Old silos, new silos, no silos

2,956 views
2,617 views

Published on

SWIB12, Köln, 2012. Video recording: http://www.scivee.tv/node/55381

0 Comments
6 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
2,956
On SlideShare
0
From Embeds
0
Number of Embeds
35
Actions
Shares
0
Downloads
68
Comments
0
Likes
6
Embeds 0
No embeds

No notes for slide
  • Questions, questions, questions
  • (Wikipedia) Information silos: systemsincapable of reciprocal operation with other, related information systemsFor who is this a problem in the library world?
  • Silo: data within a system, onlyaccessible via internal system calls.Betternow: direct access to data without goingthrough the system
  • Lots of “silo” systems provideAPIs, but theyonlygive access to systemfunctionality, notdirectly to the data
  • Focus in this talk willmainlybe on traditional library contentThis is were the silos are
  • A couple of stories. Here is one, it starts 7 yearsagoSFX/MetaLib User Group – informalorganisation, Ex Libris customersSmug= selbstgefälligSMUG 4 EU NewsletterinitiativeThis: “funny” conversationbetween 3 conference delegates, addressing issues facinglibrariesanddistributed search
  • A couple of currentbuzzwordsalready!
  • SMUG + ICAU mergedinto IGeLU 2006Issue: Integration
  • Summary of issues from SMUG-4-EU 2005-2007Issues currentlyverymuchalive
  • The origins of silos in the library data world
  • Paper,physicalcataloguestransposed to digital, standalone world at firstMARC was actually the first attempt to solvelibrary silo problem ;-)
  • Cataloguesmoved to online web world
  • Scholarly databases appeared, first standalone, CD-ROM
  • Scholarly databases, eJournalsmoved to the online web world
  • All these systems kept the closedintrovertednature of pre-web: silos, data trappedinside systems
  • Summary of the origins of silos, in the libraryworldBesides “old” technologyand commercial reasons:also important psychologicalmindset, traditions. Libraries want to defendtheir “authoritative” positions as much as publishersandvendors want to defendtheir commercial positions
  • This is aboutabout data, including metadataStartingwith traditional bibliographic data containers.
  • Problemswithfederated search:PerformanceTechnical bottlenecksInconsistent indexes (author, subject), results, multilingual issues
  • On top of low level systems, aggregations.Twoareas:Locallibrary data (books, journals)Shared databases, ejournalsExtra “super” levels; WorldCat etc. – Google Scholar
  • Federated search alsopossiblewithaggregations
  • Or new level of federated search: aggregated search?
  • New level of aggregation: discoverylayersCombininglocalandglobal data sources: local+sharedindexesNew super level of silos + redundancy
  • Illustration of hugeredundancy of bibliographic data of one ‘book’, one ‘article’.
  • Different perspective: where are library data of different typesstored?Bibliographic data; everywhereHoldings data: local systems, aggregated systems; commercial vendorsCirculation data (usage data): local systems, commercial vendors, open access?
  • Andanother type of systems/silosemerge: (multiple) next generation ILS, combination of aggregations of shared bibliographic data + shared ‘local’ holdings + localcirculation data.
  • Somelocalsilos+redundancywilldisappear, data moved to Next Gen silos
  • Intermediateconclusion:thissituation is notveryefficient.For librarians, librarycustomers, system librarians
  • Like banking crisis: library crisis.No EU to take measures. Maybe we need a “Bad Library” (like a “bad bank”) to put all the MARC records in, and get on withrestructuring
  • All types of combinationspossibleAll kinds of technologiespossible
  • Oneuniversal silo ;-)Ideally no redundancy
  • Next Gen ILS (Alma,Worldshare, Intota, Kuali OLE, etc.)Basic design: shared zone (with data for FRBR type Work, Expression, Manifestation)Library zone: holdings, items etc.For Print+ElectronicSimilartoexistingunioncatalogues/shared cataloguing/Worldcat
  • LoCBibliographic Framework TransitionInitiative: Mapping of Bibliographic data tooneWork level.Just a framework, no model/implementation
  • Design of Next Gen ILS, andexistingunion/sharedcataloguingcanbemappedto the Framework.All kinds of aggregationpossible.
  • For instance: all next gen ILS,instead of becoming new silosuseWorldcat as Community/Shared zone for printuse Google Scholar as Community/Shared zonen forelectronic/articles
  • New frameworkbased on linked open data makesitpossibletoconnectlibrary data toanythingelse, eliminating the old silo mindset
  • Allvery non-committal, without real effect
  • Stop thinking systemsStart thinking data
  • Old silos, new silos, no silos

    1. 1. Old silos, new silos, no silosFrom redundancy to aggregation or distribution?Lukas KosterLibrary Systems CoordinatorLibrary of the University of Amsterdam@lukaskl.koster@uva.nlSWIB12, Köln, November 26-28, 2012 http://www.flickr.com/photos/alpoma/4786094569
    2. 2. “System incapable of reciprocal operation with other, related What is a silo? information systems” http://en.wikipedia.org/wiki/Information_silo Old silos, new silos, no silos - SWIB12 - 2 @lukaskhttp://www.flickr.com/photos/7716310@N03/1472185959
    3. 3. “System containing data, only accessible via internal system calls” access dataSILO NO SILO Old silos, new silos, no silos - SWIB12 - 3 @lukask
    4. 4. Application Programming Interface API Direct accesssystem call data call (LOD) http://twitter.com/hochstenbach/status/268033173366124544 http://twitter.com/jindrichmynarz/status/269089964103434242 Old silos, new silos, no silos - SWIB12 - 4 @lukask
    5. 5. FocusTraditional library content: publicationshttp://commons.wikimedia.org/wiki/File:Latin_dictionary.jpg Old silos, new silos, no silos - SWIB12 - 5 @lukask
    6. 6. StoryOld silos, new silos, no silos - SWIB12 - 6 @lukask
    7. 7. http://igelu.org/wp- content/uploads/2010/10/smug4eu_issue2.pdfOld silos, new silos, no silos - SWIB12 - 7 @lukask
    8. 8. ! 7 years ago!A: Federated search systems are deadBut you won’t get those kids back that you have already lost. They will staywith GoogleB: But what is the alternative? Build big databases like Google does?Harvest data till you drop?Probably we do need both. Big databases as well as federated search.C: Perhaps only a limited number of distributed harvested indexesaround the world.Federated search can survive and be used for accessing those indexes! Old silos, new silos, no silos - SWIB12 - 8 @lukask
    9. 9. ! 7 years ago!A: Federated search systems are deadBut you won’t get those kids back that you have already lost. They will staywith GoogleB: But what is the alternative? Build big databases like Google does?Harvest data till you drop?Probably we do need both. Big databases as well as federated search.C: Perhaps only a limited number of distributed harvested indexesaround the world.Federated search can survive and be used for accessing those indexes! Old silos, new silos, no silos - SWIB12 - 9 @lukask
    10. 10. http://igelu.org/wp- content/uploads/2010/10/smug4eu_issue4.pdfOld silos, new silos, no silos - SWIB12 - 10 @lukask
    11. 11. If libraries describe their philosophy they’ll certainly also talk about integration– integration of services, integration of assets and resources.To make an excellent service all these tools need to function as an integratedsystem. Not only with each other, but also interacting with the rest, such as theinstitutional website, the institutional portal, third party’s learning environments.Technically speaking, integration may result in a functioning or a unified whole.A “functioning whole” may be understood as an integrated structure ofcooperating tools in which the different parts (like in this case SFX, MetaLib orother tools) can be distinguished as such, whereas a “unified whole” wouldmean that the end users experience one “new” system or user interface thatconceals the integrated parts.Combining different databases in real or virtual collections that can beaccessed by a number of integrated or nonintegrated tools.When libraries are talking about integration, they are not only thinking aboutthe user’s perspective, but also about the backend of their business, whereinternal workflows are in focus. In general, our aim is to reduce duplicatingwork, to store the same information only once, to cooperate closely. Old silos, new silos, no silos - SWIB12 - 11 @lukask
    12. 12. Libraries – integration of services, assets and resources.Integrated systems • Functioning whole: an integrated structure of cooperating, distinguishable tools • Unified whole: one new system of concealed integrated partsCombined databases • Real collections • Virtual collectionsData management • Harvest data • Big data • Distributed harvested indexesBackend integration of internal workflows • Reduce duplicating work • Store the same information only once • Cooperate closely Old silos, new silos, no silos - SWIB12 - 12 @lukask
    13. 13. StoryOld silos, new silos, no silos - SWIB12 - 13 @lukask
    14. 14. http://www.flickr.com/photos/88526197@N00/4218544261/Libraries: local collections of physical printed publications Old silos, new silos, no silos - SWIB12 - 14 @lukask
    15. 15. http://www.flickr.com/photos/88526197@N00/4218548045/ Catalogues: finding publications and their locations Old silos, new silos, no silos - SWIB12 - 15 @lukask
    16. 16. http://www.flickr.com/photos/cat-sidh/457060770Old silos, new silos, no silos - SWIB12 - 16 @lukask
    17. 17. http://en.wikipedia.org/wiki/File:Screenshot_of_Dynix_library_automation_software_in_green.png http://www.flickr.com/photos/dukeyearlook/7046582583 Old silos, new silos, no silos - SWIB12 - 17 @lukask
    18. 18. Old silos, new silos, no silos - SWIB12 - 18 @lukask
    19. 19. http://www.flickr.com/photos/oliver_leitzgen/2967000266/ 19 Old silos, new silos, no silos - SWIB12 - @lukask
    20. 20. Old silos, new silos, no silos - SWIB12 - 20 @lukask
    21. 21. Silos Data trapshttp://www.flickr.com/photos/msueasrecords/6242093298/ Old silos, new silos, no silos - SWIB12 - 21 @lukask
    22. 22. Silo originsTechnologyEconomicsLicenses http://www.flickr.com/photos/kotomi-jewelry/5035479051/TraditionPsychology Old silos, new silos, no silos - SWIB12 - 22 @lukask
    23. 23. Serious Story Old silos, new silos, no silos - SWIB12 - 23 @lukask
    24. 24. ILS = Library CatalogueREPO = Institutional RepositoryDB = DatabaseEJ = EJournal Bibliographic data ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 24 @lukask
    25. 25. ILS = Library CatalogueREPO = Institutional RepositoryDB = DatabaseEJ = EJournal Federated search ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 25 @lukask
    26. 26. ILS = Library CatalogueREPO = Institutional RepositoryDB = DatabaseEJ = EJournalAGG = Aggregation Aggregations Google Scholar Worldcat Union Repository Catalogue Gateway AGG AGG AGG ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 26 @lukask
    27. 27. ILS = Library CatalogueREPO = Institutional RepositoryDB = DatabaseEJ = EJournalAGG = Aggregation Federated search Google Scholar Worldcat Union Repository Catalogue Gateway AGG AGG AGG ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 27 @lukask
    28. 28. ILS = Library CatalogueREPO = Institutional RepositoryDB = DatabaseEJ = EJournalAGG = Aggregation Aggregated search Google Scholar Worldcat Union Repository Catalogue Gateway AGG AGG AGG ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 28 @lukask
    29. 29. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer DI Discovery DI DI Google Scholar Worldcat Union Repository Catalogue Gateway AGG AGG AGG ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 29 @lukask
    30. 30. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer DI DI DI Google Scholar Worldcat Bibliographic data Union Repository Catalogue Gateway AGG AGG AGG ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ Old silos, new silos, no silos - SWIB12 - 30 @lukask
    31. 31. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer B B B C H C H C H DI DI DIB Bibliographic data B B HH Holdings data B Google Scholar H HC Circulation data Worldcat B H B H Repository Union Catalogue Gateway AGG AGG AGG B B B B B H H H C H C ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ B B B B B B B B B B B B H H H H H H C H Csilos, new silos, no silos - SWIB12 - Old C C C C @lukask C 31
    32. 32. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer B B BNG = Next Generation ILS C H C H C H DI DI DIB Bibliographic data B B HH Holdings data B Google Scholar H HC Circulation data NG Worldcat B NG H NG B H Repository Union Catalogue Gateway AGG AGG AGG B B B B B H H H C H C ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ B B B B B B B B B B B B H H H H H H C H Csilos, new silos, no silos - SWIB12 - Old C C C C @lukask C 32
    33. 33. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer B B BNG = Next Generation ILS C H C H C H DI DI DIB Bibliographic data B B HH Holdings data B Google Scholar H HC Circulation data NG Worldcat B NG H B NG B H C B C H Repository H Union Gateway B Catalogue H C AGG AGG AGG B B B B B H H H C H C ILS REPO ILS REPO ILS REPO DB DB DB EJ EJ EJ B B B B B B B B B B B B H H H H H H C H Csilos, new silos, no silos - SWIB12 - Old C C C C @lukask C 33
    34. 34. ILS = Library CatalogueREPO = Institutional Repository DL DL DLDB = DatabaseEJ = EJournalAGG = AggregationDI = Discovery IndexDL = Discovery Layer B B BNG = Next Generation ILS C H C H C H DI DI DIB Bibliographic data B B HH Holdings data B Google Scholar H HC Circulation data NG Worldcat B NG H B NG B H C B C H Repository H Union Gateway B Catalogue H C AGG AGG AGG B B B B B H H H C H C REPO REPO ILS REPO DB DB DB EJ EJ EJ B B B B B B B B B B H H H H Old C H Csilos, new silos, no silos - SWIB12 - C C @lukask C 34
    35. 35. Not very efficient Free market leads to efficiency? Not really! Huge redundancies Missing dataObscure, cloudy situation Unequal access Old silos, new silos, no silos - SWIB12 - 35 @lukask
    36. 36. Library crisis 245 $ a Old silos, new silos, no silos - SWIB12 - 36 @lukask
    37. 37. Silutions?http://www.flickr.com/photos/thelightningman/5494666930 Old silos, new silos, no silos - SWIB12 - 37 @lukask
    38. 38. Stand-alone Aggregation Redundant ComplementaryStand-alone Redundant Complementary Redundantaggregation Aggregations aggregations Complementary Old silos, new silos, no silos - SWIB12 - 38 @lukask
    39. 39. Completely distributed• Less redundancy • Redundancy risk• Up to date • Performance risk• Trust • Persistence risk• Share workload • Inconsistent indexes• … • Trust • … http://en.wikipedia.org/wiki/Fallacies_of_Distributed_Computing Old silos, new silos, no silos - SWIB12 - 39 @lukask
    40. 40. Is linked data the answer? Karen Coyle at EMTACL12: “The web is awash in bibliographic data” “Most of what we would contribute would be duplication of data that already exists” “It is library holdings that is key to providing service and furthering knowledge creation”http://www.kcoyle.net/presentations/thinkDiff.pdf Old silos, new silos, no silos - SWIB12 - 40 @lukask
    41. 41. Completely centralised Old silos, new silos, no silos - SWIB12 - 41 @lukask
    42. 42. Completely aggregated• Speed • Persistence risk• Consistency • Trust• No redundancy • …• Share workload• … Old silos, new silos, no silos - SWIB12 - 42 @lukask
    43. 43. Distributed aggregation• Speed • Persistence risk• Consistency • Reduncancy risk• Share workoad • Silos• … • … Old silos, new silos, no silos - SWIB12 - 43 @lukask
    44. 44. Next Generation ILS in the cloud FRBR Worldcat Community Work /shared Expression Union zone Manifestation Catalogue Library Items ILS zone Holdingshttp://thoughts.care-affiliates.com/2012/10/impressions-of-new-library-service.html Old silos, new silos, no silos - SWIB12 - 44 @lukask
    45. 45. LoC Bibliographic Framework Work Items Expression Manifestation Holdings http://www.loc.gov/marc/transition/ Old silos, new silos, no silos - SWIB12 - 45 @lukask
    46. 46. WorldcatCommunity /shared Union Catalogue zone Library ILS zone Old silos, new silos, no silos - SWIB12 - 46 @lukask
    47. 47. Worldcat Community Google Scholar /shared zone Library zone Old silos, new silos, no silos - SWIB12 - 47 @lukask
    48. 48. Brave new world Old silos, new silos, no silos - SWIB12 - 48 @lukask
    49. 49. Library roles• Reconsider objectives• Shift focus• Communicate with system vendors• Communicate with content vendors• Engage in framework transition• Cooperate Old silos, new silos, no silos - SWIB12 - 49 @lukask
    50. 50. Don’t think systems Think data http://www.flickr.com/photos/57402879@N00/261930788/ Old silos, new silos, no silos - SWIB12 - 50 @lukask

    ×