 All right, I think we're ready to start. Thanks for coming to this session about this call the orphans project I'm Herbert van der Sompel and I will be presenting with Martin Klein. This is actually the first presentation I'm doing with my new affiliation. I used to work at Los Alamos With Martin, but I now work in the Netherlands at an institute called dance Which stands for data archiving and network services But the work that we're reporting on is work that I did while I was still at the lab with Martin and then my team over there So it's called the orphans is a project funded by the Mellon Foundation And we have a team at Los Alamos and the team at Old Dominion University Guided by Michael Nelson who's here and who's giving a keynote tomorrow as a matter of fact not to miss so the consideration at Basis of this work is something we all know is that increasingly Scholars are using web portals To do go about their business basically They do that because these portals have you know attractive characteristics and even if our institutions would Locally provide similar platforms for productivity The researcher would still be interested to go in the web version in the flow because there you basically are socially embedded You have global collaboration. You have global visibility of what is going on as opposed to local institutional visibility So just two examples here. Here is one researcher. She has an orchid and then she has accounts Identities in many different productivity portals So here a github and slide share in fixed share Poublon's and then probably what is a Personal website and here we have Sean Jones. He's at the Old Dominion University a personal site He also has a github Account these things in slide share does a lot in Wikipedia actually and does some blogging also These are just you know examples. You all know that this is going on Now the net result of them using these platforms is of course that they are depositing artifacts over there and These platforms they can be dedicated to scholarship, right? So think share and Poublon's are commercial examples They could also be not for profit like the open science framework or as a nodo But they also work in general purpose Platforms again both commercial and not for profit. So Wikipedia wiki data we talked about Github and slide share are examples So there's wide variety of these portals that they're actually using and they're depositing their materials there and There's three Problems actually with that regard so first of all typically the Institutions where these researchers work they have no clue that their researchers are depositing and they definitely do not have copies of the things That are being deposited. So they just use these platforms They drop their intellectual property there and the institution doesn't really Have any copy or any knowledge really of it at the systematic level The other thing is that as we know where portals come and go Long-term access to these materials definitely is not Guaranteed When you look at the commercial platforms No one knows when they're going to change their business model and when suddenly access may not be free anymore They clearly don't have a commitment to long-term preservation. You just look at the terms and conditions on these sides It's all about we can stop our service at any moment There's nothing about we'll persist that information for a long term and when you look at the nonprofit portals Where they have an unpredictable funding stream and they could go out of business for other reasons and many of them Live from ground to ground really and then the third platform is a problem with these platforms is that the content there is typically not systematically archived frameworks like locks and portico they focus on the journal literature not on these artifacts that are deposited on the web There are certain of these portals where one can actually drop those kind of artifacts and Researchers will typically do so not because they care about long-term preservation But because they get a DOI in return that can then can beside it and they do that of course not for all their artifacts, but for things for which they really want to do I This institutional repositories, but obviously we cannot expect our researchers to upload stuff in a web portal and then also in an institutional repository and Then we know from the hyperlink research for example with two papers in a plus one Resulting from that research that these artifacts are typically also not collected by your typical web archiving activities like the Internet Archive and so on so we have evidence in papers about that that Materials that are referenced that are web materials meaning not scholarly papers and so that are referenced in scholarly papers Typically are very poorly web archived. I Won't go into the details of the hyperlink research. I'll just show some anecdotal evidence here This is Emma again And so she has a presentation in slide share and then at the right hand side here You have a screenshot of the time travel service. This is a search engine that goes across all public web archives About 25 or so at the moment I think are covered and you see that in none of these archives is there a copy of this presentation and Then another example here Sean has a Github repository right and we see exactly one copy of it in the Internet archive Just anecdotal of course and again as I said in the hyperlink work We found very hard evidence that these things are not well web archives. So these are the three problems Related to our scholars using these web productivity Portals and so the scholarly orphans project is all about so Understanding that these problems exist understandings our researchers are doing it. How are we going to archive that material? and in this project we took a perspective that is Institution driven so it starts from the premise basically that Institution should really be interested in that material because after all it's intellectual property of their materials of their researchers it's also Collecting these materials. I think aligned with the mission of academic libraries and Then one other Reason for taking this perspective is the fact that academic institutions typically do have a long shelf life Right many of them have existed for hundreds of years So these are reasons that we take this institutional perspective to try and address the problem So here you see a depiction We have the institution researchers in the institution the web with productivity portals and These people these researchers have identities in these different platforms and they create artifacts And so we are interested in this framework in actually capturing the artifacts created by these institutional researchers and we think again we are this is our hypothesis that institutions should be Interested in actually trying to capture these things Second perspective is that we take a web archiving perspective on this one because of the scale of the problem Second I already mentioned you can't expect your research to upload all that material in an institutional repository And then there are some examples out there of bilateral agreements between For example github and fig share or github and Zenodo where it is possible to transfer Material that sits in github for long-term preservation in another environment But those bilateral agreements you can do that for a couple of portals But probably not you know doesn't scale up if you want to do it generally So basically we're taking a web archival perspective and basically going to say at the end of the institution Processes will run or on behalf of the institutions Processes will run that will attempt to well first find these artifacts and then capture them and put them in the Long-term repository. That's the perspective that we're taking here We've not only just talked about this. We've actually built something that does what I just described And I'll go into what we call our prototype pipeline for the scholarly orphans project here And Martin will give us a demonstration in a bit. So basically the pipeline consists of three components The first is tracking the artifacts. So that's basically Polling the API's of different systems out there to figure out did our researchers deposit something new okay When we have that information that indeed something new was deposited for example in slide chair that information Then goes to a capture engine where basically a web crawler is instructed to go and Find that slide chair presentation bring it across and put it in an institutional archive And then in this step we go even one step further Once we have the result of the capture process We actually also deposit it in a web archive that is cross institutional So that would be a scholarly web archive where many institutions deposit the results of these capture activities in All of this in our case is orchestrated By a component that basically makes sure that these things do their work when they have done their work They call the next one and the next one and song and all of the information of these things happening ends up in an event database And when you'll see the demo that Martin will do and he'll show you landing pages for captured objects That is actually serialization of information that sits in that event database So basically all these steps for each of the artifacts that are found are recorded in this event database So I'll go step-for-step Into each of these so first of all tracking the artifacts So in order to be able to track that your researchers have deposited something somewhere First of all, you'll need their web identity in the portals that you're going to track We've done previous work Using algorithmic discovery. We have a paper in code for lip about that. I think that was called the ego system Works well if you have a name like Herbert van der Sompel for you if you're Paul Smith, right? Discovery via registry. We have a paper in JCDL about that where we looked at the orchid Registry basically to see whether researchers actually Provide information about all their web identities in there answer being no not really so currently in our setup we still start from the perspective that one would have to manually collect these or Significantly improve on the algorithmic discovery. So that's one thing you need those web identities Another thing that you obviously will need is apis at the end of these portals where you can access using the web identity of your scholar and Hopefully also where it allows you to say show me all the new contributions since a certain moment Okay, and the result of all of this activity this tracking activity is one or more URLs of artifacts that a researcher deposited in the portal Challenges that we've encountered when building this is that Most of these portals Actually do support access by web identity. So give me everything by this person with this handle, right? Except for the ones that are dedicated to scholarship Very few no, I mean all the social platforms out there in a slide chair and all you can go You know use HVD some to handle and you get all the stuff back you go to the scholarly portals They don't support that that functionality obviously what with the proliferation of Orchids it would be really nice if they would all support functionality for this orchid what you have what you have deposited since and so on yeah Another problem, which Michael is also very interested in I know is a lot of people use these portals both for professional and for personal reasons and making that distinction can sometimes be really hard and Since you're focusing on you know institutional materials really It's a problem that needs to be solved and then of course how how frequently do you want to track? Right You want to track rather frequently because you want all the new information before it's stale and so but if you track too much You put too much load on your environment and so so that it's not a big problem, but it is an issue that does come up All right, we've done the tracking. We're now going to the capturing aspect This is where we actually start with the URL of an artifact that needs to be captured And we're going to use web archiving techniques To go and get it and the result of this is a work file. That's how you know One stores the results of web archiving activities for this new artifact and it is deposited in an institutional archive very interesting Challenges there also Of two of two types really One is what we call the boundary problem, you know when you have a URL of the artifact Typically, there's more than only that URL that needs to be captured and what really needs to be captured Well, the truth there is in the eye of the beholder one curator would say you need this this and that you need all these Related pages also and I would say what fine if you just have the landing page. That's good enough We have a trace of you know, what was going on there and the other something that The people from the internet archive obviously know very well is that it is getting extremely hard to capture web pages because of a lot of interactive content and So these are the problems that we were actually facing and while thinking about these problems We actually made a bit of a breakthrough Which I'm not going to talk in detail about because Martin will probably for the next C&I do we talk about that? It's called the memento tracer approach Again, not a lot of detail. But basically what we're doing is we let the curator Interact with a page of a certain type of a certain class an example page Let's say a slide share page and by interacting the curator basically indicates What needs to be captured for this specific page and the curator has a web browser plug-in That records all that information in an abstract manner almost like in a set of instructions Okay, for this kind of page you need to capture this this and that so he does that by clicking and so on That we call that a trace that information can then be used by a headless browser That operate server side to actually collect the materials with high fidelity Okay, so again, that's momentum tracer There's a website for that if you're interested in more information at tracer dot momentum web dot org Third component as I said really icing on the cake because at this moment We've already brought the material into the institutional boundary So icing on the cake is we're going to run a web archive in which many institutions can deposit The results of these activities and now you have a scholarly web archive basically So it's not really an awful lot to be said about that Once the material the work file was ingested into the institutional repository We know the URI of the work file that is then being sent over to the cross institutional web archive Comes and collects the web part of the work file and ingested into the web archive Want to do that with off-the-shelf web archive in software like pie web or open way back And the result of that is of course that you have mementos you have snapshots of these resources in a web archive Didn't really run into major challenges with this we did try and run this on IPFS So using a distributed file system basically with software from all dominion But we ran out of time, but this really should be possible the idea being that rather than having a centralized web archive You know The notion would be that institutions collaborate and each provide their own IPFS node But still they form a community that they become really one big virtual web archive Again, they should be doable, but we didn't have time to complete this work And with this I'm going to hand over to Martin who is going to do a demo of all of this because it's actually real I Think you heard it so I will take maybe the remainder of our time to give you a little bit of a demo of the system that we've implemented on a prototype level and I will follow up by Showing you sharing some statistics with you that are based on our prototype Give you giving you an idea of what we have collected who has contributed and what these Interactions basically look like right over these artifacts look like that we have collected so We were we thought if you know John Oliver can establish a church we can establish an academic institution and so we did That has a representation at my research Institute and We were able to recruit 14 scholars to our fictitious institution depicted here on the left and of course Herbert and I signed up immediately as well to that institution and What all these scholars you have in common is they have an orchid so they're good citizens, right? Because as Herbert alluded to we're using orchid as our primary identifier for scholars All of these researchers furthermore use these productivity portals are a number of those productivity portals that Herbert had mentioned and Create different artifacts in these portals right so 16 researchers that were basically Enrolling in our institution and we are as a representative for these institution are interested in tracking Capturing and archiving that their scholarly artifacts on the web, right? So which portals which productivity portals are we tracking currently well those 11 actually rather 10 Plus one more so you recognize. I'm sure most of those icons, right? We have github fixed share slide share hypothesis blogger as a platform for people to you know write blogs, of course pooblons a portal that lets you share peer reviews Stack overflow for computer scientists mostly medium as another blogging platform Wikipedia WordPress as more platforms of that sort We also for good measure basically included tracking their scholars Websites personal homepages basically and Herbert alluded to this. There's of course this Debates to be had that this is mostly scholarly content there or personal content But we decided to punt on this ball because you know if I'm uploading a puppy video to YouTube. It's probably Not scholarly. However, if I'm a veterinarian and upload the same thing it might be scholarly, right? So we didn't want to go there So we started tracking artifacts in these portals At the end of August of last year so all the way through now and we went back to August 1st So up until now basically you have eight months of tracking data of scholarly artifacts collected, right? Okay, so let me Switch gears here and Show you the representation of our research Institute my research Institute if you do reference this URL you'll get a Website that looks like this. So you see a matrix of Individual researchers and each of these icons represents one artifact that we have tracked from the corresponding researcher Right and they're ordered in reverse chronological order starting on the top left Moving over to the right dropping down So you go through your labyrinth like this right in reverse chronological order meaning the most recent tracked artifact It's that one on top left the least recent on the bottom right So there is different filters that I can apply To this view because this is now the researcher view right what are what sort of? Researchers have done some something in productivity portals that we were able to track I can also filter our portals that gives me an overview of what the portals were that were Interacted with what we captured tracked captured an archived Artifacts from and you already see a little bit of a dominance for github some personal web pages here some Wikipedia artifacts there and without Going for the detail. There's another filter that I can apply if I'd like to see only artifacts from a particular point in time I can put a temporal filter in this interface and I can also since right now if I scroll down here We only show 100 the most recent 100 artifacts. I can increase that number by to let's say 5000 and So to run and then I'll get the most recent 5000 Artifacts tracked right you get the idea of how this works. All right, so a couple of examples there like to share This one for example So if you click on one of those heads one of those researchers you get to the landing page as her it has a little to all of that tracked captured an archived artifact in this case it is a an artifact deposited by the researcher Daniel in github it is a source code contribution basically and the researcher created or Contributed to to the particular repository in github So we further show a number of metadata items basically surrounding this so belonging to this to this artifact The original URL where it lives in github the daytime of when this artifact was published So when did the research deposit something? What's the real-world name of that researcher the orchid in this case his orchid? and the the portal in which the artifact was tracked and also very important the username of that researcher in that particular portal All right, because again, we need to since for example github does not Support the search by orchid. We need to still translate orchid to username in a particular portal Let's scroll down on This one a few pieces few more pieces of information for one the URL of the memento So this is the archival snapshot that we have created right based on the original URL right here and if you sue Fancy you can also download the walk file that we created for the institutional on an institutional level basically, right? That's a distinction that Herbert alluded to the institutional repository the institutional archive versus the archive that can potentially be across institutions, right and Herbert had also shown the diagram of our pipeline and Between all these individual components and the orchestrator the brain of the thing in the center There's always messages being exchanged send back and forth if you'd like to look at those messages They're linked here as well. So you can actually take a look and another shout out to a development work done at ODE on the Right-hand side memento embed is basically a social card of that created archived snapshot To give you like a sneak peek basically of what you can expect if you if you look at that resource so Just to show you that there's really actually something done has been done here so I can Open up the original URL and I open up the archived copy of it the memento of it There you go So this is a github.com right the entry page basically of the repository in github looks quite nice and The memento our archive copy Looks very similar right just by by visual observation basically however you see the banner indicating this is an archive copy this is not life be aware and now I can you know Click on these links to show you that we actually really did archive stuff Takes a little bit to load. So there's really something there and something that were Somewhat proud of is that we're able to to archive Resources based on interactions with these resources right so I can actually in the archive copy click on this clone our download link And I can click on the download zip link in order to actually get a copy of a zipped copy of that repository Right I can click on it and I'll be getting the download dialogue Right so that is something where the the notion of high-quality capture of the scholarly artifacts comes in In other example, oops, let's get this way. I want to share with you is Emma and Emma's slide share presentation that she uploaded to Slice.net and I'll just click on The original and also again on the memento For two reasons a to show you that the original has a 62 slides right so archiving that Basically manually with an approach like web recorder requires you to click 62 times right to get every single slide with our With our approach the moment to trace your approach we can automate that come on and Show you the 62 slides in our web archive you know and Click through the slides right so again a notion of high-quality Trying to strike the balance between high-quality captures and also scalability of the approach, right? so one last really quick example here is From Emma again and Emma's also active writing peer reviews and sharing that peer those peer reviews on poob lines So I'll Show the original and the live version and I showed the memento again the reason I'm showing this to you is because The now that the manuscript nor the review are publicly available right so Emma has chosen to share that review But not publicly so it's an upload It's an artifact in arguably but it's not available to the public It's not open to the public and hence of course our approach cannot make that public all of a sudden And so you'd also only get this landing page of the review, right? So that's going in the direction of saying, you know, we cannot archive anything that is not public That's the point here to to take out All right, so I invite you to to To play with the system to do your searches yourself see what sort of artifacts we've tracked And archived and captured and see what sort of researchers are doing what? All right with that I'll go back to my Presentation and I'll I'll share some statistics with you that are as I mentioned earlier based on our pilot our prototype that has been running for the last eight month and Based on these 16 Excuse me 16 researchers and based on these 11 scholarly portals that we've tracked, right? So this is our data based on our pilot your mileage may vary in this regard, right? Because we've made some executive decisions what portals what frequency and what researchers so but The reason why we thought it would be interesting to share these these data with you is if you were in a position If you find yourself in a position to say this is an interesting sort of concept I'd like to try this in my department in my Library in my university. What can I expect right in terms of? load basically and Distribution so let's start with the distribution of scholarly portals Total right and you see this is dominated by github Three quarters of artifacts that we've tracked come from github So that is interesting, but it's also not overly surprising me because github is intuitively speaking a portal where you frequently interact with right. It's one commit after another In and on top of that we're tracking different artifact types on github So you can leave a comment, right? You can do a pull request. You can do all sorts of forking and whatnot So there's different interactions different types of interactions different types of artifacts that we're tracking in the same portal Versus compared to slide share for example, there's really only one interaction of our tracking which is an upload of a slide deck So that's not quite ever samples, right? But you still see, you know three quarters of github personal websites 18% of of our facts make up 18% of our facts Wikipedia to my surprises up there with 5% So, you know something maybe not necessarily a scholarly productivity portal that you would have thought of I didn't and you see a little bit of fixture and slide share Less than 1% but still still there, right? There's this There's something going on All right, so then we ask the question is this distributed equally across all researchers, right? All researchers, you know equally productive in this context and of course the answer is no Even though we chose those researchers Somewhat randomly we try to you know account for diversity and so on to fourth But yeah, so there's obviously one winner that contributes the most amount of artifacts But there's a good number of folks that you know do a thousand fifteen hundred a thousand artifacts So that's an interesting observation to make I think when you're looking at you know What as an institution from an institutional point of view? What do my scholars do and who's who's there the the winner and in our case We've kind of de-anonymized that and maybe some of you have all know Daniel He's very active on a github and Wikipedia actually so he is clearly the winner in our little competition here and For full disclosure, her and I are not Right, but this is of course very biased because we didn't track all portals right so that's my excuse anyways, all right So then Next question that we tried to answer is no snow again from an institutional point of view. What sort of load can I expect how often? Do I actually find? artifacts in scholarly productivity portals how much do I have to do and This the result basically looks like an EKG right so there's peaks and valleys the total frequency of artifacts tracked per day for the last eight month and There's two interesting observations from my point of view three a is it goes up and down right Be the peaks don't go beyond a hundred and ten total artifacts per day And again, that's this is not this is biased right towards the researchers and artifacts and so on to fourth portals rather than be chose But that's it there's a decent there's a good average probably right and then thirdly not all that surprising during the holidays No one's doing anything The so the obvious dip there in this in the typical frequency curve, right? so The next slide then I show you is trying to answer the question is there Does this happen? Distributed on an uneven level so are the all the artifacts in GitHub coming at the same pace basically all the Artifacts coming for slight share at the same pace and they're also the answer is no and this might be a little bit hard to see but each of those little sticks Represents an artifact tracked on a particular day The the color shape of the shade of the of the stick represents how many of those artifacts were there per day, right? so the regular Bright red basically is somewhere between 1 and 22 artifacts on that particular day and the darker the shade gets the more artifacts on that Particular day were tracked right so it goes all the way to 110 as confirmed in the previous plot So a couple of interesting observations from my point of view right the first one is GitHub is basically constant right this is a constant stream and again This might be due to the fact that we're tracking several different artifact types in this portal, but still Confirming intuition that GitHub is a portal that you interact frequently with one commit after the other This is a constant stream of artifacts that we were tracking Wikipedia Mostly so let's say but then fixed shares slight share poob loans WordPress those seem to be isolated by those seem to be almost clustered So, you know three here and then maybe the next week to there so that At least that would mirror that does mirror my behavior on fixture for example just a couple weeks ago over a paper Oh, I need to share my data set so I uploaded you know five datasets to fixture within 20 minutes And so that's kind of the behavior that I think is mirrored here Medium is Interesting because you have exactly one artifact from medium. We expected a bit more there, but man something is weird another interesting observation from my point of view obviously this sort of a Plot cannot tell you What is the next hot productivity platform to emerge? But it may be able to tell you in retrospect. Well, what is on the descent right? Which platform has been abandoned sunsetting and so on so forth, but for example hectic happened to stack overflow No idea, but there's basically nothing happened since the beginning of October So again from an institutional point of view this may provide you interesting insights into what you allocate resources What to focus on? Maybe you want to increase or decrease the frequency or get up tracking those sort of questions can potentially be answered with this, right? all right and now we looked inside and Came up with some numbers that reflects our pilot our prototype and what it has done basically over the last eight months, right? So they have tracked a total of 10,187 unique artifacts over the course of these eight months The event database that Herbert mentioned right on the bottom of this Of this diagram that basically knows everything about everything that we do within the pipeline It's not more than 41 megabytes right now. So this is again all metadata, right? That hence it's it's manageable We have 61 gigabytes of works. Of course, that's a little bit larger depending on you know The size of your slide decks and so on to forth and the web archive index is much smaller than that So it's just a couple of pieces of information there Then we're interested in Given the pipeline that we have the tracker the captor and the archiving process Which of these components takes the most amount of time and again that is of course biased because it's our implementation And obviously this is you know a research project So we don't don't have the resources mirroring like the inner archive for example So obviously this is a small scale, but in terms of proportion there was still interesting to see it. So on average For 50% of the artifacts that we tracked the capture process is done after nine minutes But that that's good or bad. I don't really know right? There's no I don't have a benchmark But it's good to see that this is not four days again in terms of scale scalability and manageability Of an approach like this, right? Our standard deviation there is pretty bad because it's a research project things go wrong So if you if you look at 75% of the artifacts that duration is much much longer, right? The archiving process of the indexing in the potentially cross institutional Web archive also takes roughly ten minutes nine to ten minutes for 50% of the artifacts to be there so you know in an ideal scenario these These these two processes don't take crazy long and it's still manageable on an on an institutional level All right, but they're also If you look at more, you know 75% of artifacts 100% of artifacts the duration increases because things did go wrong during these eight months all right, so two more slides one is Briefly summarizing what I think we've done and I think what we're trying to to accomplish for this for this project and so for the scholarly offerings effort what we're really trying to do is explore an institutional perspective of capturing Scholarly artifacts that researchers deposit on the on the open web, right? So we're taking the institutional perspective given the notion that institutions intuitively have a long shelf life and capturing the intellectual property of their researchers should be aligned with their mission and vision statements, right? So to make this concrete I looked it up actually the University of St. Louis has two thousand two hundred researchers So round this up with the number of postdocs graduate students other people that publish 5000 5000 researchers that you as an institution are tracking in We did 11 Maybe it's two dozen. Maybe it's 30 portals. That is still a manageable scale, right? compared to how many researchers are there worldwide trying to if you're trying to to Grab everything that they produce. That's that's a Google scale effort, right? so and on top of that we're focusing on these artifacts that are not necessarily in scope of efforts like locks and portico Regular, let's call it incidental web archiving And the pilot that we shared here represented at my research institute is really there to demonstrate This can be done, right? This is the technology is there The methods are there the API is there ish there could be of course much much better and Facilitate much better collaboration there or cooperation there, but it does work There there there are opportunities as I think have shown with the stats provided But there are of course also challenges as Herbert has eluded to So the last slide that I will share with you is Actually my favorite Because we did ask all these 16 researchers beforehand before we started this product process For their permission. We send them an email briefly outlining what we're trying to do and ask for their consent And we get all positive responses All right, and here are a few that I'd like to point out right the first one the first researcher respond That's awesome Thanks for letting me know carry on as usual and feel free to monitor away I'll try not to change my behavior or anything now with this knowledge, right? So giving you the Giving us the confirmation that this researchers researcher knows that now we're doing some monitoring on his or her output, right? But still he or she is perfectly cool with that the second Statement that I want to share is this is fine since everything you're capturing is public to start with Yeah, go ahead. I also wonder if you know about software heritage So again the notion of researchers aware that we're only capturing stuff that's public to begin with and be also pointing us at other scholarly portals that we could look at and or pointing out other projects that are Real that is not dissimilar to ours, right a Little bit of a pet on our shoulder and very comfortable with being part of this very important research project So we've all done But researchers can be hip I'm cool with it smiley face and then my absolute favorite actually came in two emails In response to our us paying this research at the first email says Interesting project. I'm happy to participate And then a few minutes later the same researcher sent another follow-up email saying oh one more thing Is it possible to get the copy of the URI ours? So the URLs of the scholarly artifact That you guys detected so I can feed them into an archive of my choice It's a two things there a this researcher also has no idea what he or she is depositing in any of these portals, right? It's completely lost track Rightfully, so I mean that's understandable and be this researcher also does not trust the Scholar infrastructure that we currently have in terms of sustainability and long-term access of the research right so I thought that was interesting to to to see to share and It gave us a little bit of peace of mind right starting the project So now okay, we did the right thing. We asked them and everything is cool. So we started right so with that I'll close here and Opening up for for question for her to myself and thank you so much for listening