Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
David Walker<br />California State University<br />From metasearch to metaservices<br />
Introduction<br />Part philosophy<br />What are we trying to do w/ metasearch?<br />Where are the problems?<br />How might...
Rethinking metasearch?<br />We’ve been at it long enough now<br />Google Scholar<br />Newer systems<br />Next-generation c...
New metasearch models<br />Metalib X-Server<br />Design your own interface<br />Incorporate other systems and data<br />Ex...
Xerxes<br />PHP /  XSLT / MySQL<br />DB<br />X-Server<br />Metalib<br />KB<br />
What are we trying to accomplish?<br />“We offer a fragmented set of systems to search for published information . . . eac...
Overview of the literature<br />1990-1999<br />“Is There a Chance for a Standardized User Interface?” –Fletcher<br />2000-...
Overview of the literature<br />2004-2008<br />“Is Metasearching Dead?” –Tennant<br />“Metasearching: Not as Good as We&ap...
Arguing about metasearch<br />“[C]ross-database search tools . . . are the correct solution for unifying access to a varie...
Arguing about metasearch<br />Broadcast argument (pro)<br />“You can search multiple databases simultaneously!”<br />Nativ...
Isn’t it ironic?<br />Some librarians dislike metasearch because it is too much like Google; othersbecause it is not enoug...
Usage statistics<br />SFX as a proxy measure?<br />Query # 2: Requests by source (SID)<br />Not all databases or clicks<br...
Usage statistics<br />Fullerton<br />Native: 297,602  (54%)<br />Xerxes: 252,238  (45%)<br />Google: 5,340         (1%)<br...
Native: 22,654   (48%)
Xerxes: 22,275    (48%)
Google: 1,944       (4%)</li></ul>Native:  416,141 	52%<br />Xerxes:  367,047 	46%<br />Google: 20,98	   3%<br />
Isn’t it ironic?<br />Usage of Google Scholar may depend in large part on whether librarians promote it or not!<br />
Usability studies<br />Universities<br />Boston College –BYU – Carnegie Mellon – Maryland – Mississippi – Northwestern – O...
Usability studies<br />70% of the students in BYU study preferred metasearch over native interfaces<br />“[B]oth [metasear...
On search times<br />“Eight [out of 18] students measured the speed of the search processing as reasonable and only five f...
On the interface<br />“I found that both were not very user friendly.” – BYU  Student<br />“[Non-federated search] lent it...
Broadening our goals<br />“[T]he point of federated searching is to make searching as simple as possible” –Cervone<br />Is...
Selecting the right resources<br />“Nothing slows the user&apos;s scanning momentum more than encountering results that ar...
Selecting the right resources<br />Why not search everything?<br />Impractical technically<br />Impractical presentational...
Once you start down the Quick Search path, forever will it dominate your destiny.<br />
Desperately seeking search box<br />“I just want a search box on the homepage.” 					– Your users<br />First Rule of Usabi...
Desperately seeking search box<br />“[T]here is something inherent in the site&apos;s design that causes users to choose t...
Isn’t it ironic?<br />The search box dominates the opening screens, then disappears!<br />
Changing queries<br />“Nearly all students repeated their searches, changing terms or subject categories, so the interface...
Services for results<br />Spell check<br />Peer reviewed flag<br />Full-text look-up<br />Full-text linking<br />Format<br...
Avoid pogosticking<br />“66% of purchases on [e-commerce] sites happened without any pogosticking . . . the more [users] p...
Services with results<br />Save and export<br />Citation formatting<br />Tagging, editing, and sharing<br />Expert researc...
Going back to search<br />Problems with broadcast searching<br />Slow, dropped connections<br />Lowest-common denominator ...
Metasearch<br />Not all targets are created equal<br />
Not all targets are created equal<br />Metasearch<br />
Problems searching the catalog<br />Z39.50 searching not great<br />Limited search options<br />Browse searching not inher...
WorldCat and Ebsco APIs<br />WorldCat API<br />Free (to OCLC members) web service to WorldCat<br />Just ended pilot phase<...
A-9 style Search Results Page<br />Catalog(s)<br />Acad Search<br />Metalib<br />
A-9 style Psychology Search Results Page<br />Catalog(s)<br />PsycInfo<br />Metalib<br />
Hybrid<br />System<br />Metasearch<br />Other databases<br />Catalog<br />Main database<br />
From the catalog out?<br />Endecca<br />VUFind<br />Primo<br />Encore<br />WorldCat Local<br />
Todd Miller, Webfeat<br />
Library portal?<br />Metalib not flexible enough<br />Interface<br />Adding functionality<br />Integrating with other syst...
Upcoming SlideShare
Loading in …5
×

From metasearch to metaservices

1,784 views

Published on

Rethinking the goals and benefits of federated search

Published in: Technology, Education
  • Be the first to comment

From metasearch to metaservices

  1. 1. David Walker<br />California State University<br />From metasearch to metaservices<br />
  2. 2. Introduction<br />Part philosophy<br />What are we trying to do w/ metasearch?<br />Where are the problems?<br />How might we do it better?<br />Part practical<br />Waxing philosophical is not enough!<br />What can we do now with the tech we have?<br />
  3. 3. Rethinking metasearch?<br />We’ve been at it long enough now<br />Google Scholar<br />Newer systems<br />Next-generation catalogs? <br />Enterprise search? <br />Discovery layers?<br />
  4. 4. New metasearch models<br />Metalib X-Server<br />Design your own interface<br />Incorporate other systems and data<br />Experiment!<br />Xerxes Project<br />Developed by CSU and John Hopkins<br />Implemented at 20+ universities<br />
  5. 5. Xerxes<br />PHP / XSLT / MySQL<br />DB<br />X-Server<br />Metalib<br />KB<br />
  6. 6. What are we trying to accomplish?<br />“We offer a fragmented set of systems to search for published information . . . each with very different tools for identifying and obtaining materials. For the user, these distinctions are arbitrary.” – UC Bib Serv Taskforce<br />
  7. 7. Overview of the literature<br />1990-1999<br />“Is There a Chance for a Standardized User Interface?” –Fletcher<br />2000-2004<br />“Trumping Google? . . .” –Luther<br />“Talking about a Revolution? . . . ” –Nicholas<br />“The Answer to all of our Problems? . . .” –Groenewegen<br />“The Right Solution . . .” –Tennant<br />
  8. 8. Overview of the literature<br />2004-2008<br />“Is Metasearching Dead?” –Tennant<br />“Metasearching: Not as Good as We&apos;d Like It” –NLAQ<br />“Why Librarians Hate Metasearch” –McHale<br />“Plotting a New Course for Metasearch” –Breeding<br />
  9. 9. Arguing about metasearch<br />“[C]ross-database search tools . . . are the correct solution for unifying access to a variety of information resources.” –Roy Tennant<br />“Metasearching, then, is a step backward, a way of avoiding the learning process.” – William Frost<br />“[M]etasearch … cannot stand up to search systems based on centralized indexes” –Marshall Breeding<br />“Part of me keeps hoping [metasearch] will go away, but nope, it&apos;s still there.” – Andrew Pace<br />
  10. 10. Arguing about metasearch<br />Broadcast argument (pro)<br />“You can search multiple databases simultaneously!”<br />Nativist argument (anti)<br />“The search is not advanced enough!”<br />Immature technology argument (anti)<br />“The search is too slow!”<br />“Google Scholar is faster!”<br />
  11. 11. Isn’t it ironic?<br />Some librarians dislike metasearch because it is too much like Google; othersbecause it is not enough like Google.<br />
  12. 12. Usage statistics<br />SFX as a proxy measure?<br />Query # 2: Requests by source (SID)<br />Not all databases or clicks<br />Apples-to-apples comparison<br />Example Cal State campuses<br />Cal State Fullerton – general – 38,000 students<br />Cal Poly – science + engineering – 20,000 students<br />Sonoma State – liberal arts – 8,500 students<br />
  13. 13. Usage statistics<br />Fullerton<br />Native: 297,602 (54%)<br />Xerxes: 252,238 (45%)<br />Google: 5,340 (1%)<br />Cal Poly<br />Native: 95,885 (47%)<br />Xerxes: 92,534 (46%)<br />Google: 13,697 (7%)<br /><ul><li>Sonoma
  14. 14. Native: 22,654 (48%)
  15. 15. Xerxes: 22,275 (48%)
  16. 16. Google: 1,944 (4%)</li></ul>Native: 416,141 52%<br />Xerxes: 367,047 46%<br />Google: 20,98 3%<br />
  17. 17. Isn’t it ironic?<br />Usage of Google Scholar may depend in large part on whether librarians promote it or not!<br />
  18. 18. Usability studies<br />Universities<br />Boston College –BYU – Carnegie Mellon – Maryland – Mississippi – Northwestern – Oregon State – Rochester – Texas A & M –Colorado, Denver –California, Santa Cruz<br />Systems<br />Metalib – Encompass – Serials Solutions – Webfeat – LibraryFind <br />
  19. 19. Usability studies<br />70% of the students in BYU study preferred metasearch over native interfaces<br />“[B]oth [metasearch and searching native interfaces individually] produce citation sets of similar quality” – BYU <br />“Graduate students and faculty . . . all located citations they had not previously found” –Texas A&M<br />
  20. 20. On search times<br />“Eight [out of 18] students measured the speed of the search processing as reasonable and only five found the system too slow.” –Maryland<br />“Users are willing to wait as long as they think that they will get useful results. Their perceptions of time depend on this belief.” –Santa Cruz<br />“When people accomplish what they set out to do on a site, they perceive that site to be fast . . . If people can&apos;t find what they want on a site, they will regard the site as a waste of time (and slow).” – Perfetti, Landesman<br />
  21. 21. On the interface<br />“I found that both were not very user friendly.” – BYU Student<br />“[Non-federated search] lent itself to more abstracts . . . With [federated search] I was relying more on the title which can sometimes be misleading.” –BYU Student<br />“I would have to search through every single one of these to find which one is a scholarly article and which one is just a newspaper article.” – Maryland Student<br />
  22. 22. Broadening our goals<br />“[T]he point of federated searching is to make searching as simple as possible” –Cervone<br />Is it all about search?<br />What happens before you search?<br />What happens after you search?<br />Re-search is more than just searching<br />
  23. 23. Selecting the right resources<br />“Nothing slows the user&apos;s scanning momentum more than encountering results that are irrelevant . . . many users view it as a digital equivalent of Tourette&apos;s Syndrome, where the system just spits out random items, unrelated to their search.”<br /> – Jared Spool <br />
  24. 24. Selecting the right resources<br />Why not search everything?<br />Impractical technically<br />Impractical presentationally<br />What you don’t search is equally as important as what you do search<br />Why metasearch systems get this wrong<br />(Overly-) focused on the search box<br />Defining is not the same as limiting<br />
  25. 25.
  26. 26.
  27. 27.
  28. 28. Once you start down the Quick Search path, forever will it dominate your destiny.<br />
  29. 29.
  30. 30.
  31. 31.
  32. 32. Desperately seeking search box<br />“I just want a search box on the homepage.” – Your users<br />First Rule of Usability? Don&apos;t Listen to Users!<br />What users say they do is sometimes different from what they actually do<br />“The Google phenomenon”?<br />
  33. 33.
  34. 34.
  35. 35.
  36. 36. Desperately seeking search box<br />“[T]here is something inherent in the site&apos;s design that causes users to choose the search engine or the links, not a hard-and-fast preference of the user” <br />“[U]sers often gravitated to the search engine when the links on the page didn&apos;t satisfy them in some way.”<br /> – Jared Spool <br />
  37. 37. Isn’t it ironic?<br />The search box dominates the opening screens, then disappears!<br />
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42. Changing queries<br />“Nearly all students repeated their searches, changing terms or subject categories, so the interface needs to make this easy.”– Maryland <br />“Each new piece of information [users] encounter gives them new ideas and directions to follow and, consequently, a new conception of the query …  [T]he query itself (as well as the search terms used) is continually shifting, in part or whole.” –Bates <br />
  43. 43. Services for results<br />Spell check<br />Peer reviewed flag<br />Full-text look-up<br />Full-text linking<br />Format<br />Foreign language<br />
  44. 44. Avoid pogosticking<br />“66% of purchases on [e-commerce] sites happened without any pogosticking . . . the more [users] pogosticked, the less likely the session would result in a purchase.”<br />“The best search results pages will prevent pogosticking by providing the relevant content before the user chooses a specific result.”<br /> – Jared Spool <br />
  45. 45.
  46. 46.
  47. 47. Services with results<br />Save and export<br />Citation formatting<br />Tagging, editing, and sharing<br />Expert research help<br />
  48. 48. Going back to search<br />Problems with broadcast searching<br />Slow, dropped connections<br />Lowest-common denominator searching<br />Problems with central indexing<br />Not easy<br />Requires software, hardware, money, haggling<br />A middle ground?<br />
  49. 49. Metasearch<br />Not all targets are created equal<br />
  50. 50. Not all targets are created equal<br />Metasearch<br />
  51. 51.
  52. 52. Problems searching the catalog<br />Z39.50 searching not great<br />Limited search options<br />Browse searching not inherent in Metalib<br />Innovative ILS<br />Hit counts are wrong<br />Keyword results return results in bib id order<br />Not getting fixed any time soon.<br />
  53. 53. WorldCat and Ebsco APIs<br />WorldCat API<br />Free (to OCLC members) web service to WorldCat<br />Just ended pilot phase<br />Ebsco Integration Toolkit<br />Free (to Ebsco customers) web service to all Ebsco databases<br />Available now<br />
  54. 54.
  55. 55. A-9 style Search Results Page<br />Catalog(s)<br />Acad Search<br />Metalib<br />
  56. 56. A-9 style Psychology Search Results Page<br />Catalog(s)<br />PsycInfo<br />Metalib<br />
  57. 57. Hybrid<br />System<br />Metasearch<br />Other databases<br />Catalog<br />Main database<br />
  58. 58.
  59. 59.
  60. 60. From the catalog out?<br />Endecca<br />VUFind<br />Primo<br />Encore<br />WorldCat Local<br />
  61. 61. Todd Miller, Webfeat<br />
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68. Library portal?<br />Metalib not flexible enough<br />Interface<br />Adding functionality<br />Integrating with other systems<br />Xerxes should be<br />Everything is XML-based<br />Open source<br />
  69. 69. Toward a services layer<br />Adding value beyond the native interface<br />Consolidation before distribution<br />Come to the RSS presentation on Friday<br />Saving, tagging, sharing<br />The interface is the system<br />Metasearch, centrally indexed, hybrid . . .<br />Still need a good interface<br />
  70. 70. Conclusion<br />“You can search multiple databases at the same time” is not a compelling enough argument<br />We need to focus on the whole research process , not just search<br />Add value and layer functionality on top of the results<br />
  71. 71. Conclusion<br />We need an experimental platform to try new things, and an open source community to allow that to happen<br />xerxes.calstate.edu<br />

×