WEBVTT Kind: captions Language: en 1 00:00:00.359 --> 00:00:05.420 >> DEANNA MARCUM: Welcome to those of you on the webcast. I hope you are following along 2 00:00:05.420 --> 00:00:12.660 and getting a lot from the program. It's very exciting here in person, largely because we 3 00:00:12.660 --> 00:00:18.050 have so many great comments from all of you, and we want to continue that with this panel 4 00:00:18.050 --> 00:00:30.170 as well. We've talked a lot about what is going on right now, what are the gaps, we've 5 00:00:30.170 --> 00:00:37.440 had lots of comments about successful projects, but in this panel, we're going to try to go 6 00:00:37.440 --> 00:00:45.989 a little deeper into those successful projects and think about what would be needed to bring 7 00:00:45.989 --> 00:00:53.690 these successful efforts to a level of scale that begins to feel like part of this national 8 00:00:53.690 --> 00:01:01.000 platform. I'm very happy to introduce the panelists we have talking about this, and, 9 00:01:01.000 --> 00:01:15.290 let's see, in the order in which they are speaking, Jon Voss, who is the Director of 10 00:01:15.290 --> 00:01:22.420 Strategic Partnerships at History Pin. The second speaker will be Anne Wootton, who is 11 00:01:22.420 --> 00:01:30.580 the CEO of the pop up archive. The third speaker is Mary Flanagan, who is a professor at Dartmouth 12 00:01:30.580 --> 00:01:37.490 College, and the final speaker will be Trevor Munoz, who is the assistant dean at the University 13 00:01:37.490 --> 00:01:46.740 of Maryland libraries. When we've talked about gaps earlier today, we've heard indications 14 00:01:46.740 --> 00:01:57.329 of some of the things that are needed, crowd sourcing solutions, computational approaches, 15 00:01:57.329 --> 00:02:06.650 linked data. These are some of the things that our panelists will describe today, and 16 00:02:06.650 --> 00:02:12.730 I hope that as they talk about their projects and how they think they might be expanded 17 00:02:12.730 --> 00:02:21.940 to become something bigger, I'd be very interested in having them comment on what are the emerging 18 00:02:21.940 --> 00:02:31.230 technologies that make these solutions possible, what kinds of criteria can we apply to other 19 00:02:31.230 --> 00:02:39.709 projects to see if they fit into this national landscape that we've been describing, and 20 00:02:39.709 --> 00:02:47.910 what needs to happen to move projects from individual research projects or individually 21 00:02:47.910 --> 00:02:56.480 funded projects to something bigger, and then I have another question I'd like all of you 22 00:02:56.480 --> 00:03:05.209 to think about. We have a number of national organizations that have a mission to serve 23 00:03:05.209 --> 00:03:14.849 our needs, and there are lots of them, but, you know, just for sake of argument, there's 24 00:03:14.849 --> 00:03:23.110 the Library of Congress, the National Archives, OCLC. Internet Archive falls into this category, 25 00:03:23.110 --> 00:03:31.209 although with a different funding model, I will have to say. We have organizations with 26 00:03:31.209 --> 00:03:37.849 national missions. What do we need from them? And I'd really love to hear the panelists 27 00:03:37.849 --> 00:03:47.650 talk about that too. What can we do to make sure that all of our efforts are leading in 28 00:03:47.650 --> 00:03:54.349 at least a parallel direction and pointing to something bigger that we can all be very 29 00:03:54.349 --> 00:04:01.030 proud of? So, with that, I am going to turn to the panelists for your comments. 30 00:04:01.030 --> 00:04:09.260 >> JON VOSS: I guess I will start. I'm Jon. I wear two hats usually, at least, history 31 00:04:09.260 --> 00:04:15.230 pen which is about building community around local history , and on the other side, I co 32 00:04:15.230 --> 00:04:25.759 founded the linked open data libraries, museums and archives summit, and it's the latter part 33 00:04:25.759 --> 00:04:30.280 that I'm going to talk a little bit about, with linked open data, and I have to say, 34 00:04:30.280 --> 00:04:34.919 it's amazing to be in a room with so many of you here and on the Twitter feed as well. 35 00:04:34.919 --> 00:04:39.529 This panel is some of my favorite people that I'm still getting to know, and knowing that 36 00:04:39.529 --> 00:04:45.820 I came completely from outside of the world, somebody passionate about stealing all of 37 00:04:45.820 --> 00:04:50.229 this stuff, and it turns out that people wanted to make it more available, so I'm excited 38 00:04:50.229 --> 00:04:55.439 to be a part of that, and I also say right now I'm a bit sad with Nepal and Baltimore 39 00:04:55.439 --> 00:05:03.819 and extreme poverty, racism, systemic rather, that it's hard to think that what we're doing 40 00:05:03.819 --> 00:05:10.339 is so important and takes such precedent, but at the same time, libraries are a lifeblood 41 00:05:10.339 --> 00:05:18.029 to so many people, and I think it really heartens me to know that particularly public libraries 42 00:05:18.029 --> 00:05:23.110 are there when all other government intervention has failed and to see that again and again, 43 00:05:23.110 --> 00:05:26.949 so it's an honor to be here. Linked open data is something that gets thrown around quite 44 00:05:26.949 --> 00:05:34.300 a bit, so for a second, what is it? And I'll say without getting into any technical detail, 45 00:05:34.300 --> 00:05:39.919 it's ultimately a combination of technology, policy and law that has the potential to build 46 00:05:39.919 --> 00:05:45.180 a more useful, discoverable and interoperable ecosystem, but more than anything, it's part 47 00:05:45.180 --> 00:05:50.630 of a fundamental cultural shift. So, linked open data uses the technology of the world 48 00:05:50.630 --> 00:05:56.599 wide web to share resources in a machine readable and predictable way. Those are the two key 49 00:05:56.599 --> 00:06:03.029 pieces to it, and technically, it's a combination of standards, tools and protocols, but also, 50 00:06:03.029 --> 00:06:08.789 licensing and data normalized to scale. So, this is hopefully moving us from what started 51 00:06:08.789 --> 00:06:15.460 as a web of interlinked documents, remember, 1991, 92, to a web of interlinked data, and 52 00:06:15.460 --> 00:06:21.409 that is ultimately the potential. The environment and the culture of linked open data sees the 53 00:06:21.409 --> 00:06:26.840 web less as a traditional publishing platform, and I think this is part of a cultural shift, 54 00:06:26.840 --> 00:06:31.689 and more as a living interoperable network that provides new and entirely different ways 55 00:06:31.689 --> 00:06:36.719 of engaging and collaborating with communities to enrich collections, connect them to the 56 00:06:36.719 --> 00:06:40.050 universities of knowledge, and build new things from them, which you're going to hear from 57 00:06:40.050 --> 00:06:43.800 the other panelists here and the exciting opportunities around that. 58 00:06:43.800 --> 00:06:48.800 What is it not then? One, it's not magic. You can't write a grant and then say it's 59 00:06:48.800 --> 00:06:52.809 going to be linked open data too. That would be awesome. There actually aren't computer 60 00:06:52.809 --> 00:06:57.000 wizards that can turn you content data into something that it's not. It just doesn't work 61 00:06:57.000 --> 00:07:02.249 that way. There are standards, and there are systems and protocols that allow us to do 62 00:07:02.249 --> 00:07:06.669 things and make things more easily accessible, but it's not going to happen just like that. 63 00:07:06.669 --> 00:07:11.719 You also can't say this is going to solve everything. You can't load lamb something 64 00:07:11.719 --> 00:07:17.280 and expect it to solve all these problems. Finally, it's not something that you can just 65 00:07:17.280 --> 00:07:21.759 buy. You can't say to a vendor, throw the load lamb on it, that'll make it work, it's 66 00:07:21.759 --> 00:07:28.639 got to be linked open data. We can push our vendors to build tools to make things more 67 00:07:28.639 --> 00:07:31.569 interoperable, and that's something we really have to do, but that's more about the policy 68 00:07:31.569 --> 00:07:38.289 bit of it. The potential is huge here. Obviously, unlocking the world's cultural heritage data 69 00:07:38.289 --> 00:07:45.309 for good is a big deal, one big library. You know, this is ultimately where we can start 70 00:07:45.309 --> 00:07:49.909 to bring things together, use them together in ways that we're not just going to one place, 71 00:07:49.909 --> 00:07:55.899 but that we have at our fingertips all the world's knowledge. Well, it's a dreamy big 72 00:07:55.899 --> 00:08:02.770 idea, but crazy enough, the technology is already here. The shifts for us here in cultural 73 00:08:02.770 --> 00:08:07.679 heritage are really more about licensing and policy. On the enterprise data, linked data 74 00:08:07.679 --> 00:08:13.800 is being used in major, massive corporations around the world to link different data sets 75 00:08:13.800 --> 00:08:21.449 around their regions, etc. One minute? Geez. But for us, I think it can be something bigger. 76 00:08:21.449 --> 00:08:25.550 So, let me dig into, you've heard mention of, obviously today Digital Public Library 77 00:08:25.550 --> 00:08:30.169 of America, Europiana, Digital Enzed, these are different massive international aggregates 78 00:08:30.169 --> 00:08:34.580 that are starting to make things useful in more ways. So, what I expect us to see in 79 00:08:34.580 --> 00:08:40.400 my last minute and hope to see is more specifically, and we can dig into this in the Q & A probably, 80 00:08:40.400 --> 00:08:45.761 more multi institutional collaborations focused on specific domains. So, there are more and 81 00:08:45.761 --> 00:08:53.839 more projects that are taking, A, a discreet domain, two, identifiable resources, and hopefully 82 00:08:53.839 --> 00:08:58.870 for multi institutions and building applications and visualizations around them. It doesn't 83 00:08:58.870 --> 00:09:03.379 have to be huge, it doesn't have to be major resource suck. It's experimental. It's 1993 84 00:09:03.379 --> 00:09:08.269 in the world of the web, basically, when we think about linked data. So, I'm hoping to 85 00:09:08.269 --> 00:09:13.569 see more of this community approach, and finally, really, the dream about linked open data is 86 00:09:13.569 --> 00:09:17.202 a community approach in which we can take this technology, use the culture of the world 87 00:09:17.202 --> 00:09:27.790 wide web and let it change everything. >> ANNE WOOTTON: Hi, everyone. I'm Anne, and 88 00:09:27.790 --> 00:09:34.220 I'm one of the co founders of Pop Up Archive, and I think if there's one take away you get 89 00:09:34.220 --> 00:09:38.790 from my remarks today, it should be that I would like to offer all of your moms an API 90 00:09:38.790 --> 00:09:44.410 tutorial, and I think that any of these computational, scalable projects that are worth funding should 91 00:09:44.410 --> 00:09:48.611 do the same, even if they're not able to give you or your mom the tutorial themselves, they 92 00:09:48.611 --> 00:09:55.519 should have extensive and interactive documentation so you can teach yourselves. Some quick background 93 00:09:55.519 --> 00:09:59.949 on pop up archive, and I think the reason I'm sharing my experience right now is because 94 00:09:59.949 --> 00:10:07.360 of the work that we're doing specifically through a hybrid business model to help scale 95 00:10:07.360 --> 00:10:14.100 both preservation of digital audio, we work with disparate sources from individuals to 96 00:10:14.100 --> 00:10:20.420 very large institutions and partner with the Internet archive in doing that, and then also, 97 00:10:20.420 --> 00:10:26.220 accessibility, in part through computational creation of meta data, specifically automatically 98 00:10:26.220 --> 00:10:34.560 generated transcripts and semantic analysis of audio and speech. So, audio/visual collections 99 00:10:34.560 --> 00:10:41.120 have received a lot less attention to date than text and image based ones, understandably, 100 00:10:41.120 --> 00:10:47.850 and fundamental improvements in computational capacity are definitely changing that, but 101 00:10:47.850 --> 00:10:52.519 it's really just the tip of the iceberg. I could talk all afternoon, and I will touch 102 00:10:52.519 --> 00:10:59.279 briefly on some of the recent significant improvements in speech detect software, for 103 00:10:59.279 --> 00:11:04.360 example, but I don't think that my role here today is to share all of those details. I'm 104 00:11:04.360 --> 00:11:09.589 happy to talk about it more, you know, offline, or online. 105 00:11:09.589 --> 00:11:14.519 So, you know, significant improvements in speech detect software that we see already 106 00:11:14.519 --> 00:11:20.350 being applied commercially, whether it's through voice recognition and command, like Siri, 107 00:11:20.350 --> 00:11:24.769 or medical transcription, which Nuance is a major commercial provider of. There are 108 00:11:24.769 --> 00:11:31.149 trade offs, of course, between accuracy and search. One of our engineers wrote a blog 109 00:11:31.149 --> 00:11:36.200 post on the trade offs between accuracy and search for public media, if you're curious 110 00:11:36.200 --> 00:11:41.899 about that, and then the potential for this technology when we apply it to large AV collections 111 00:11:41.899 --> 00:11:48.259 that have previously only been searchable at best via limited meta data and cataloging. 112 00:11:48.259 --> 00:11:53.569 On that note, I would like to mention the importance of combining computational approaches, 113 00:11:53.569 --> 00:12:01.670 whether speech to text or other wave form analysis, with crowd sourcing efforts like 114 00:12:01.670 --> 00:12:05.170 meta data games and tilt factory, which I'm sure Mary will tell you more about with her 115 00:12:05.170 --> 00:12:11.410 work at Dartmouth, so that is automation plus human intelligence, which is critical, and 116 00:12:11.410 --> 00:12:16.290 then BBC's work on the world service archive is a really great example of there are open 117 00:12:16.290 --> 00:12:22.730 source components that they're slowly releasing as a part of that work, but more fundamentally, 118 00:12:22.730 --> 00:12:29.600 we built pop up archive because everyone at small and medium sized organizations was asking 119 00:12:29.600 --> 00:12:34.319 for drag and drop. They wanted something that was as easy to use as Facebook, they wanted 120 00:12:34.319 --> 00:12:38.589 it to integrate with the commercial cloud storage they were already using, and they 121 00:12:38.589 --> 00:12:42.910 wanted to be able to tweet it with the click of a button. So, we had already spent a year 122 00:12:42.910 --> 00:12:48.009 at the UC Berkeley School of Information, my co founder Bailey and I, doing user research 123 00:12:48.009 --> 00:12:54.180 and surveying the broad ecosystem of archival audio. At the same time, bigger organizations 124 00:12:54.180 --> 00:13:00.170 want to handle this computational work themselves, and they have more complex access policies 125 00:13:00.170 --> 00:13:04.750 and rights considerations and should absolutely still be able to benefit from taxpayer funded 126 00:13:04.750 --> 00:13:10.350 projects. So, it requires a very thoughtful balance of outputs, which in my experience 127 00:13:10.350 --> 00:13:15.730 so far, is frankly very hard to strike. So, for this reason, in terms of criteria 128 00:13:15.730 --> 00:13:20.060 and what needs to happen, I would highlight, as many others have today, the importance 129 00:13:20.060 --> 00:13:27.810 of interoperability, accessibility, reusability and educational resources, and this is generally 130 00:13:27.810 --> 00:13:32.910 where API's come into play. More specifically, successful computational projects need to 131 00:13:32.910 --> 00:13:40.319 take into account international standards for AV collections, whether meta data schemers 132 00:13:40.319 --> 00:13:46.611 on the AV side, which are also leveraging linked data to talk to each other, integrate 133 00:13:46.611 --> 00:13:53.120 with national databases, like the DPLA, linked data efforts like what Jon just talked about, 134 00:13:53.120 --> 00:13:57.769 and then arguably one of the most important roles we have played is as a straight forward 135 00:13:57.769 --> 00:14:04.190 web interface instead of educational resources that help producers and archivists ultimately 136 00:14:04.190 --> 00:14:11.250 benefit from research that would sit dormant indefinitely in an open source re-po hub. 137 00:14:11.250 --> 00:14:15.129 The Internet archives TV NEWS archives also comes to mind as a good example of that type 138 00:14:15.129 --> 00:14:20.120 of project. We talked a lot about the various merits and drawbacks as software as a service 139 00:14:20.120 --> 00:14:31.660 versus downloadable open source packages. We work with a variety of these technologies, 140 00:14:31.660 --> 00:14:36.380 and I'll pass on the detail there because we don't have a lot of time, but our experience 141 00:14:36.380 --> 00:14:41.290 is that bringing these research ideas to broad implementation requires a combination of these 142 00:14:41.290 --> 00:14:45.399 technologies, and ultimately, what we're up against, not surprisingly, is that regardless 143 00:14:45.399 --> 00:14:51.060 of size, the organizations that stand to benefit from this type of scaling are limited in terms 144 00:14:51.060 --> 00:14:55.879 of their resources and their aversion to risk. So, I think it's pretty standard for government 145 00:14:55.879 --> 00:14:59.559 funded projects to be open data projects, but there's a big spectrum in terms of open 146 00:14:59.559 --> 00:15:04.649 data that museums and libraries and archives can use and open data that they don't even 147 00:15:04.649 --> 00:15:15.290 have the resources to handle. Small and medium sized organizations don't even have, you know, 148 00:15:15.290 --> 00:15:19.810 they can barely setup their own my sequel database to run Omecca, but then bigger institutions 149 00:15:19.810 --> 00:15:24.959 have all kinds of rules about firewalls and access. The DPLA is doing some really great 150 00:15:24.959 --> 00:15:29.709 work to review copyright policies and publish findings, and we need that work desperately, 151 00:15:29.709 --> 00:15:34.209 because no matter how much we're capable of computational, it hardly matters if no one 152 00:15:34.209 --> 00:15:42.100 is ever going to hear it or see it. So, I'll just close by saying to get these grant funded 153 00:15:42.100 --> 00:15:46.559 projects from research to implementation, it requires resources and vision and will 154 00:15:46.559 --> 00:15:52.329 to act, and in our experience with media organizations in particular, that often means connecting 155 00:15:52.329 --> 00:15:58.029 the potential and vitality of the archive with the immediate pressing revenue concerns 156 00:15:58.029 --> 00:16:07.139 of the leadership, and that is really is a question of technology. 157 00:16:07.139 --> 00:16:09.160 >> MARY FLANAGAN: It's great to follow up with Anne. It's like we're having this very 158 00:16:09.160 --> 00:16:16.139 intense conversation also in our own minds. I'm really pleased to be here representing 159 00:16:16.139 --> 00:16:25.559 the Meta Data Games project which is how I ended up as a designer. Also, the work that 160 00:16:25.559 --> 00:16:30.399 we're doing on a national level building a crowd sourcing consortium, a conversation 161 00:16:30.399 --> 00:16:35.589 about what we can actually do to implement, to come up with best practices and somehow 162 00:16:35.589 --> 00:16:40.779 bridge some of these challenges together. So, folks on the panel, in one way or another, 163 00:16:40.779 --> 00:16:45.209 are really deeply engaged with creating these practical solutions, for engaging with the 164 00:16:45.209 --> 00:16:49.339 public, for making systems that are smart, for making lightweight systems that can appeal 165 00:16:49.339 --> 00:16:55.600 to a variety of institutions, and that was one of the big motivators behind making meta 166 00:16:55.600 --> 00:17:03.180 data games, which is a software platform that actually de couples game interfaces from the 167 00:17:03.180 --> 00:17:09.530 core software architecture so that we can actually rescan front end games and put games 168 00:17:09.530 --> 00:17:15.390 out into the public that solve a variety of challenging issues, such as collecting meta 169 00:17:15.390 --> 00:17:22.270 data tags, verifying or correcting in the case of our work with the bio heritage diversity 170 00:17:22.270 --> 00:17:27.760 library, correcting transcriptions that conflict with one another or that are run through OCR 171 00:17:27.760 --> 00:17:31.929 and have errors, so you can use game interfaces to engage with people in a variety of ways, 172 00:17:31.929 --> 00:17:36.279 and this is all, of course, inspired by Luis von Ahn and his great work with recapture 173 00:17:36.279 --> 00:17:44.159 stuff and Google image search, but we're really trying to keep using these things for an open 174 00:17:44.159 --> 00:17:49.760 source platform. So, the platform is designed to kind of have people make their own game 175 00:17:49.760 --> 00:17:58.520 interfaces as well and own their own data. To date, the project served 45 different collections, 176 00:17:58.520 --> 00:18:03.970 so instead of serving a silo, it's moving across institutions, including work with the 177 00:18:03.970 --> 00:18:14.809 British library, the Holocaust museum, so there's this way we're trying to work across 178 00:18:14.809 --> 00:18:22.140 institutions, and trying to meet the demand of scale is really interesting. Dartmouth 179 00:18:22.140 --> 00:18:41.789 gave us some funding to do some Amazon cloud testing, and we found some really interesting 180 00:18:41.789 --> 00:18:46.630 stuff. Most of the organizations we're working with want us to host their data. 181 00:18:46.630 --> 00:18:51.350 The system is designed so that the data can live anywhere, right, in content hubs, but 182 00:18:51.350 --> 00:18:57.909 actually, people really didn't want their content data to actually merge with, you know, 183 00:18:57.909 --> 00:19:02.470 they want to keep their data kind of pristine and separate, so when it comes to what needs 184 00:19:02.470 --> 00:19:08.260 to happen to make bigger impact, I think in some ways, we need to buy in and trust some 185 00:19:08.260 --> 00:19:16.120 of the techniques to make really strong and accurate information, and also, inclusive 186 00:19:16.120 --> 00:19:21.870 information. It's been touched on earlier today about the way that archives can be democratized 187 00:19:21.870 --> 00:19:27.140 through participation and community engagement, and I think crowd sourcing is a really amazing 188 00:19:27.140 --> 00:19:32.039 way to do this, but we have to let it happen, and we have to figure out how that goes. So, 189 00:19:32.039 --> 00:19:43.610 because of that project and other conversations with Bob Horton, I kind of took it on to start 190 00:19:43.610 --> 00:19:51.440 escalating this conversation, and, so, creating a, in summer of 2014, began to work on using 191 00:19:51.440 --> 00:19:59.900 an IMLS forum award to create the crowd sourcing consortium of libraries and archives, the 192 00:19:59.900 --> 00:20:09.410 SCCLA. So, we've done a number of things in this space. We conducted two national environmental 193 00:20:09.410 --> 00:20:13.360 scans. You may have, some people in the room have probably filled out one of our surveys, 194 00:20:13.360 --> 00:20:16.559 aimed at libraries and archives to understand the field. 195 00:20:16.559 --> 00:20:22.769 So, the first survey with 350 participants found that over 70 percent of the respondents 196 00:20:22.769 --> 00:20:28.660 thought that a lack of technical expertise was a barrier to try crowd sourcing in institutions. 197 00:20:28.660 --> 00:20:33.640 All this stuff is on crowdconsortium.org. It's really interesting information about 198 00:20:33.640 --> 00:20:38.570 an environmental scan. The second survey uncovered perceived barriers to crowd sourcing more 199 00:20:38.570 --> 00:20:45.830 deeply, and that was about 160 participants. In that survey, less than 10 percent of the 200 00:20:45.830 --> 00:20:51.760 participants reported that their institution had incorporated linked open data, so this 201 00:20:51.760 --> 00:20:57.549 is, and the numbers were low about projecting into the future what was going to happen with 202 00:20:57.549 --> 00:21:08.760 that. So, we also did some webinars with crowd sourcing 101, and did some webinars, 203 00:21:08.760 --> 00:21:12.670 and they were quite well attended. People are interested. They want to know about how 204 00:21:12.670 --> 00:21:17.850 to do this stuff, what it would take to scale, and, so, next week, we have planned a culminating 205 00:21:17.850 --> 00:21:23.440 meeting of this project where we have cross institutional support as Brett had mentioned. 206 00:21:23.440 --> 00:21:34.070 The University of Maryland is hosting, and Dartmouth will host engaging the public. So, 207 00:21:34.070 --> 00:21:42.200 there are probably about 50 people, but we also, and we'll be webcasting that too, so 208 00:21:42.200 --> 00:21:48.519 it'll be available, same kind of thing, and so it's really to uncover best practices, 209 00:21:48.519 --> 00:21:52.640 and we're doing a facilitative conversation with a design firm who will really help the 210 00:21:52.640 --> 00:21:58.480 communities pull out and actually get some work done in a practical way. There are a 211 00:21:58.480 --> 00:22:01.769 lot of makers, there are people who really want to try a new way of thinking, so we're 212 00:22:01.769 --> 00:22:09.720 working on that way to figure out how we can think about engaging the public and consolidate 213 00:22:09.720 --> 00:22:13.880 our national knowledge. So, I like the idea that was brought up earlier of what are our 214 00:22:13.880 --> 00:22:18.950 top 10 points. Maybe we can also incorporate that into this conversation since these are 215 00:22:18.950 --> 00:22:26.470 happening so closely. So, I think that's all I need to say right now. We'll talk more about 216 00:22:26.470 --> 00:22:32.400 things in the Q & A. >> TREVOR MUNOZ: Well, thank you to IMLS for 217 00:22:32.400 --> 00:22:38.320 inviting me. It's a really almost impossible task to try to follow these three great people 218 00:22:38.320 --> 00:22:43.799 on this panel, and when we were planning this panel and our phone conversation about it, 219 00:22:43.799 --> 00:22:47.900 we were kind of trying to figure out, you know, kind of the relationship of what I was 220 00:22:47.900 --> 00:22:52.200 going to talk about to the three things we've heard about already, because part of why I 221 00:22:52.200 --> 00:22:56.640 ended up on this panel was to talk a little bit about data curation as a practice that 222 00:22:56.640 --> 00:23:02.299 we might scale up, and I think through that lens, data curation is really interesting, 223 00:23:02.299 --> 00:23:11.580 because I think it's less yet another set of protocols or tools or methodologies and 224 00:23:11.580 --> 00:23:16.640 more something that sits in relationship or that's the strategy that helps us think about 225 00:23:16.640 --> 00:23:21.380 how we deploy all the things we've already heard about, computational techniques and 226 00:23:21.380 --> 00:23:25.331 crowd sourcing and linked open data, and then I think in this regard, it's most useful to 227 00:23:25.331 --> 00:23:32.850 think of data curation as kind of our strategic guide for how we, as cultural heritage organizations 228 00:23:32.850 --> 00:23:37.450 specifically, you know, deploy and interact with some of these, the things that we've 229 00:23:37.450 --> 00:23:46.289 already heard about, because I think the interesting role that a sort of data curation focus plays 230 00:23:46.289 --> 00:23:51.830 in this conversation around the national digital platform is that it asks us to think both 231 00:23:51.830 --> 00:23:59.460 about repeatable units in that kind of business and technology sense of scale that we heard 232 00:23:59.460 --> 00:24:04.120 a little bit about this morning, right? If part of curation is preservation, then we 233 00:24:04.120 --> 00:24:08.570 need to get lean and fast and efficient when we're collecting material from the web or 234 00:24:08.570 --> 00:24:12.940 we're collecting audio/visual material and making sure that that lasts, but I think what's 235 00:24:12.940 --> 00:24:18.909 interesting is preservation being a component of curation, right? We can never actually 236 00:24:18.909 --> 00:24:22.490 get to curation by just collecting more stuff, right? 237 00:24:22.490 --> 00:24:25.880 That preservation is an important part of our mission, but that curation asks us to 238 00:24:25.880 --> 00:24:32.789 do that and to think differently about the materials that we're bringing in, and I think 239 00:24:32.789 --> 00:24:37.769 that's what makes it really interesting and perhaps what might shape how we think about 240 00:24:37.769 --> 00:24:43.760 data curation projects as part of the national digital platform funding priority. In other 241 00:24:43.760 --> 00:24:49.809 words, a lot of the things that we're seeing so far are about helping institutions participate 242 00:24:49.809 --> 00:24:56.289 in the platform to bring stuff to the national digital aggregation or aggregations, and I 243 00:24:56.289 --> 00:25:01.750 would also like us to see proposals and funded projects that are focusing on, you know, in 244 00:25:01.750 --> 00:25:07.990 parallel with that important task, working with communities of users, you know, the absence 245 00:25:07.990 --> 00:25:12.750 of our sort of end users has come up today already, working with communities of those 246 00:25:12.750 --> 00:25:17.990 users to be making sort of active reuse of those materials in parallel with all the sort 247 00:25:17.990 --> 00:25:22.769 of coordinated institutional work that we're doing to bring it in and preserve it. So, 248 00:25:22.769 --> 00:25:27.169 just a couple quick examples of what that might look like, I have a colleague at University 249 00:25:27.169 --> 00:25:34.440 of Maryland who is really interested in virtual reunification of image archives, and, you 250 00:25:34.440 --> 00:25:37.900 know, to see work like that of going, all right, we're bringing materials into these 251 00:25:37.900 --> 00:25:44.450 large national collections, can we also be working at the same time with people who want 252 00:25:44.450 --> 00:25:49.520 to begin to tend those collections at the national level, at that large scale, to think 253 00:25:49.520 --> 00:25:56.830 about the shapes of those collections, about their gaps, so that not only the small and 254 00:25:56.830 --> 00:26:02.340 excellent team at DPLA or at Haughty Trust is charged with enhancing the meta data or 255 00:26:02.340 --> 00:26:06.460 the findability of those materials, but that we also have outside people partnering with 256 00:26:06.460 --> 00:26:11.700 them who have their own investment of the material to also be adding meta data or adding 257 00:26:11.700 --> 00:26:16.350 description, adding value. You know, how do we have projects that bring these meta data 258 00:26:16.350 --> 00:26:23.559 mobs to DPLA and Haughty Trust sooner rather than later, because I think it would be a 259 00:26:23.559 --> 00:26:26.760 mistake to think about staging these sequentially, right? 260 00:26:26.760 --> 00:26:30.300 To get all the stuff together and then worry about, you know, how people are going to be 261 00:26:30.300 --> 00:26:35.899 making use of it, but I think curation urges us, as we think about our active management, 262 00:26:35.899 --> 00:26:40.539 to think about inviting those people in who are saying they want to work on materials 263 00:26:40.539 --> 00:26:46.659 that are already in our national digital platform and begin to enrich them in perhaps more idiosyncratic 264 00:26:46.659 --> 00:26:53.820 ways once they're already there. So, the image reunification is one project. I think seeing 265 00:26:53.820 --> 00:26:59.690 more work around round tripping data back to something like the national digital platform, 266 00:26:59.690 --> 00:27:04.539 right, it's there and I get it and I enhance it, how do I give it back? That's a huge challenge 267 00:27:04.539 --> 00:27:12.860 that we really haven't graveled with. I think, you know, this has come up, you know, getting 268 00:27:12.860 --> 00:27:17.779 serious about, you know, the usability and collection development and sessioning around 269 00:27:17.779 --> 00:27:22.860 our web archives, that that would be another area we could be working on right away. So, 270 00:27:22.860 --> 00:27:29.990 those are some near term priorities that I think we might bring into the conversation. 271 00:27:29.990 --> 00:27:35.330 I think I'll stop there and we can go to discussion. >> DEANNA MARCUM: Great. Thank you all for 272 00:27:35.330 --> 00:27:48.590 really interesting comments. We'll take questions in just a minute. How hard has it been for 273 00:27:48.590 --> 00:28:00.309 you in your individual projects to work with libraries, archives, museums? 274 00:28:00.309 --> 00:28:11.279 >> TREVOR MUNOZ: It's awesome. I guess I'll just say this, and this is about, data curation 275 00:28:11.279 --> 00:28:18.059 is a strategy, right? I work in a library, so, you know, it's great to work with libraries 276 00:28:18.059 --> 00:28:23.770 and archives and museums, but I think, you know, a sense of changing the priorities of 277 00:28:23.770 --> 00:28:29.399 what we spend more of our time on relative to other things has been a challenging conversation, 278 00:28:29.399 --> 00:28:36.159 you know, and having leadership in these organizations, sort of helping shape this difficult conversation 279 00:28:36.159 --> 00:28:40.221 of maybe there are things we need to stop doing so that we can do other things, that's 280 00:28:40.221 --> 00:28:42.929 something I guess I'd like to see more of that would make my life easier. 281 00:28:42.929 --> 00:28:46.259 >> DEANNA MARCUM: Good. >> ANNE WOOTTON: I can say, and I may be digging 282 00:28:46.259 --> 00:28:52.110 my grave even deeper here, but you can add public media organizations. What I'm going 283 00:28:52.110 --> 00:28:58.130 to say, you know, it really depends. Can it be tough and frustrating? Absolutely, but 284 00:28:58.130 --> 00:29:06.620 I was going to mention this actually, there are internal silos, so it, you know, it really 285 00:29:06.620 --> 00:29:11.700 depends which group you're working with, which person you're working with, and part of the 286 00:29:11.700 --> 00:29:16.510 take away's that I've had over the last few years are that the outputs of these projects 287 00:29:16.510 --> 00:29:22.500 need to be understandable and parsible by the internal advocates at those organizations, 288 00:29:22.500 --> 00:29:28.240 or even the people who may be, like, on the verge, tippable, but they need to see, like, 289 00:29:28.240 --> 00:29:32.930 just what they can leverage and benefit from and what that implementation will entail, 290 00:29:32.930 --> 00:29:38.870 and then it can really, you know, open all kinds of doors, because everybody's learning 291 00:29:38.870 --> 00:29:43.379 and it is a question of what the learning curve is and how much you can expect to educate. 292 00:29:43.379 --> 00:29:52.669 >> JON VOSS: I'd say just it varies so greatly by size and the facilities that they have 293 00:29:52.669 --> 00:29:57.480 and the tech resources that they have, person resources, as a lot of people have called 294 00:29:57.480 --> 00:30:02.779 out. At Historypin, we have very different strategies with small and medium institutions. 295 00:30:02.779 --> 00:30:09.950 They can be historical societies, but they have a great deal of access to the social 296 00:30:09.950 --> 00:30:13.470 good, because they're so connected in the local community, but on the technical side, 297 00:30:13.470 --> 00:30:17.889 obviously, we're dealing with big challenges. On the other side of it, on the national/international 298 00:30:17.889 --> 00:30:22.740 aggregators, we can start talking about the linked data. It's very different. 299 00:30:22.740 --> 00:30:29.450 >>MARY FLANAGAN: Yeah, I guess I would say, as a, you know, working outside of library 300 00:30:29.450 --> 00:30:35.419 and institutional archives, the meta data games actually emerged out of collaboration 301 00:30:35.419 --> 00:30:43.809 with an archivist at Dartmouth, and originally, I hope he's not in trouble if I say this, 302 00:30:43.809 --> 00:30:47.290 he said just stay away from the library for awhile, stay away from the archive, think 303 00:30:47.290 --> 00:30:56.870 outside of it. It's by being an outsider in some ways that his argument was you're going 304 00:30:56.870 --> 00:31:03.070 to see this stuff differently, and I would agree with that. I think it's, you know, again, 305 00:31:03.070 --> 00:31:09.410 I'm speaking as an outsider, but I would agree that it's been really helpful, and I've had 306 00:31:09.410 --> 00:31:14.570 nothing but really strong interest in the project, but most institutions do seem to 307 00:31:14.570 --> 00:31:18.639 feel overwhelmed about trying something new when they feel like there are many, many things 308 00:31:18.639 --> 00:31:23.809 to catch up, and they can't necessarily decide which is going to get them ahead faster. 309 00:31:23.809 --> 00:31:38.240 >> DEANNA MARCUM: Let's open it up to the audience for your questions, comments. Did 310 00:31:38.240 --> 00:31:45.509 lunch put you to sleep? >>QUESTION FROM AUDIENCE: I guess I'll say 311 00:31:45.509 --> 00:31:52.440 something. Hi. Ben Virshbaum, New York Public Library and Web PL Labs. Been working with 312 00:31:52.440 --> 00:31:57.049 most of you folks for years on all this stuff, and, yeah, so many great points have been 313 00:31:57.049 --> 00:32:01.639 raised. I think that, yeah, we find ourself in an interesting point, with crowd sourcing 314 00:32:01.639 --> 00:32:06.470 especially. Where, I come from an institution where we did have an appetite to try it and 315 00:32:06.470 --> 00:32:12.190 see if it could be done and there was traction with users, and we now have produced some 316 00:32:12.190 --> 00:32:17.950 significant corps of data that are messy but interesting or not even that messy in some 317 00:32:17.950 --> 00:32:24.879 cases but in some ways curatorial interest has been reprioritized or moved on but public 318 00:32:24.879 --> 00:32:29.580 interest has not so I think and there is this connotention of going after the trend and 319 00:32:29.580 --> 00:32:34.750 going after the long distance run that this kind of data curation, data stewardship implies 320 00:32:34.750 --> 00:32:41.070 and requires for it to really have the impact we all want it to have. Trevor has been extending 321 00:32:41.070 --> 00:32:46.100 that chain of custody with a colleague at University of Pennsylvania, Katy R. that of 322 00:32:46.100 --> 00:32:49.630 curating the menus data from what is in the menu transcription project, and now we are 323 00:32:49.630 --> 00:32:54.529 trying to think how do we get that round trip, how do we get that back, how do we build a 324 00:32:54.529 --> 00:33:01.250 new data infrastructure where we can attach these new forms of data to core records, that 325 00:33:01.250 --> 00:33:08.070 integration question is so important, so this is really just comments. I think Clifford 326 00:33:08.070 --> 00:33:14.320 Lynch spoke about the need to be able to message the public more effectively in a more unified 327 00:33:14.320 --> 00:33:19.620 way about these efforts I think there is a bigger narrative about this kind of work that 328 00:33:19.620 --> 00:33:23.929 we need to work on developing. There is a call to action involved but also what can 329 00:33:23.929 --> 00:33:28.700 we do. This is a great data migration project we are all involved in, I have a lot of metaphors 330 00:33:28.700 --> 00:33:33.790 in mind. Community gardening, but that participation in the act of stewardship of the forward migration 331 00:33:33.790 --> 00:33:39.920 of the data of these collections develops incredible bonds with the public and raises 332 00:33:39.920 --> 00:33:44.090 awareness of these materials, so it does a lot of good with these missions. Trevor has 333 00:33:44.090 --> 00:33:49.379 written eloquently on this before. So it think that that messaging and the narrative of that 334 00:33:49.379 --> 00:33:56.549 work of moving data. It sounds geeky and inaccessible at first but it is translated into very grand 335 00:33:56.549 --> 00:34:06.570 public service narratives that we can develop. And to call back to the mix of human and automated 336 00:34:06.570 --> 00:34:13.110 approaches is really powerful, and that is well established with OCR, with text recognition 337 00:34:13.110 --> 00:34:19.190 and the cleanup. we are seeing very interesting developments in the audio visual space, going 338 00:34:19.190 --> 00:34:23.960 from messy transcripts to with text recognition, to targeted extracted entities, and the cleanup, 339 00:34:23.960 --> 00:34:30.130 with cleaner meta data. We are doing some work with historical maps with the same premise 340 00:34:30.130 --> 00:34:35.850 where we are doing a good first pass on show pass of maps, urban scale maps getting kind 341 00:34:35.850 --> 00:34:40.990 of building shapes and footprints and then breaking the quality control and validation 342 00:34:40.990 --> 00:34:45.190 into smaller pieces. So, I think these are all really important elements, the human computer 343 00:34:45.190 --> 00:34:50.700 collaboration, the narrative of this work, and that figuring out the long distance kind 344 00:34:50.700 --> 00:34:56.380 of stamina required to really see these things through, because I think we're well past now 345 00:34:56.380 --> 00:35:02.400 the question of do people want to do it or will they even participate. They will, and 346 00:35:02.400 --> 00:35:07.250 I think more could. It could scale further. So, those are just a couple of reflections. 347 00:35:07.250 --> 00:35:10.320 Wonderful presentations. Thank you. >> DEANNA MARCUM: Thanks. Comments? 348 00:35:10.320 --> 00:35:16.130 >>TREVOR MUNOZ: Well, just one follow up comment on that. I think the insight of people who 349 00:35:16.130 --> 00:35:21.750 have worked deeply in crowd sourcing is the importance of having the community of people 350 00:35:21.750 --> 00:35:26.750 who want to do something, you know, with the things that are produced from crowd sourcing, 351 00:35:26.750 --> 00:35:31.110 right? They have a research project, whether we're talking about the galaxy folks, where 352 00:35:31.110 --> 00:35:37.380 they're trying to identify galaxies, or as we work with the menus data, part of cleaning 353 00:35:37.380 --> 00:35:41.200 it up is actually figuring out, you know, what matters to the people who are trying 354 00:35:41.200 --> 00:35:46.990 to do some kind of scholarly, or sort of who are hobbyists or enthusiasts about it, what 355 00:35:46.990 --> 00:35:52.870 matters to them, so that community of interest, being there and working with that community 356 00:35:52.870 --> 00:35:58.830 and sort of cultivating them is a real key part of that, and that goes beyond just the 357 00:35:58.830 --> 00:36:02.590 kind of mechanics of, you know, sharing data out and data in. 358 00:36:02.590 --> 00:36:08.220 >>MARY FLANAGAN: I think we can also raise the profile and help with this narrative that 359 00:36:08.220 --> 00:36:14.520 Ben speaks of if we're also, you know, focusing some time on the interoperability and the 360 00:36:14.520 --> 00:36:21.140 round tripping and making sure that organizations can trust, oh, wow, I've seen the map, I understand 361 00:36:21.140 --> 00:36:26.770 the loop, and it's working, and it can work for other groups, and once we can start to 362 00:36:26.770 --> 00:36:31.170 unite the projects, I think that's another way that that can become visible. 363 00:36:31.170 --> 00:36:38.110 >>DEANNA MARCUM: And it certainly helps with the issue of scale, because we'll never be 364 00:36:38.110 --> 00:36:43.410 able to do all of these things if we try to do them ourselves, right? We have to open 365 00:36:43.410 --> 00:36:50.811 it up to others. >>JON VOSS: Ben, I really liked your point 366 00:36:50.811 --> 00:36:58.860 about the timeframes. I think that's a huge bit. We started to consider knowledge communities, 367 00:36:58.860 --> 00:37:05.480 right? So, these were enthusiasts or people who know a location really well, etc. It takes 368 00:37:05.480 --> 00:37:10.640 time for them to trust you, for starters, and then there's also this idea of co creating, 369 00:37:10.640 --> 00:37:15.930 so kind of adding this design theory to not only how we do things but how the community 370 00:37:15.930 --> 00:37:20.850 does things with us, and that's a matter of trust, it's a matter of building their confidence, 371 00:37:20.850 --> 00:37:25.830 but it is definitely a longer term timeframe than a two year grant. So if you get these 372 00:37:25.830 --> 00:37:32.010 grants and you are about to write the reports, now you are ready to get busy with the community. 373 00:37:32.010 --> 00:37:34.331 >>DEANNA MARCUM: Thanks. >>QUESTION FROM AUDIENCE: I want to see what 374 00:37:34.331 --> 00:37:38.610 this group has to say about some other issues than what you've talked about. I think it's 375 00:37:38.610 --> 00:37:45.530 interesting that our panel is primarily addressing things that I would say most are not completely 376 00:37:45.530 --> 00:37:52.210 core yet. We are interested in both those things, trying very hard to work in that space, 377 00:37:52.210 --> 00:37:57.180 but it is still somewhat less core than other things like collective research volumes or 378 00:37:57.180 --> 00:38:07.010 circulating, so one question I have is you have thought about how to scale these, perhaps 379 00:38:07.010 --> 00:38:12.550 becoming core library work, what lessons have you learned that you think might apply to 380 00:38:12.550 --> 00:38:17.640 also thinking about approaches to scaling and sharing other perhaps, you know, even 381 00:38:17.640 --> 00:38:25.840 more fundamental or essential functions of things that, to me, scream out, things like 382 00:38:25.840 --> 00:38:31.900 account management, to have, you know, best in class user accounts and best in class privacy 383 00:38:31.900 --> 00:38:35.890 around that, things like shared discovery layers or tools for creating your own discovery 384 00:38:35.890 --> 00:38:40.370 layer. Are there lessons you've learned from your work to scale these projects that could 385 00:38:40.370 --> 00:38:51.010 be applied in those other problems? >> JON VOSS: Well, a lot of this, for Historypin 386 00:38:51.010 --> 00:38:57.410 particularly, is looking at, you know, increasingly, just design thinking around users, and I think 387 00:38:57.410 --> 00:39:04.090 that's been huge. I don't have a great concept of how things work behind the scenes at libraries. 388 00:39:04.090 --> 00:39:17.160 You know, is there a way you can tell us what you have, tell us how 389 00:39:17.160 --> 00:39:24.970 we can use it, and I think, in the last five years of working on this, it has been, it 390 00:39:24.970 --> 00:39:29.190 started with this is obviously what you need and this is how you use it, and now we're 391 00:39:29.190 --> 00:39:33.610 really going back and redesigning everything and thinking, okay, how is it that you work 392 00:39:33.610 --> 00:39:37.721 now and understand these work flows. A lot of the work that you guys are doing, you think 393 00:39:37.721 --> 00:39:43.190 of partnerships with D schools and design, kind of really looking at that element is 394 00:39:43.190 --> 00:39:51.160 a huge part of how we should approach these problems with the user in mind. 395 00:39:51.160 --> 00:40:00.630 >> ANNE WOOTTON: Successful complex systems are essentially much more simple, and to start 396 00:40:00.630 --> 00:40:08.580 building something complex from scratch, it's almost doomed to fail, and the reason pop 397 00:40:08.580 --> 00:40:14.850 up art exists is because Bailey and I spent a year plus in graduate school at a school 398 00:40:14.850 --> 00:40:20.590 of information which was very focused on design thinking and iterative software development, 399 00:40:20.590 --> 00:40:26.260 and it was totally new to both of us. We were humanities people, former journalists and 400 00:40:26.260 --> 00:40:32.640 didn't even realize just how steeped in it we were until we left, but having that time 401 00:40:32.640 --> 00:40:37.450 to go out and do user research, and, I mean, I'm finding surveys that we did in 2011 that 402 00:40:37.450 --> 00:40:41.800 I don't even remember that are still relevant and informative to the work that we're doing 403 00:40:41.800 --> 00:40:49.380 now, and I think, you know, in the day to day, at most museums, libraries and archives, 404 00:40:49.380 --> 00:40:53.400 there just isn't time to do that kind of work, and when you're applying for grant funding, 405 00:40:53.400 --> 00:40:59.350 it's often the component that ends up kind of hitting the cutting room floor. I know 406 00:40:59.350 --> 00:41:04.560 there was some conversation about this earlier today too. I think it's really important, 407 00:41:04.560 --> 00:41:10.730 not just for documentation and maintenance to be a part of these projects, but also for 408 00:41:10.730 --> 00:41:16.460 user research and education to be a part of it too. 409 00:41:16.460 --> 00:41:20.080 >> TREVOR MUNOZ: I guess I want to pick at the question a little bit, because I feel 410 00:41:20.080 --> 00:41:32.060 like there are a lot of things that libraries are doing where they are trying to share some 411 00:41:32.060 --> 00:41:39.560 of these core things, a number of these activities, but I guess what I find that might be translatable 412 00:41:39.560 --> 00:41:45.230 is a lot of those activities, whether it's, you know, the tools around integrated library 413 00:41:45.230 --> 00:41:49.870 systems or shared print repositories or something, they feel, at the moment, like they're being 414 00:41:49.870 --> 00:41:58.190 done in a kind of back of the house, let's sort of keep the bother out of the site and 415 00:41:58.190 --> 00:42:03.160 our communities, and that, I think, is a different spirit than a lot of these things that aren't 416 00:42:03.160 --> 00:42:11.010 yet core but we want to become core. I mean, I think a lot of the case that many of us 417 00:42:11.010 --> 00:42:15.230 are making are about, you know, the excitement or, you know, the engagement that we can get 418 00:42:15.230 --> 00:42:19.530 out of doing this new thing, and that's a lot of the argument that we're making to our 419 00:42:19.530 --> 00:42:24.980 leadership, and I don't see that same kind of argument being made around some of these 420 00:42:24.980 --> 00:42:29.880 other really key important things, like, you know, what should the library discovery layers 421 00:42:29.880 --> 00:42:35.230 be like or our e book activities or shared print, so I wonder if that's the thing that 422 00:42:35.230 --> 00:42:42.000 needs to come across and that sort of thinking about, you know, how is this not a back of 423 00:42:42.000 --> 00:42:46.380 the office task but a part of our sort of engagement with our community of users and 424 00:42:46.380 --> 00:42:50.911 people that care passionately about all this stuff that we have. I don't know, that's a 425 00:42:50.911 --> 00:43:03.620 sort of half formed >>QUESTION FROM AUDIENCE: (Off mic.) 426 00:43:03.620 --> 00:43:14.420 >>TREVOR MUNOZ: Well, I mean, just, I think it's, to me, it resonates more with actually 427 00:43:14.420 --> 00:43:19.190 the point you were making about e books, right? Like, in a lot of sense, I think, you know, 428 00:43:19.190 --> 00:43:25.550 it's to not be dependent on vendors, but we haven't gone all the way through to saying, 429 00:43:25.550 --> 00:43:29.470 like, what would it mean for us to have sort of self determination in this area, right? 430 00:43:29.470 --> 00:43:33.730 Like, the ILS could become a powerful and exciting thing, that could be a bit much, 431 00:43:33.730 --> 00:43:41.490 but could be kind of a positive thing if we thought about it in terms of taking back ownership 432 00:43:41.490 --> 00:43:46.630 of our relationship with our communities of use and having that be sort of the central 433 00:43:46.630 --> 00:43:51.080 narrative of the project instead of just being, like, let's be another entrance into a marketplace. 434 00:43:51.080 --> 00:43:57.230 >>ANNE WOOTTON: I'm just going to qualify what I said earlier. When Pop Up Archive started, 435 00:43:57.230 --> 00:44:01.150 we never thought that machine transcription would even be a part of what we did, so we 436 00:44:01.150 --> 00:44:05.980 were taking people back to square one and saying fundamentally, for a lot of producers 437 00:44:05.980 --> 00:44:11.580 who found themselves archivists, what is your ideal relationship with this material that 438 00:44:11.580 --> 00:44:17.100 you're creating and becoming stewards of, whether you like it or not, that you recognize 439 00:44:17.100 --> 00:44:23.470 the value in, but fundamentally, you know, what relationship to this would work for you, 440 00:44:23.470 --> 00:44:27.391 how can we, we talk about making archiving part of the production work flow, right, and 441 00:44:27.391 --> 00:44:30.550 fundamentally, what should that look like, let's talk about that, let's ask a lot of 442 00:44:30.550 --> 00:44:36.470 people to draw their visions of what that might look like and then try to boil it down, 443 00:44:36.470 --> 00:44:41.610 and for us, the computational piece was, like, icing on the cake that has become something 444 00:44:41.610 --> 00:44:45.430 I'm happy to talk about but wasn't even why we started doing this. 445 00:44:45.430 --> 00:44:51.780 >>MARY FLANAGAN: Just really quickly, you know, in thinking about crowd sourcing and 446 00:44:51.780 --> 00:44:55.920 tagging, it seems like we're actually gathering data, and the original emphasis was to do 447 00:44:55.920 --> 00:45:03.170 that about archival material we didn't know much about, and that is still a really helpful 448 00:45:03.170 --> 00:45:08.980 thing, but I think what we're hearing over and over again as we iterate these projects 449 00:45:08.980 --> 00:45:13.550 is that we're really, software developers are really engineering the social and that 450 00:45:13.550 --> 00:45:20.140 maybe libraries, museums and archives are really far more social designed than they 451 00:45:20.140 --> 00:45:26.560 are artifact designed, and if that's the case, has anyone played trivia crack? Okay, thank 452 00:45:26.560 --> 00:45:35.190 you. Thank you. So, I was, all my students play trivia crack, it's an app, and I was 453 00:45:35.190 --> 00:45:38.970 just, this is an anecdote, I'm just going to go there though, so I was diagramming for 454 00:45:38.970 --> 00:45:42.820 my students in my game design class, okay, well, how does it work? It's a trivia game 455 00:45:42.820 --> 00:45:48.680 online. You think, oh, the majority of the task is about gathering questions, but users 456 00:45:48.680 --> 00:45:58.730 can submit trivia questions, and if you actually look at the ecosystem of trivia crack, about 457 00:45:58.730 --> 00:46:04.320 10 percent is really time spent with the trivia question, and that's not what you would think 458 00:46:04.320 --> 00:46:08.240 if you just saw the words, well, I don't know what you would think, but if you said I'm 459 00:46:08.240 --> 00:46:13.910 going to go play a trivia app, you would not think of it as 90 percent social engineering, 460 00:46:13.910 --> 00:46:18.861 right? And I think that this is something that just emerges more and more in the work 461 00:46:18.861 --> 00:46:24.260 that I've been doing. It becomes far more of a design challenge in the social sphere 462 00:46:24.260 --> 00:46:30.740 than the technological sphere, although the technology is, of course, in there, iteratively 463 00:46:30.740 --> 00:46:35.560 so. >>DEANNA MARCUM: Question down here? 464 00:46:35.560 --> 00:46:42.460 >>QUESTION FROM AUDIENCE: Michael again. Sorry. I want to throw the gauntlet down a little 465 00:46:42.460 --> 00:46:50.010 bit about scale. I wonder throughout the day and throughout this project if we're talking 466 00:46:50.010 --> 00:46:58.490 about an incremental change model or, I don't want to use the word radical, the way Google 467 00:46:58.490 --> 00:47:03.050 talks about a 10 times improvement model, and I have the feeling that we're thinking 468 00:47:03.050 --> 00:47:09.170 about scale collectively in a very small way. A lot of us were really profoundly struck 469 00:47:09.170 --> 00:47:20.120 when Ted served its billionth video in 2013. When we look at the traction, the profound 470 00:47:20.120 --> 00:47:28.410 user engagement you get on Reddit. When our curators appear on Reddit. It is many orders 471 00:47:28.410 --> 00:47:37.920 of magnitude, some of you have seen what I've been talking about with the vlog brothers, 472 00:47:37.920 --> 00:47:45.070 Hank and John Green, just tens of millions of people raising millions and millions of 473 00:47:45.070 --> 00:47:50.040 dollars, and I wonder if we're subconsciously designing a project that's just going to be 474 00:47:50.040 --> 00:47:55.150 a little bit better, sort of a digital version of the levels of scale that we get with traditional 475 00:47:55.150 --> 00:47:58.360 libraries, or if we're really thinking about, you know, within a decade, we're going to 476 00:47:58.360 --> 00:48:04.190 have seven or eight billion people online, the kinds of scale that could and should be 477 00:48:04.190 --> 00:48:12.160 designed with them in mind. Most of the next five billion are going to be from kind of 478 00:48:12.160 --> 00:48:18.710 marginal governments, most of them are going to be poor, most of them are going to be hungry 479 00:48:18.710 --> 00:48:25.291 for the kinds of resources that these organizations stand for. What are we going to do when they 480 00:48:25.291 --> 00:48:28.730 arrive? What are we going to do now so that when they arrive, there's something for them 481 00:48:28.730 --> 00:48:35.750 that's good? >> JON VOSS: I'll start. Thanks for that, 482 00:48:35.750 --> 00:48:43.410 Michael. Yeah, man, let's do it this way. I mean, the things, and I've sworn to myself 483 00:48:43.410 --> 00:48:48.670 I'm going to make it through without swearing once, so the stuff that Seb is doing that 484 00:48:48.670 --> 00:48:54.750 is so exciting to me is that if you close your building down for three years and you 485 00:48:54.750 --> 00:48:59.600 reinvent everything, what would you come up with, and that's the kind of stuff that people 486 00:48:59.600 --> 00:49:04.060 are coming up with, and what would happen if we closed our libraries down completely 487 00:49:04.060 --> 00:49:07.650 for three years? I mean, obviously, you wouldn't want to do this, but that's the kind of thinking 488 00:49:07.650 --> 00:49:11.901 where we could get it to scale, where we could start to say how do we build everything either 489 00:49:11.901 --> 00:49:15.870 around the content, around the people, what are the assets that we have, what does that 490 00:49:15.870 --> 00:49:21.460 look like. That's a very exciting conversation to have. Let's do it right now. 491 00:49:21.460 --> 00:49:27.210 >>DEANNA MARCUM: What are those features? >> JON VOSS: I don't know. I think to me, 492 00:49:27.210 --> 00:49:30.660 I would start by building something in the environment of the web. I mean, that's what 493 00:49:30.660 --> 00:49:35.010 I love about what Seb and Aaron and the whole team has done over there, is that the database 494 00:49:35.010 --> 00:49:40.150 was the key to everything. You could surface things, you could change things around, and 495 00:49:40.150 --> 00:49:44.120 I would say it wouldn't be just what I have in my own institution, but what everyone else 496 00:49:44.120 --> 00:49:48.200 has. So, if we all started collectively to go at it that way, now you've got something 497 00:49:48.200 --> 00:49:53.360 very different, and I would also say these institutions have huge social value, and it's 498 00:49:53.360 --> 00:50:00.580 been completely underestimated, and that's something that I would want to focus on as 499 00:50:00.580 --> 00:50:04.080 well, you know, what do they really do to strengthen the communities. Nobody really 500 00:50:04.080 --> 00:50:09.740 talks about libraries in that way, but that's a huge element so the evaluation part of that 501 00:50:09.740 --> 00:50:15.290 is a factor too. >>DEANNA MARCUM: Mary, are you ready to jump 502 00:50:15.290 --> 00:50:17.470 in? >>MARY FLANAGAN: Well, I think it's a great 503 00:50:17.470 --> 00:50:26.710 question, and it's always in a planning meeting, you know, how far can we go, right? The futurists 504 00:50:26.710 --> 00:50:34.510 have different visions of what 2030 will look like. So, I think, regardless of what kind 505 00:50:34.510 --> 00:50:38.461 of radical change we have, we know that there are certain pain points anyway, and that may 506 00:50:38.461 --> 00:50:46.220 be the place to focus, so that we can actually allow for, like, talking, having data be able 507 00:50:46.220 --> 00:50:54.400 to be free and talk to each other. We're going to need that no matter what happens, so being 508 00:50:54.400 --> 00:51:00.470 able to incorporate the public in a new way has to be a part of this conversation. So, 509 00:51:00.470 --> 00:51:05.470 I think that we have some really interesting seeds for that kind of radical conversation. 510 00:51:05.470 --> 00:51:14.960 >>ANNE WOOTTON: I want to be at the table for that conversation. I don't want to repeat 511 00:51:14.960 --> 00:51:20.150 myself, but I think we, you know, I like to frame our work often as, you know, what we 512 00:51:20.150 --> 00:51:25.990 can do to make audio findable by Google, right, and so I come at it, especially working in 513 00:51:25.990 --> 00:51:31.460 public media and, like, RSS feeds, this open protocol where we can go out and identify 514 00:51:31.460 --> 00:51:40.781 all of this media that is, you know, digital radio of today, and, you know, assess its 515 00:51:40.781 --> 00:51:46.781 value independently and incorporate it into the work that we do for that reason, and I 516 00:51:46.781 --> 00:51:51.990 think, yeah, you know, I'd love to be a part of that conversation about making it happen 517 00:51:51.990 --> 00:51:57.380 more systemically. >>TREVOR MUNOZ: I was reflecting on something 518 00:51:57.380 --> 00:52:04.150 that came up in the conversation earlier this morning about, you know, if we look at the, 519 00:52:04.150 --> 00:52:09.860 you know, the communities that many of the institutions that we work at consider their 520 00:52:09.860 --> 00:52:15.060 constituency, it's actually not a national constituency, it's something much more local 521 00:52:15.060 --> 00:52:23.230 to them, and there's a great, you know, sort of, there's a great inertia to stay, you know, 522 00:52:23.230 --> 00:52:28.390 to preserve the institution which serves that local constituency that provides our little 523 00:52:28.390 --> 00:52:34.550 local resources, and I don't know, maybe part of the way to break that open, you know, because 524 00:52:34.550 --> 00:52:38.220 I think it would be difficult for me, you know, to tell the state of Maryland, for instance, 525 00:52:38.220 --> 00:52:42.110 you know, like, you know, the state of Maryland will get some good stuff, but we're off on 526 00:52:42.110 --> 00:52:46.790 this national thing, you know, this is part of the problem that we all have in many of 527 00:52:46.790 --> 00:52:51.970 our institutions, and that part of perhaps what some of us are struggling toward here 528 00:52:51.970 --> 00:52:57.040 is the way in which something like the national digital platform, though I think the portfolio 529 00:52:57.040 --> 00:53:01.970 or some other term might work better here, is, you know, the narrative that allows us 530 00:53:01.970 --> 00:53:06.770 to sort of make that argument that helps us kind of break open our local institution a 531 00:53:06.770 --> 00:53:12.220 little bit more, and I wonder if that helps us toward some of the challenge, Mike, that 532 00:53:12.220 --> 00:53:16.750 you're pointing us to. On a technical level, right, I think it is about, like, that the 533 00:53:16.750 --> 00:53:21.140 web is the way to go, you know, that that's kind of a proven direction, but, I mean, there's 534 00:53:21.140 --> 00:53:27.510 so many other dimensions to this about sort of, you know, the institutions of realities, 535 00:53:27.510 --> 00:53:32.340 and we have to break those open too in order to sort of really attack your problem of scale. 536 00:53:32.340 --> 00:53:40.140 >>MARAY FLANAGAN: I love the idea of, like, thinking about, okay, why isn't, the ask me 537 00:53:40.140 --> 00:53:48.510 anything, why isn't that a reference library. To really engage with the public and interface 538 00:53:48.510 --> 00:53:52.360 where people are at and the tools are at is really exciting. 539 00:53:52.360 --> 00:53:56.940 >>DEANNA MARCUM: I think you're the next person. >> QUESTION FROM AUDIENCE: Blain Dessey, Library 540 00:53:56.940 --> 00:54:04.910 of Congress, and I represent one of those institutions that Deanna mentioned, but I 541 00:54:04.910 --> 00:54:12.360 have a very practical question, and coming at this from a national institution, or even 542 00:54:12.360 --> 00:54:19.570 if I were from a funding agency, when we talk about scaling to practice, what does that 543 00:54:19.570 --> 00:54:29.090 mean tangibly? I'm not a technologist, I'm just a bureaucrat.(Laughing.) That's what 544 00:54:29.090 --> 00:54:38.700 I am. It's not that funny. But, you know, if you were to come to me and say we got a 545 00:54:38.700 --> 00:54:46.550 scale to practice, this is critical, what does that mean? What would I do? Do I start 546 00:54:46.550 --> 00:54:54.600 convening meetings tomorrow? Do I buy people software? Do I write reports? What are we 547 00:54:54.600 --> 00:55:01.360 talking about in terms of people in positions of influence being able to move this forward? 548 00:55:01.360 --> 00:55:08.031 >>JON VOSS: It might be to get out of the way, honestly. That might be part of it. I 549 00:55:08.031 --> 00:55:14.320 do think that the working on the web and the way of the web is a totally different cultural 550 00:55:14.320 --> 00:55:19.100 construct, and, I mean, it goes along with Michael's thing here too, if we're going to 551 00:55:19.100 --> 00:55:24.920 talk about this kind of a scale, we're making, we are making a lot of small incremental changes, 552 00:55:24.920 --> 00:55:30.590 and even if you look at the meta data that we're dealing with, you know, we've gone from 553 00:55:30.590 --> 00:55:34.330 card catalog records to something very different now when we're starting to put things on the 554 00:55:34.330 --> 00:55:38.970 web, that's a whole different thing, and we don't want to make those little adjustments, 555 00:55:38.970 --> 00:55:44.610 we want to do something quite a bit differently, and I think, for me, from somebody who's a 556 00:55:44.610 --> 00:55:50.730 policy maker or somebody in a major institution, it's how do you enable the people on your 557 00:55:50.730 --> 00:55:57.910 teams to do this quite a bit differently and experiment with things quite a bit, you know, 558 00:55:57.910 --> 00:56:03.440 and not necessarily radical, but entirely different ways. That is how people are communicating 559 00:56:03.440 --> 00:56:10.580 differently. Without also alienating, or keeping in mind your end user or ultimate goals, but 560 00:56:10.580 --> 00:56:17.730 it's not to be an institution. >>TREVOR MUNOZ: I mean, I would suggest that, 561 00:56:17.730 --> 00:56:21.920 like, if we wanted to talk about scaling up practice, I mean, maybe you would start off 562 00:56:21.920 --> 00:56:25.730 with a meeting in which you'd bring together everyone in the organization and say here's 563 00:56:25.730 --> 00:56:29.720 where we're trying to get, and we're going to change everybody's job, we're not just 564 00:56:29.720 --> 00:56:34.220 going to add a new person to do this new thing and the other nine of you are going to keep 565 00:56:34.220 --> 00:56:38.660 doing what you were doing before. I think that would be a really critical thing, if 566 00:56:38.660 --> 00:56:42.210 you were going to start talking about scaling up practice. You know, it's got to be everybody's 567 00:56:42.210 --> 00:56:48.220 job. >>ANNE WOOTTEN: We are all pretty wary of 568 00:56:48.220 --> 00:56:52.000 not reinventing the wheel, and when it comes to scale, that's, like, paramount, so to give 569 00:56:52.000 --> 00:56:57.040 your team, to have that meeting and say let's go out and see what's been done, and this 570 00:56:57.040 --> 00:57:02.000 is why I talked about the importance of any of these projects, especially when they're 571 00:57:02.000 --> 00:57:06.390 federally funded, having outputs that are digestible by the people at organizations 572 00:57:06.390 --> 00:57:11.140 who are going to be able to take that information and figure out how they work it into what 573 00:57:11.140 --> 00:57:15.610 they're trying to scale at their organization and how they can leverage that or latch on 574 00:57:15.610 --> 00:57:20.120 to that and then in turn, you know, provide back to their work more information, more 575 00:57:20.120 --> 00:57:24.310 tidbits that others can add to sort of the collective knowledge. 576 00:57:24.310 --> 00:57:33.010 >>MARY FLANAGAN: I want to take this to the next week meeting. This is a great question. 577 00:57:33.010 --> 00:57:37.110 >>DEANNA: Tom. >> QUESTION FROM AUDIENCE: Tom Cramer from 578 00:57:37.110 --> 00:57:48.901 Stanford University. There was something that happened, Jon, while you were speaking. I 579 00:57:48.901 --> 00:57:55.750 just wanted to talk about briefly the value of verbs or the value of tools, which is the 580 00:57:55.750 --> 00:57:59.920 way, I think, you scale practice, and the way you make things accessible to a greater 581 00:57:59.920 --> 00:58:04.390 group of people is by giving them a simple tool that does a job well and that people 582 00:58:04.390 --> 00:58:09.990 understand the value, and so you can pop up archive, you can Google things, you can Reddit 583 00:58:09.990 --> 00:58:14.560 things, you can even do data curation, I think we know what that is. I think from someone 584 00:58:14.560 --> 00:58:22.670 who is working on link data, It's still too abstract. Crowd sourcing is kind of a mindless 585 00:58:22.670 --> 00:58:38.280 panacea. I think from a national digital platform perspective, one of the things that this group 586 00:58:38.280 --> 00:58:43.990 might do is identify what those little tools are and really drive the identification and 587 00:58:43.990 --> 00:58:48.640 adoption of those tools and really get more people playing in the milieu. 588 00:58:48.640 --> 00:58:51.930 >>ANNE WOOTTEN: Great. >>DEANNA MARCUM: Jim 589 00:58:51.930 --> 00:58:58.880 >>QUESTION FROM AUDIENCE: Yeah, I wanted to talk a little bit about Blain's question, 590 00:58:58.880 --> 00:59:03.740 what should the national organizations do, and I'd like to, I think there are four areas 591 00:59:03.740 --> 00:59:09.570 where they can be very helpful. One is the area of advocacy, talk this up, making it 592 00:59:09.570 --> 00:59:15.310 clear that this is a national priority. Two, information policy, helping us get the right 593 00:59:15.310 --> 00:59:19.751 legal and legislative framework so we can better serve our users through content and 594 00:59:19.751 --> 00:59:26.410 technology. Third is professional development. We'll hear more about that, but I think these 595 00:59:26.410 --> 00:59:31.760 organizations have the ability to build the type of leadership and professional development 596 00:59:31.760 --> 00:59:39.110 capacities for the wider community, and fourth is global leadership, positioning us around 597 00:59:39.110 --> 00:59:43.470 the world in terms of how we're going to knit together these various initiatives and to 598 00:59:43.470 --> 00:59:49.290 help us address the question of Europeana is easy, but how do we deal with the developing 599 00:59:49.290 --> 00:59:55.000 worlds of Africa, Asia, South America, and really begin to address those questions, because 600 00:59:55.000 --> 00:59:59.110 these organizations have important connections in those areas. 601 00:59:59.110 --> 01:00:10.660 >>DEANNA MARCUM: Great. Thanks, Jim. Any response from the panel? Just gratitude, I think. 602 01:00:10.660 --> 01:00:16.030 >>QUESTION FROM AUDIENCE: I'm Jeff Resnic. I am with the National Library of Medicine. 603 01:00:16.030 --> 01:00:27.740 I direct the history of medicine division. Particularly, as many know, I was talking 604 01:00:27.740 --> 01:00:32.910 with Brewster about it earlier today, about NLM and the national institute of health leading 605 01:00:32.910 --> 01:00:38.280 the way in open access, and this is something we continue to do despite challenges, and 606 01:00:38.280 --> 01:00:45.980 we hope that those efforts and the resources can be of use to this community. Looking forward, 607 01:00:45.980 --> 01:00:51.480 you know, if there are resources, the national library of medicine is one of America's three 608 01:00:51.480 --> 01:00:57.170 national libraries that could be useful, I'd be happy to connect with you and have a conversation 609 01:00:57.170 --> 01:01:01.580 and invite you out to our institution. Just wanted to put in a plug there. This has been 610 01:01:01.580 --> 01:01:07.480 an incredibly engaging conversation, and from that vantage point of one of the other national 611 01:01:07.480 --> 01:01:16.330 libraries, I think that the idea of knowledge communities as they are defined in the medical 612 01:01:16.330 --> 01:01:21.050 or scientific context, certainly the national science foundation plays a role, but there's 613 01:01:21.050 --> 01:01:28.360 no other place in the United States that generates so much medical knowledge, and the national 614 01:01:28.360 --> 01:01:32.930 library of medicine is a storehouse of that through our databases and various other resources, 615 01:01:32.930 --> 01:01:37.010 so just to throw that out there, to think about, you know, as we grow these communities 616 01:01:37.010 --> 01:01:44.260 of knowledge, as we find their interconnectivity, what about the personalization of that knowledge 617 01:01:44.260 --> 01:01:49.750 as it relates to health, as it relates to well being, I think that's a dimension that 618 01:01:49.750 --> 01:01:58.610 I'll call out. Finally, one of the premier advances that's unfolding now at the national 619 01:01:58.610 --> 01:02:02.430 institute of health is the precision medicine initiative that President Obama announced 620 01:02:02.430 --> 01:02:06.690 in his State of the Union earlier this year, and the NIH director Francis Collins is firmly 621 01:02:06.690 --> 01:02:15.750 behind this, and I think that defines, in many ways, the personalized nature of data 622 01:02:15.750 --> 01:02:23.410 of information vis a vis health in this world of exploding data, so how do we navigate that, 623 01:02:23.410 --> 01:02:28.230 and I think all the questions that you're raising, you know, overlaying them on top 624 01:02:28.230 --> 01:02:35.730 of personal data, individual data, in a world of, you know, a pool of data, it puts a particular 625 01:02:35.730 --> 01:02:40.670 spin on this. >> DEANNA MARUM: Extremely good point. Thank 626 01:02:40.670 --> 01:02:43.600 you. Lucy? >> QUESTION FROM AUDIENCE: Hi. I'll just quickly 627 01:02:43.600 --> 01:02:51.420 add a fifth one to what Jim added as something that I think we could be working on, even 628 01:02:51.420 --> 01:02:59.041 as other tools and issues need to be being developed for this portfolio, I'm using the 629 01:02:59.041 --> 01:03:10.290 new word, is that we need to be actually working with the users of the future on educational, 630 01:03:10.290 --> 01:03:17.440 on their digital literacy in order to use these resources and make of them what they 631 01:03:17.440 --> 01:03:27.120 will so that they can make videos or reuse videos to make new projects or do what Trevor 632 01:03:27.120 --> 01:03:35.750 was talking about, I'm sorry, Rich does, or encourages people to do with the products 633 01:03:35.750 --> 01:03:45.820 in archives, to make new, reuse digital materials and make new things. So, I know we sponsor 634 01:03:45.820 --> 01:03:52.220 a new small grant to encourage literacy and engagement with historical materials, but 635 01:03:52.220 --> 01:03:58.380 obviously, there's a need to encourage that kind of work with data materials, all types 636 01:03:58.380 --> 01:04:06.700 of, you know, for reading, I mean, education, the department grant, education does it too, 637 01:04:06.700 --> 01:04:14.490 but, I mean, that type of development of younger people needs to carry along, or else, in a 638 01:04:14.490 --> 01:04:22.560 sense, we're building a product that you're going to say, well, they're just going to 639 01:04:22.560 --> 01:04:28.620 grow up in it, you know, they're the digital generation, they'll know how to use it, but 640 01:04:28.620 --> 01:04:34.200 I think that there's been, you know, enough work to show that they don't just know how 641 01:04:34.200 --> 01:04:41.410 to use it, but to keep working on that side, it also teaches us what we don't know about 642 01:04:41.410 --> 01:04:48.640 how they use things, which is equally important, the case user studies and all that part of 643 01:04:48.640 --> 01:04:59.820 it. So, that's my fifth part. >>DEANNA MARCUM: Are there other people ready 644 01:04:59.820 --> 01:05:08.110 to speak? >>JON VOSS: In the meantime, I just want to 645 01:05:08.110 --> 01:05:12.750 get back to the sciences part there, because I think we're pretty humanities heavy in this 646 01:05:12.750 --> 01:05:16.980 meeting, so I really like the science point, and when you talk about, particularly linked 647 01:05:16.980 --> 01:05:22.850 data, scaling things across connective archives, databases, huge advances are happening there. 648 01:05:22.850 --> 01:05:26.780 You also, you have so much more funding in those areas. I mean, when we talk about doing 649 01:05:26.780 --> 01:05:32.540 linked open data projects within humanities, the scale is so much smaller, and I think 650 01:05:32.540 --> 01:05:37.830 there's a lot to be learned from what's happening in that space that can be transferrable, and 651 01:05:37.830 --> 01:05:43.870 I love this idea of intersection, the data exploration, you know, how you visualize this 652 01:05:43.870 --> 01:05:49.520 data. It could be pretty intense. >>DEANNA MARCUM: I think we all have so much, 653 01:05:49.520 --> 01:05:54.800 we're so much in the debt of the national library of medicine for doing so much good 654 01:05:54.800 --> 01:06:00.380 work in this area, so I think, it's really great to see you here so that the humanities 655 01:06:00.380 --> 01:06:06.150 and the sciences can talk together a bit more. >>MARY FLANAGAN: Thank you, Deanna. That sounds 656 01:06:06.150 --> 01:06:11.420 like a wonderful note for us to wrap up on. Jim just asked me if there was chocolate, 657 01:06:11.420 --> 01:06:13.220 so. (Laughing.) 658 01:06:13.220 --> 01:06:17.770 >>DEANNA MARCUM: Please join me in thanking the panel for their very good comments. 659 01:06:17.770 --> 01:06:17.780 (Applause.)