that it's open, but it doesn't give any added benefit in the computer science domain and the alt metrics or something like that's contained there is a little bit boring and not really very meaningful information in my eyes.

For some reason I use dblp on a daily basis. I use Elsevier on occasion, cause there are a number of publications in computer science. I use Google Scholar. Yes. But mainly for looking for persons and not for publications. I don't actually use semantics Scholar that much, but I use it on occasion. If there's a PDF available for their... they sometimes have open copies from certain publications. On occasion I also use Semanic Scholar.

For Springer the same holds as for ACM and Elsevier. So, there's a, it's an important publisher and a lot is to be found there. I avoid using research gate. I don't find it very meaningful in computer sciences and so on. Usual search engine is essentially Google. I tried for a while using Ecosia, but I got bored by it and then switched back to Google again. Kind of trained of using Google in a certain way. I would add that I very much use the IEEE Explorer database. It grows. A lot of publications in dblp are hosted from IEEE Explore. Actually, it's the most prominent publisher. Yes. 

investigator: Okay. Are there any other systems or digital library you also use? 

Knuspermüsli: Yes, I would add Wikidata, actually. since I, for larger part start looking up the structured data there, and also structured data publications and authors. I also add openAlex, although it has not fully launched yet. I have done some queries there lately.

And I guess that's mainly it. 

investigator: Okay. That's great. Then I think we can continue to the next page. 

Knuspermüsli: Yes. 


### TASK 1 ###


investigator: To task one. Now you should consider the following task. Find two experts on the topic of your liking. Example topics could be "domain specific query languages" or "hashing functions", but should be from the broader area of computer and information science.

So, it can be any topic you want, and you do not really have to do this task, but we are going to talk about how you would do this task right now. 

Knuspermüsli: Okay. 


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


investigator: So, the first question is what is your chosen topic? 

Knuspermüsli: Should I choose a topic where already would know the answer? Something, I wouldn't know the answer?

investigator: Whatever you like.

Knuspermüsli: Oh, God. good question. So, I take a topic that I'm familiar with and then we choose Bregman clustering. 


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


investigator: Okay. So how familiar are you with this topic? 

Knuspermüsli: I'm very familiar with it, it's the topic of my thesis. 


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


investigator: Okay, perfect. And the third question is how would you define an expert? 

Knuspermüsli: I'd say it's somebody who has published repeatedly addressing that topic. So, kind of a record of published work in this area. Also, somebody, who would also be citing that related work also who's using essentially maybe would also call an expert in a certain sense. 

investigator: Okay. That's nice. Do you want to include anything else to this definition?

Knuspermüsli: Okay, of course it doesn't have to be limited to formal publications. This could also, of course be pivoted by given talks or having other auxiliary material on the web that's findable for Google searches. Something like this. 


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


investigator: Okay. So, should we continue to question four where you describe how you solve this task?

Knuspermüsli: I guess I would start in a sense of using something like Google and Googling for the general topic. Or the name of the topic, essentially. Maybe I would even go to Google scholar to have more focus on published results. At the same time, I would also look up in dblp.

Although dblp is more limited in that sense since it only needs perfect string matches in titles to reach results. So, it doesn't have the fuzziness of a query from Google. So, Google can be more convenient here and I think I just heard my headset complaining that it needs more power... there is the cable.

So, starting from there, if I find something where I'm confident with, then I maybe go for home pages and so on from the people I find via Google Scholar or something like this, if I'm not satisfied with the results that came up there, I would maybe we use also Something like the ACM or the IEEE or something.

It depends on the topic. So if it's something like I said clustering, I guess I would have more look into induction of ACM or even Elsevier or Springer to see what has been published there. I wouldn't look so much in IEEE, which is more in the electrical engineering side of computer science. But I would look up conferences, so, which are related to the topic, and I would've at least a skim through, or a small search, whether something has been published in these latest Iterations of the conferences and so on, that relates to the topic. 

investigator: Okay. How would you check for the conferences that are related to the topic?

Knuspermüsli: So, I of course know dblp. So, I would maybe start there. For clustering, for instance, I would know that I have to look at something like "NIPS", as an information theoretic side of clustering. I would have to look at something like "COLT" or... other conferences that come up from the learning side. I would start focusing, I guess, on the A ranked stuff, that staff that everybody knows is in the community that this is good quality and then maybe rarely go to let's say the B level or so of the conferences. I wouldn't mind so much for the small workshops and so on. Would have been too many, I guess, to track them all. And usually using dblp I would know where to look or even maybe just have a search, like something like learning or so in the conference search of something like IEEE or so. Where I can maybe discover something or remind myself of stuff that I have forgotten.

investigator: Okay. And from these conferences you've found, how do you find the expert or the experts? 

Knuspermüsli: I would identify the proceedings. In the proceedings I would go to the table of contents. I would skim through it. I would maybe look up for some hot terms in the titles, something like Bregman, which is the name of a influential researcher. About everything that's related to his clustering problems should have maybe his name in the topic, title, or maybe in the abstracts or something like this. And then I would have a small look at maybe just the abstract or the metadata of those items, and then maybe identify stuff that's interesting.

Have a look in there and probably if I'm not content with just recognizing names, in a certain sense that I see, then maybe I could even go to the references section and on, and maybe see who are the people citing. So, who are those people who are working on a topic are thinking who is relevant when working with this topic? Sometimes even when the time for this is given, I could also go into the paper in the related work section and read maybe just related work section to get an overview about results or so in the area. So that maybe from there you can identify what is relevant. 

investigator: Okay. And how exactly do you determine if someone is an expert or not? What's your decision criterion here?

Knuspermüsli: Or people should have contributed a significant result in a certain way I think. This is maybe the most striking evidence of expertise. If they do so repeatingly then maybe they're true experts of the field.

investigator: Okay. Do you wanna add anything else to this task? Or is this depicting your workflow here? 

Knuspermüsli: Yes, this is what I would do. 


### TASK 2 ###


investigator: Okay, perfect. So, then we can continue to the next page where you should see task two. Consider the following task, find relevant papers from a topic of your liking, which appeared after 2017. Example topics could be "paper recommendation" or "author disambiguation", but should be from the broader area of computer and information science. So, you can pick the same topic as before, or you can choose a completely new one. 


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


So, question one is what is your chosen topic? 

Knuspermüsli: So, the topic has been appearing after 2017 or the papers have been after?

investigator: Papers.

Knuspermüsli: Let's see, what have been the hot topics of the past few years?

investigator: Oh, but it can also be an older topic, but the papers should be quite... 

Knuspermüsli: Yeah, but let's go with something that appears after 2017, something like... "cryptocurrency". 

investigator: Okay. 

Knuspermüsli: Very important topic. Everybody should read about. 


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


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

Knuspermüsli: Absolutely not, as I know, it's nonsense.


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


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

Knuspermüsli: Pretty similar to before in the end. And so, I'm looking for paper again, not purposes. So, the relevancy of paper. That's a little more tough to tell, but of course citation is a form of the paper. I guess, in a certain sense the default sorting of Google would be highly influential in this task for me. Essentially who goes to the second page on Google. So, in the end whether I like it or not, I would maybe... relevancy would be in part defined by what search engines recommend to me, but if it's a new topic and new papers... okay.

What's a relevant paper, that's just one or two years old? It's pretty tough to tell. I don't know, actually, if this can be judged by anybody who hasn't read the content of the paper. But let's say it's a topic that emerge from 2017 then I would maybe be interested in what's the crucial paper that introduced this topic. What's the one paper everybody else is relying on to find their results. So, in the end for who's the ancestor of most publications out there. Would be something like relevance, yeah. 

investigator: Okay. Are you happy with this definition or do you want, add anything else to it? 

Knuspermüsli: It's okay. 


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


investigator: Okay, perfect. So, then we can continue to question four. How would you solve the task of finding relevant papers from your topic chosen, the cryptocurrency topic, which appeared after 2017? 

Knuspermüsli: So, the lazy method of course, would be to rely on what Google and so on would bring up the most relevant results of what they give there. So, in the end, this would maybe end up starting point. And I guess if I were really looking for the crucial paper behind stuff, that's not something that tools do support currently. That's something, I guess, that will only be reached by maybe finding those first suggested results from search engines and looking into inside of the papers.

Maybe have a look at the related results section to get an overview of the field and then move from there. Find the influential up links from this paper, what has been cited in this paper. And yeah. And then maybe have a look in this paper as well and find whether this has been influenced the others' papers before it. So, looking upwards from paper to paper in a way. And especially if I look for a number of papers, for five or so, and there are some papers who are mentioned in each and every one of them, it's also good indicator of this. If I want to use something with tool support, then there has at least at launch been the concept of influential citations in Semantic scholar. But I don't know, actually, after the data quality really holds the promise of what they want to implement with this feature. And also, I don't know whether this feature is still in semantic Scholar or has been there after launch only. 

investigator: Okay. If you would use this feature weather it exists still or not, would you enter keywords or would you go from a specific paper to find the most influential citation? 

Knuspermüsli: I guess I could also use the Semantic scholar or so to make a keyword search to identify the papers, but essentially this feature is after I've identified a paper. Then I would use this as a starting point to look upwards, the citation chain and maybe also don't rely on one paper to start with, but four or five or so to see something like a common co-citation. Something like this. 

investigator: Okay. Do you wanna add anything else to this question or is this how you find relevant papers, the ones that appear often? 

Knuspermüsli: Okay. That's how I would try to do. 


### Thank you ###


investigator: Okay, then I think I can stop the recording.

