Research interests 9/12
In reading the article by Paul Cron on selecting research topics & developing an overall framework of research, I began thinking about all of the ideas that I have written down over the last year and was wondering whether or not some of these ideas work together, represent good research areas, and whether or not any of them are specific enough to warrant further consideration. I decided to put them all in this entry and attempt to group/sort them out. . .
1. The integration of Traditional MARC, Amazon, Google, Ebay and other vendor webservice provided data to create more rich data set for resource description. Started with idea to link to images in Amazon, grew to include migration of three standards to new MD scheme (MARCXML, AMAZONMD, QDC), representation of information in faceted contexts. I modified a small movie search page based on this investigation that harvests metadata from Amazon through their API & merges it with the MARC data for videos in the ZSR library catalog. The result was promising but I still don't see the 'question' here.
2. To investigate the notion that there is a gap between the information organization skills in a librarian and the common man (whether subject expert or not). Marcia bates presents this view in her paper - the invisible substrate of information science -. I have two projects in the works related to this idea, the Kent Mcconnell project and the digital Forsyth project both of which are involving having users perform cataloging on primary source material. How does this mesh with my research focus on how users approach raw metadata? The hypothesis is that users are becoming more data/information focused and less container/interface focused and that this shift is a natural transition reflected in the shift from print to e, from info as object to info as knowledge, from webpage to wiki? Is it that people bring information organization skills to an interaction and as a result, it makes little sense to prescribe an info organization rule on them via some interface ? that usability is inevitably caught up in providing a least common denominator approach to presentation? (think Google - which answers questions very easily but does not always encourage further development of new questions).
In the Dr. McConnell project, we are teaming with a class to complete content description of digitized primary source materials. 13 students will complete the description of 300+ articles selected by Dr. McConnell. Their description will involve some cataloging principles (verification of already entered data, application of structured headings) and will involve some unstructured description (application of topical refinements, entry of an open 'aboutness' description, definition of new headings). Fortunately or not the system lacks some Authority control at the moment but I plan to allow the users to collectively define Authority by building their own vocabulary list. How does this turn into research? Do I compare the work done by the students with the work done by our catalogers on similar resources? Do I compare their descriptions against the work done previously in indexing the resource? I am looking forward to discovering how much/what sorts of intellectual effort the students apply to this project.
The digital Forsyth project is in its infancy but is a collaborative project between libraries in Forsyth County. My vision is to not only build a collective library of digital resources but to enable user submission & tagging of these images. This project will be several years in the making.
3. How would you use collaborative tagging with iterative automation to build and refine controlled vocabularies? There has got to be a host of methods (automatic determination of hierarchy, processing of personal names, user based metadata tagging of concepts) to make this process possible. What if you gave the user tools to mark important concepts in the text (forget tags, give them an editing tool) like highlighting with different colors to represent tags. A highlight process that when stopped would allow you to pick which tag you wanted to associate with (author, title, subject concept, quote, etc).
I am hoping to try some of these methods on the Dr. McConnell project as we are emphasizing to students that they should follow their own descriptive ideas. I am hoping to be able to extract some good data from this process.
4. What automatic metadata generation functionalities are currently supported in Digital Library systems? Do these functionalities center on cross-application themes or are they specific to the application? What AMG functions do users expect/want & how would they use them? To what extent can AMG facilitate obtain/organize/present functions? A sub-question that came from this investigation is: What are the primary purposes of existing DL systems? How does their system orientation impact intended uses & workflows? Are these systems being designed within a single framework or multiple? What could research contribute to this process? Is this a situation like the ERMS/DLF initiative that would benefit from working around a central standard//framework? Are those guidelines definable?
I was fortunate enough to work on this project with Dr. Greenberg over the summer. During the project I reviewed ~6 open source digital library systems and at the least developed a foundation understanding of each system's workflows, design focus & base functionality. I am continuing this research with Dr. Greenberg in the fall.
5. In what ways do current information & library systems emphasize interoperability? What actions are enabled by data interoperability in these systems (data harvesting, resource identification, resource location, seamless service request, database building, time saving???). Is OpenURL a sufficient standard/technology to allow significant interoperability and user initiated data harvesting? What would a system built on interoperability centric standards (such as OpenURL & RSS) look like? What features would it have & how would it fit within current library system models? What other standards would be required? What features would it support? Does such a system represent a significant change in the approach of system use in research?
This question drove an evolving paper from my first semester in Dr. Greenberg's metadata class. During the investigation process, I designed & built a prototype system which emphasized the use of interoperability based standards (OpenURL, RSS, NCIP) in creating a central resource & service request system. The underlying idea was that the traditional ILS public applications still operated as endpoint silos where Internet based research required users to dynamically connect information from multiple systems. The system I developed emphasized the utility of the data rich linking capabilities of OpenURL and implemented resource obtainment & service request / delivery within this model. Some example services I included were, development of a temporary resource list, export of data in multiple formats (EndNote, RSS (XML), HTML), service request initiation from a data rich link, back-end resource availability resolution, back-end circulation functionality. The idea was based on the idea of merging OpenURL functionality (an application sitting between disparate search engines) and Metasearch applications (Single place to go for information). I stated at some point in the paper, that while Metasearch applications were fundamentally flawed because they started with homogenizing data that an OpenURL based meta-service application allowed the user to get Meta services with data rich links, putting centrality to the research where appropriate, not at the resource discovery level but at the resource management, procurement, and research-management process. I enjoyed working on this topic & actually got a publication & poster presentation from it but am not sure how to move further past the 'interesting idea' phase.
The article on Emotive cognitive psychology (Clark) discussed how good information systems changed the way people interact with information & make a higher interaction/achievement possible by providing some service that would otherwise consume a large amount of time/energy. While I suspect that this system serves this purpose, I don't know how I would test that idea. . .My suspicion is that as users are becoming more data aware they not only implicitly understand the idea of interoperability - they are starting to expect it.
6. How do individuals choose to integrate information systems into their work systems? What reasons do they have for either following the integrated model or choosing other systems in a more decentralized approach? What consequences does this mean for system design? This idea comes in part from the Cron reading today and was one that took hold. In my own experience, I have seen a wide level of adoption of information systems. More often than not, the large unwieldy information system fails a full implementation because it includes too many assumptions. it seems to me that smaller (necessarily interoperable!) components would represent a better structure. How is this studied though? Do i choose a type of system (say acquisitions departments in academic libraries) and investigate thier workflows & information systems use in order to see where/why they diverge? I suppose that this could be extractable to some model (probably already been done!).
7. And finally, In the 881 class I recently articulated an interest in discovering how users are approaching the use of data outside of traditional interfaces. My underlying thought is that as technology has developed that people are becoming more familiar with the idea of described data (meta) & are starting to expect not only user-centric design but actual access to the data. For instance, Banks now make xls & quicken exports available to their web-users, Google & Amazon offer APIs for anyone who is interested. This is based in part on the idea of remix culture & mashups and in part on the idea that the best interface/information system is the one that a person identifies/builds for themselves. I suspect that my interest in this might only be shared by others in my position & not by the more general user but am not sure.
I was thinking that some research in this area could include, a survey of data export/interoperability functionality for a given topic/function area and an investigation of the use of those features as well as a investigation of extended (non-traditional) uses of the data.
8. Related to the above concept is the idea that personal information management is a growing field. While this is not some area I know much about, I am wondering what approaches are being used in PIM scenarios. Is this an area (like privacy) that is ultimately difficult to pin down or is there some simple research that could be conducted.
1. The integration of Traditional MARC, Amazon, Google, Ebay and other vendor webservice provided data to create more rich data set for resource description. Started with idea to link to images in Amazon, grew to include migration of three standards to new MD scheme (MARCXML, AMAZONMD, QDC), representation of information in faceted contexts. I modified a small movie search page based on this investigation that harvests metadata from Amazon through their API & merges it with the MARC data for videos in the ZSR library catalog. The result was promising but I still don't see the 'question' here.
2. To investigate the notion that there is a gap between the information organization skills in a librarian and the common man (whether subject expert or not). Marcia bates presents this view in her paper - the invisible substrate of information science -. I have two projects in the works related to this idea, the Kent Mcconnell project and the digital Forsyth project both of which are involving having users perform cataloging on primary source material. How does this mesh with my research focus on how users approach raw metadata? The hypothesis is that users are becoming more data/information focused and less container/interface focused and that this shift is a natural transition reflected in the shift from print to e, from info as object to info as knowledge, from webpage to wiki? Is it that people bring information organization skills to an interaction and as a result, it makes little sense to prescribe an info organization rule on them via some interface ? that usability is inevitably caught up in providing a least common denominator approach to presentation? (think Google - which answers questions very easily but does not always encourage further development of new questions).
In the Dr. McConnell project, we are teaming with a class to complete content description of digitized primary source materials. 13 students will complete the description of 300+ articles selected by Dr. McConnell. Their description will involve some cataloging principles (verification of already entered data, application of structured headings) and will involve some unstructured description (application of topical refinements, entry of an open 'aboutness' description, definition of new headings). Fortunately or not the system lacks some Authority control at the moment but I plan to allow the users to collectively define Authority by building their own vocabulary list. How does this turn into research? Do I compare the work done by the students with the work done by our catalogers on similar resources? Do I compare their descriptions against the work done previously in indexing the resource? I am looking forward to discovering how much/what sorts of intellectual effort the students apply to this project.
The digital Forsyth project is in its infancy but is a collaborative project between libraries in Forsyth County. My vision is to not only build a collective library of digital resources but to enable user submission & tagging of these images. This project will be several years in the making.
3. How would you use collaborative tagging with iterative automation to build and refine controlled vocabularies? There has got to be a host of methods (automatic determination of hierarchy, processing of personal names, user based metadata tagging of concepts) to make this process possible. What if you gave the user tools to mark important concepts in the text (forget tags, give them an editing tool) like highlighting with different colors to represent tags. A highlight process that when stopped would allow you to pick which tag you wanted to associate with (author, title, subject concept, quote, etc).
I am hoping to try some of these methods on the Dr. McConnell project as we are emphasizing to students that they should follow their own descriptive ideas. I am hoping to be able to extract some good data from this process.
4. What automatic metadata generation functionalities are currently supported in Digital Library systems? Do these functionalities center on cross-application themes or are they specific to the application? What AMG functions do users expect/want & how would they use them? To what extent can AMG facilitate obtain/organize/present functions? A sub-question that came from this investigation is: What are the primary purposes of existing DL systems? How does their system orientation impact intended uses & workflows? Are these systems being designed within a single framework or multiple? What could research contribute to this process? Is this a situation like the ERMS/DLF initiative that would benefit from working around a central standard//framework? Are those guidelines definable?
I was fortunate enough to work on this project with Dr. Greenberg over the summer. During the project I reviewed ~6 open source digital library systems and at the least developed a foundation understanding of each system's workflows, design focus & base functionality. I am continuing this research with Dr. Greenberg in the fall.
5. In what ways do current information & library systems emphasize interoperability? What actions are enabled by data interoperability in these systems (data harvesting, resource identification, resource location, seamless service request, database building, time saving???). Is OpenURL a sufficient standard/technology to allow significant interoperability and user initiated data harvesting? What would a system built on interoperability centric standards (such as OpenURL & RSS) look like? What features would it have & how would it fit within current library system models? What other standards would be required? What features would it support? Does such a system represent a significant change in the approach of system use in research?
This question drove an evolving paper from my first semester in Dr. Greenberg's metadata class. During the investigation process, I designed & built a prototype system which emphasized the use of interoperability based standards (OpenURL, RSS, NCIP) in creating a central resource & service request system. The underlying idea was that the traditional ILS public applications still operated as endpoint silos where Internet based research required users to dynamically connect information from multiple systems. The system I developed emphasized the utility of the data rich linking capabilities of OpenURL and implemented resource obtainment & service request / delivery within this model. Some example services I included were, development of a temporary resource list, export of data in multiple formats (EndNote, RSS (XML), HTML), service request initiation from a data rich link, back-end resource availability resolution, back-end circulation functionality. The idea was based on the idea of merging OpenURL functionality (an application sitting between disparate search engines) and Metasearch applications (Single place to go for information). I stated at some point in the paper, that while Metasearch applications were fundamentally flawed because they started with homogenizing data that an OpenURL based meta-service application allowed the user to get Meta services with data rich links, putting centrality to the research where appropriate, not at the resource discovery level but at the resource management, procurement, and research-management process. I enjoyed working on this topic & actually got a publication & poster presentation from it but am not sure how to move further past the 'interesting idea' phase.
The article on Emotive cognitive psychology (Clark) discussed how good information systems changed the way people interact with information & make a higher interaction/achievement possible by providing some service that would otherwise consume a large amount of time/energy. While I suspect that this system serves this purpose, I don't know how I would test that idea. . .My suspicion is that as users are becoming more data aware they not only implicitly understand the idea of interoperability - they are starting to expect it.
6. How do individuals choose to integrate information systems into their work systems? What reasons do they have for either following the integrated model or choosing other systems in a more decentralized approach? What consequences does this mean for system design? This idea comes in part from the Cron reading today and was one that took hold. In my own experience, I have seen a wide level of adoption of information systems. More often than not, the large unwieldy information system fails a full implementation because it includes too many assumptions. it seems to me that smaller (necessarily interoperable!) components would represent a better structure. How is this studied though? Do i choose a type of system (say acquisitions departments in academic libraries) and investigate thier workflows & information systems use in order to see where/why they diverge? I suppose that this could be extractable to some model (probably already been done!).
7. And finally, In the 881 class I recently articulated an interest in discovering how users are approaching the use of data outside of traditional interfaces. My underlying thought is that as technology has developed that people are becoming more familiar with the idea of described data (meta) & are starting to expect not only user-centric design but actual access to the data. For instance, Banks now make xls & quicken exports available to their web-users, Google & Amazon offer APIs for anyone who is interested. This is based in part on the idea of remix culture & mashups and in part on the idea that the best interface/information system is the one that a person identifies/builds for themselves. I suspect that my interest in this might only be shared by others in my position & not by the more general user but am not sure.
I was thinking that some research in this area could include, a survey of data export/interoperability functionality for a given topic/function area and an investigation of the use of those features as well as a investigation of extended (non-traditional) uses of the data.
8. Related to the above concept is the idea that personal information management is a growing field. While this is not some area I know much about, I am wondering what approaches are being used in PIM scenarios. Is this an area (like privacy) that is ultimately difficult to pin down or is there some simple research that could be conducted.
posted by Erik Mitchell at 5:40 AM
0 Comments:
Post a Comment
<< Home