Considering Dynamic, Non-Textua Content when Migrating Digita Asset Management Systems Aya Stein; University of Iinois at Urbana-Champaign; Urbana, Iinois USA Santi Thompson; University of Houston; Houston, Texas USA Abstract The researchers investigate how important managing and providing access to dynamic, non-textua media was to institutions that were migrating from one digita asset management system to another. Drawing upon data from a arger survey, the researchers used responses to categorica questions addressing extensibiity, content management, and metadata to answer two queries: "Do survey resuts suggest that institutions are considering dynamic non-textua content when migrating to a new DAMS?" and "What trends or features (deveoped or added through migrating), identified from survey resuts, coud provide soutions to this probem?" The researchers concude that whie there is an interest in managing and providing access to dynamic content, there seems to be a disconnect between respondent priorities, desired functionaity, and system capabiities. Motivation As institutions continue to provide access to digitized and born digita coections and data, some have eected to migrate from one digita asset management system (DAMS) to another for a variety of reasons. The data referenced in this paper is drawn from "Identifying Motivations for DAMS: Migration: A Survey," which identified 13 topica categories for migrating from one digita asset management system (DAMS) to another. This paper anayzes a subset of the responses which focus on topics associated with managing and providing access to dynamic, nontextua media. With the proiferation of digitized and born digita, non-textua, dynamic content, it is necessary to design systems and services to meet this growing demand. The researchers beieve that resuts from this study may give insight into whether or not ibraries are considering impementing systems with improved access to non-static content or features that enabe interaction with content in new ways. Probem Some information professionas have suggested that as digita objects become more compex and dynamic, and as the research conducted with digita content evoves over time, existing DAMS have strugged to keep pace, especiay in tenus of dynamic content access, storage, and system scaabiity [1][2][3]. A number of institutions have begun addressing this issue. Indiana University (IU), one of the eariest institutions to document their process, migrated from Variations to Variations 2, a home-grown DAMS to another. IU eected to migrate to another home-grown soution because they had the capabiity to improve access to music and video content by extending existing software functionaity [2]. The Variations projects utimatey resuted in the Avaon Media System Project [4]. Preservation repositories are aso increasingy storing dynamic non-textua content, such as geo-spatia data and maps [5]. The iterature aso discusses user tags and armotations, which, whie primariy textua at first gance, require an entirey separate metadata mode than traditiona textua content [6]. The authors wi be responding to the foowing questions: Question!: Do survey resuts suggest that institutions are considering dynamic, non-textua content when migrating to a new DAMS? Question 2: What trends or features (deveoped or added through migrating), identified from survey resuts, coud provide soutions to this probem? Definitions Researchers draw upon defmitions deveoped by the Caifornia Digita Library's Gossary to frame the differences between simpe and compex objects. Additionay, they have created a defmition for Dynamic, Non-Textua data. Simpe digita objects- "Comprised of a singe content fie (and its format variants or derivatives) and the metadata for that fie. For exampe, a TIFF of the Mona Lisa, a user JPEG, a reference GIF, and the appropriate metadata woud comprise a simpe digita object" [7]. Compex digita objects- "Incudes two or more content fies (and their format variants or derivatives) and corresponding metadata. The content fies are reated as parts of a whoe and are sequenced ogicay, such as pages. For exampe, a compex digita object coud consist of a muti-page diary scarmed as TIFF images, from which are generated dispay images (JPEGs and GIFs), pus a transcription of the diary and the metadata for each fie" [7]. Dynamic, non-textua content- For the purposes of this study, defmed to be sound, fthn, digita art, and compex digita objects. Approach To compete this study, researchers anayzed a subset of data from a arger investigation that seeks to identify motivations for migrating from one DAMS to another. Using a survey as their instrument, the researchers soicited participation from eigibe institutions from Juy through September 2014. In order to quaify for the survey, respondents had to fufi one of the foowing three eigibiity categories: 1. Institutions had competed migration from the "Od DAMS" to the "New DAMS" 2. Institutions were currenty migrating from the "Od DAMS" to the "New DAMS" 3. Institutions seected a "New DAMS" but had not started the migration process. After removing entries that were not eigibe or were not fuy competed, the researchers had 49 responses to anayze for this Archiving 2015 Fina Program and Proceedings 119
------- study. Over haf of the eigibe responses came from academic ibraries. Tabe 1: Which of the foowing best describes your ibrary? Response Type N %!Academic Library 30 61 Research Library 8 16 Pubic Library 4 8 Specia Library 2 4 Specia Coections Libraries or Archives 2 4 Government Library 2 4 Other 1 0 \i~~~l_1'1 Lib~t:~_ry 0 0 To create the survey, the researchers crafted specific questions armmd 13 topics [8] reated to DAMS evauation, incuding: Impementation & Day-to-Day Costs User Administration Organizationa Viabiity Technica Support System Administration Extensibiity Information Retrieva & Access Content Management Preservation User Interface Customization Interoperabiity Reputation Metadata Standards Survey questions for these topics were designed to be either a Likert scae of 1 [Not Important] to 4 [Very Important] or seect a that appy. The survey asked for key demographic information to hep the researchers understand how institutions prioritized potentia motivations, and it aso tracked that movement when institutions migrated from the "Od DAMS" to the "New DAMS.'' Demographic questions required respondents to seect and/or sefidentify the "Od DAMS" and "New DAMS." Additionay, the survey asked respondents to choose the top five motivations from one of the 13 topics isted above and then prioritize those five seections in order from most important to east important. Respondents woud then answer questions in the survey that focused on the five topics they identified. Because the survey had over 100 questions, the researchers used this method to reduce the overa time needed to compete it. Since the scope of this paper is to understand the reationship between dynamic content and DAMS migration, the researchers identified questions from the survey that addressed topics reated to acquiring, describing, preserving and making accessibe dynamic, non-textua media. The questions most reated to these areas fe in the foowing survey topics: Extensibiity: The abiity to incorporate additiona functionaity and capabiities to the "New DAMS's" via viewing and manipuating the system code base, APis, socia media integration, or other measures. 120 Content Management: Coection content and administration in the "New DAMS", incuding fie formats, ingest issues, scaabiity, and rights management information Metadata: The "New DAMS's" support of estabished metadata standards, user generated metadata, and inked \ data technoogies Researchers used the survey reports feature in Quatrics to generate descriptive statistics for the seected questions and drew \ upon these reports to formuate concusions and identify future research areas. Resuts Researchers anayzed the number of responses, mean of the responses, and standard deviation from survey questions to better understand the roe that dynamic media pays when migrating from one DAMS to another. The date in the foowing tabes aggregate responses to questions in the Extensibiity, Content Management, and Metadata topic areas, respectivey. Researchers considered statements that registered mean responses higher than 3.0 and a standard deviation ofess than 1.0 to be important considerations for institutions migrating to a new DAMS. This means that a questions identified in this paper, excuding "The size imit for ingesting content" were important considerations for survey respondents. Tabe 2: Survey questions reated to extensibiity, content management. and metadata Question N Mean so Institutions can create their own modues/puginswidgets/etc for the "New 21 3.67.58 DAMS".!rhe "New DAMS" has an avaiabe API. 22 3.64.58!The abiity to support various fie formats 28 3.61.79 r-he code base is avaiabe for anyone to see and use (open source}. 22 3.55.60!The capacity to dispay both simpe and ~mpex digita objects 29 3.48.78!The storage capacity to hod both simpe ~nd compex digita objects 28 3.46.79 r-he "New DAMS" supports digita object "dentifiers. 22 3.23.97!The governing organization creates modues/pugins/ widgets/etc. to fit the 22 3.00.76 needs of your institution. The size imit for ingesting content 28 2.86 1.01 The foowing responses are to questions that shed additiona ight into the types of materia and fie formats as we as the kinds of administrative and technica metadata that a new DAMS shoud support. 2015 Society for Imaging Science and Technoogy
Tabe 3: What types of objects did you desire the system to dispay? ~nswer N % Manuscripts 24 83 Images 24 83 Maps 23 79 Books 22 76 ~ideo Content 21 72 ~udio Content 21 72 Spreadsheets 11 40 Graphs 9 31 ~D Objects 8 28 Other 5 17 Research data 2 7 K31S 1 3 Audio and video content are, by far, the most popuar dynamic, non-textua content in which ibraries are interested in DAMS supporting. 72% of respondents designated that DAMS shoud dispay audio and video content, suggesting that future systems shoud render some dynamic content just as we as it renders non-dynamic types of content, incuding maps, books, manuscripts, and static images. Tabe 4: What fie formats did you desire the "New DAMS" to support? Answer N % PDF 28 97 JPEG 26 90 MP3 22 76 JPEG2000 21 72 TIFF 21 72 MP4 19 66 MOV 17 59 csv 16 55 DOC 13 49 DOCX 12 41 KML 2 7 WAV 2 7 GIS 1 3 KMZ 1 3 ArcGIS fies 1 3 Of a the fie formats reating to dynamic, non-textua content, those reating to audiovisua formats were the most popuar. Desired support for specific audio and video formats varied. The most desired audio format was MP3, with 76% of responses, foowed by 7% of responses for W A V fies. MP4 video fies taied 66% whie MOV fies received 59%. Tabe 5: What administrative, preservation, structura, and/or technica metadata standards 1 did you desire the "New DAMS" to support? Response N % METS 18 90 PREMIS 15 75 TE 8 40 VRA Core 5 25 MIX 2 10 PB Core 2 10 Other 0 0 The most widey used administrative, preservation, structura, and/or technica metadata standards desired were METS metadata, at 90% of responses, foowed by PREMIS preservation metadata with 75%, and TE with 40%. There was very itte interest in administrative, structura and/or technica metadata beyond these. Discussion Question 1: Do survey resuts suggest that institutions are considering dynamic, non-textua content when migrating to a new DAMS? Anayzing survey resuts that address desired future support for specific content types shows that institutions are inconsistent when considering dynamic, non-textua content as a motivation for migrating to a new DAMS. Resuts favored the DAMS supporting a diverse number of fie formats (going beyond static images to incude audio, moving image, GIS, and digita maps). They aso signaed that institutions have an interest in managing and making avaiabe audio and video content. Researchers beieve that these fmdings demonstrate that more institutions have an expectation to make audio and video content accessibe. This deveopment is not necessariy surprising given the increase in audio/visua hodings among institutions in addition to the expanding abiities of DAMS to curate and make accessibe audio/visua fie formats. However, providing access to other dynamic content types was not as significant a concern when choosing a New DAMS. For exampe, there was no cear consensus on whether providing access to schoary output or research data was an important factor when migrating DAMS. There are most ikey severa pausibe expanations for this; content producers may deposit data into discipine base, institutiona based, or data based repositories, which is a future area of research for the researchers. Managing 3D objects in a repository did not appear as important to respondents. High costs and a ack of accessibe digitization equipment for 3D materias coud pay a roe in this resut. In addition to these expanations, readers shoud aso note that the researchers' survey may hod key imitations when it comes to Researchers aso incuded questions about descriptive metadata in the survey. Responses to these questions showed an overwheming need (a respondents saying it was either "very important"' or "important") for descriptive metadata support in a new DAMS. Beieving these resuts refected obvious concusions, the researchers eected not to incude this data in this paper. Archiving 2015 Fina Program and Proceedings 121
understanding institutiona needs around schoary and research data and 3D objects. These and other imitations wi be discussed at the end of this paper. Question 2: What trends or features (deveoped or added through migrating) identified from survey resuts coud provide soutions to this probem? Two reated trends emerged from the survey resuts. Institutions desire the abiity to expand the kinds of access and interfaces avaiabe for compex digita objects, which incude dynamic, non-textua media. This trend is driven by favorabe responses to the support of digita object identifiers and appication program interfaces (APis). Digita object identifiers offer unique identification for every object in a repository; combined with uniform resource identifiers (URis), digita object identifiers become the buiding bocks for estabishing persistent, ong-term access references. The researchers aso hypothesize that this interest can estabish the foundation for a more interactive schoary environment within the DAMS. APis are defmed as "a set of routines, protocos, and toos for buiding software appications" [9]. APis enabe institutions to generate aternative "views" and functionaity, such as visuaizations, timeines, and geo-referenced maps, with digita objects managed by DAMS. Expanding these areas provide institutions with new ways to serve content and functionaity to users seeking dynamic, non-textua content. Compementing expanding access and interfaces, institutions desire the abiity to expand system capabiities using their own resources and deveopers. Resuts show that institutions prefer the abiity to create their own toos compared to reying on the system's governing body to do so. Additionay, resuts favored having access to a code base that is avaiabe for anyone to open and use. This offers institutions the opportunity to deveop systems that hande compex digita objects and meet the research needs of their user communities in more efficient and fexibe ways. Inconcusive Resuts: DAMS functionaities that support dynamic non-textua content and compex digita objects. The researchers fee that the response data for the foowing questions cannot be readiy appied to either of the questions asked at the beginning of the paper. These are areas that wi require future investigation and research. The survey question "The size imit for ingesting content", was the ony question examined in this paper that was not considered important to the survey respondents. Out of 28 survey respondents, the mean was 2.86 with a standard deviation of 1.01. Whie most respondents, 18/28 indicated 'the size imit for ingesting content' was important or very important, the remaining 10 respondents indicated it was not important (3/28) or somewhat important (7/28), eaving no rea consensus on the matter. The researchers found this ack of consensus curious, considering that dynamic non-textua content fies such as audiovisua, digita art, and 3D objects, are often arger and more difficut to send over a network than static PDFs or image ftes. However, the ack of consensus does seem to aign with the resuts from our frrst question, that institutions are inconsistent when considering dynamic, non-textua content in DAMS migration. The second survey question the researchers fet was in need of further research was, "What administrative, preservation, structura, and/or tecmica metadata standards did you desire the "New DAMS" to support"? Ony two respondents indicated interest in PBCore, and ony five indicated interest in VRA Core. The researchers expected that if institutions were considering dynamic non-textua content in their DAMS migrations, then DAMS support for PBCore and VRA Core woud be much higher. One possibiity for these resuts may be that the generic output of PREMIS, METS, and descriptive metadata schemes coud be seen as sufficient by most institutions, so the granuarity of VRA Core and PBCore may not be seen as necessary. Limitations The survey used for this study was not designed to address dynamic non-textua content in DAMS migration excusivey. Consequenty, the researchers identified severa imitations. The design of the survey ed to an inconsistent number of responses for the 13 topics reated to DAMS evauation. When writing survey questions, the researchers did not defme some terms that coud be interpreted in severa ways, incuding 'simpe digita objects', 'compex digita objects', or 'open source software'. Aso, the wording for severa survey questions coud have been more carefuy chosen. For exampe, whie the researchers inquired about 3D content in the question asking, 'What types of object did you desire the system to dispay?', they did not incude any possibe 3D object fie formats in the question 'What fie formats did you desire the "New DAMS" to support?'. Additionay, the researchers did not eave a free-text option for 'other' in the 'Seect a the fie formats that appy' question. The survey scope specified that respondents focus on motivations for DAMS as they appy to digita ibrary or digita coections environments, not repositories designed for institutiona, schoary, or research data, which might contain dynamic, non-textua content. However, migration of institutiona, schoary, or research data repositories is an area of future investigation for the researchers. The fma imitation that the researchers identified was the Ango-centraity of the istservs to which the survey was sent. Whie most of the istservs that the survey was distributed to boast internationa subscribers, they were not necessariy focused more gobay or in non-engish anguages. In the future, the authors wi make an effort to send survey invitations to IFLA and other internationa istservs as we. Concusion The researchers beieve that when migrating DAMS there is an interest in managing and providing access to dynamic nontextua content in an environment that aows for more sophisticated user interaction and research, particuary around audio and video. However, other emerging, more compex content, incuding 3D objects, geospatia data, and research data, remain daunting for ibraries to manage, make accessibe, and preserve. Future research in the areas of institutiona and data repositories, digita research environments, and more compex data types can begin to accimate ibrarians with these emerging chaenges. References [1) C. L. Borgman, "The Digita Future is Now: A Ca to Action for the Hrunanities," Digita Hrunanities Quartery, 3, 4 (2009). Retrieved March 27, 2015, from http://www. digitahumanities.org/dhq/vo/3/4/000077 /000077.htm. 122 2015 Society for Imaging Science and Technoogy
[2] J. Dunn, M Notess, and R. Schere, "Fipping the Switch: Lessons Learned from a Major Digita Library Migration Project," Presented at the Digita Library Federation Fa2005 Fof11, Charottesvie, Virginia, November 7-9,2005. [3] K. Masood, & A Neatrour, "Digita Asset Management System Options: Report of the University of Utah Libraries DAM Review Task Force," MWDL Winter Webinar Series (2014) Retrieved March 27,2015, from http://www.mwd.org/events/dams_options.php. [4] Avaon Media System Project, "The Project", (2014). Retrieved December 7, 2014, fromhttp://www.avaonmediasystem.org/project. [5] J. Hutai, "Archives New Zeaand Migration from Fedora Commons to the Rosetta Digita Preservation System," ipres2013 Proceedings (2013) [6] C. Lagoze, D.B Kraft, S. Payette,& S. Jesuroga, "What Is a Digita Library Anyway~" D-Lib Magazine, 11, 1-22 (2005) doi: I 0.1 045/november2005-agoze. [7] Caifornia Digita Library, "Gossary" (2014) Retrieved March 30, 2015 from http://www. cdib.org/gateways/technoogy/gossary.htm. [8] Mari, 1. L., & Luczak, E. C. (2009). Evauation of digita repository software at the Nationa Library of Medicine. D-Lib Magazine, 15(5-6). doi: I 0.1 045/may2009-mari. [9] Webopedia "API- Appication Program Interface" (2015) Retrieved March 30, 2015 from http://www.webopedia.com/term/napi.htm Author Biography Aya Stein is currenty a Metadata Librarian at the University of iinois Urbana-Champaign Main Library. She hods a BA in East Asian Studies.from the University of Ari::ona and a MS Q[riformation.from the University of Michigan. Santi Thompson is currenty the Head of Digita Repository Services at the University of Houston (UH) Libraries. He hods a MLIS and MA, Pubic History from the University of South Caroina. Archiving 2015 Fina Program and Proceedings 123