Metadata Scheme Architecture and Tool Evaluation Assignment Number 3

Schema Architecture Analysis

University of Oregon Libraries Digital Images Initiative Art & Architecture Images Metadata Scheme

     The University of Oregon Libraries Digital Images Initiative Art & Architecture Images Metadata Scheme is located online at http://www.uoregon.edu/~ehteague/aai-metadata.html. I could not find a written statement or specification instruction other than the outlined schema itself. However, the unstated aims and goals are to faciliate resource discovery of approximately 5,000 images of artwork and architecture located in their library. Their metadata standards and recommended practice guidelines are taken from many sources listed at http://libweb.uoregon.edu/catdept/meta/metatools.html. I could not find a page where they discussed which part of each guideline they used. Part of their rules for facilitating resource discovery are developed within a group of education institutions they belong to known as the Digital Library Federation (DLF). This group was started and is lead by Harvard University and the Massachusetts Institute of Technology (MIT).
     Their stated domain is the faculty, staff, and students of the University of Oregon. Of note, they also assist with image discovery from the library and textbooks used in, and for, the class ARH358. This group is sharing other library collections with the citizens of Oregon and Universities around the world. However, this particular collection of images is password protected to assure limitation to the intended domain. The object format is comprised of several sources starting with Dublin Core. They also include VRA Core 3, UO-AAI, and several data standards from the Getty.
     The architectural layout includes 5 basic query fields that are sub-divided into 25 field labels. Each basic query field and field labels range from specific labels such as the supplier serial number to highly granulated fields such as keywords to allow for a range of descriptive identifiers.

Metadata Tool Evaluation

Metadata generation tool Editor-Convertor Dublin Core Metadata(Ver.-1.31) at http://www.library.kr.ua/dc/dceditunie.html. This editor assists with automatic metadata abstraction with the ability to edit the results before asking for the final code.

Site #1: http://raymondcarney.com

Orginal result before editing:

<META NAME="DC.Title" CONTENT="Welcome to raymondcarney.com" >
<META NAME="DC.Description" CONTENT="www.RaymondCarney.com Raymond Carney Rn BSN home page">
<META NAME="DC.Subject" CONTENT="raymond, carney, raymond carney, unc, chapel hill, university of north carolina at chapel hill, carolina, carolina university">
<META NAME="DC.Type" CONTENT="Text">
<META NAME="DC.Format.extent" CONTENT="25 Kbytes">
<META NAME="DC.Date.DataGathered" CONTENT="2006-12-12">
<META NAME="DC.Rights" CONTENT="All text and pictures are copyrighted and can not be reproduced except under the fair use act without getting written permission from Raymond. ">
<META NAME="DC.Identifier" CONTENT="http://raymondcarney.com">

And with editing:

<META NAME="DC.Title" LANG="en" CONTENT="Welcome to raymondcarney.com">
<META NAME="DC.Subject.keywords" LANG="en" CONTENT="raymond, carney, raymond carney, unc, chapel hill, university of north carolina at chapel hill, carolina, carolina university">
<META NAME="DC.Description" CONTENT="Home Page of Raymond Carney Rn BSN ">
<META NAME="DC.Publisher.place" LANG="en" CONTENT="Chapel Hill, NC 27514">
<META NAME="DC.Date.DataGathered" SCHEME="DCMI Period" CONTENT="2006-12-12">
<META NAME="DC.Format" SCHEME="IMT" CONTENT=”text/html; charset=iso-8859-1">
<META NAME="DC.Format.extent" SCHEME="IMT" CONTENT="25 Kbytes">
<META NAME="DC.Type" SCHEME="DCMI Type" CONTENT="Text.Homepage.Personal">
<META NAME="DC.Identifier" SCHEME="URI" CONTENT="http://raymondcarney.com">
<META NAME="DC.Source" SCHEME="URL" LANG="en" CONTENT="Http://raymondcarney.com">
<META NAME="DC.Language" SCHEME="ISO 639-2" CONTENT="en">
<META NAME="DC.Rights" LANG="en" CONTENT="All text and pictures are copyrighted and can not be reproduced except under the fair use act without getting written permission from Raymond. ">

Site #2 http://google.com

<META NAME="DC.Title" CONTENT="Google">
<META NAME="DC.Format" CONTENT="text/html, text/html; charset=windows-1251">
<META NAME="DC.Type" CONTENT="Text">
<META NAME="DC.Format.extent" CONTENT="2 Kbytes">
<META NAME="DC.Date.DataGathered" CONTENT="2006-12-12">
<META NAME="DC.Identifier" CONTENT="http://google.com">

With editing:

<META NAME="DC.Title" LANG="en" CONTENT="Google">
<META NAME="DC.Creator.CorporateName" LANG="en" CONTENT="Google, Inc.">
<META NAME="DC.Subject.CorporateName" LANG="en" CONTENT="Google, Inc.">
<META NAME="DC.Description" LANG="en" CONTENT="Search engine user interface.">
<META NAME="DC.Publisher.CorporateName" LANG="en" CONTENT="Google, Inc.">
<META NAME="DC.Date.DataGathered" SCHEME="DCMI Period" CONTENT="2006-12-12">
<META NAME="DC.Format" SCHEME="IMT" CONTENT="text/html, text/html; charset=windows-1251">
<META NAME="DC.Format.extent" CONTENT="2 Kbytes">
<META NAME="DC.Type" SCHEME="DCMI Type" CONTENT="Text.Form">
<META NAME="DC.Identifier" SCHEME="URI" CONTENT="http://google.com">
<META NAME="DC.Source" SCHEME="URL" LANG="en" CONTENT="http://google.com">
<META NAME="DC.Language" SCHEME="ISO 639-2" CONTENT="en">
<META NAME="DC.Relation.References" LANG="en" CONTENT="world wide web">

I evaluated the metadata generation tool: Metadata generation tool Editor-Convector Dublin Core Metadata(Ver.-1.31). I found the evaluation tool easier to use than DCDot because more information was refilled for each website. The fields were easy to change before asking for the final version of the code. Since this tool is mainly for generating Dublin Core metadata, I think the definitions of the elements should accompany the form so I would not have to click back and forth between web sites to determine the appropriate information for each element.
I think this form can easily be extended to include more than just the 15 basic Dublin Core elements. This would make it a tool with a much wider use while keeping within Dublin Core schemes. In both cases of the web sites I chose, the metadata generation tool was able to generate most of the final metadata information with the tools' first attempt using the existing non-Dublin Core metadata and page content. This is a good start for the general public that would otherwise not produce any standardized metadata. If the element definitions were with the tool, most people able to create a web page would be competent to create quality machine readable metadata. With most of the blanks filled in, easily editable elements, and the code ready to copy and past, the web site creators of simpler sites could be encouraged to create the metadata if the need for metadata and these tools were more well known. I think this tool is a significant step in the creation of the semantic web.

Option 2: Research Question

Research Question Assignment #4

     What personal information would current non-Internet users state they are would be comfortable having accessible in public databases if they were themselves Internet users. With the current percentage of Internet users in the teens, clearly most of the world is not online. I have read many studies and thoughts about what kind of information the current online population puts online and what privacy concerns many deal with. What I don't see is the anticipated online privacy concerns of the majority of the world yet to get online.
     I don't think one can assume the current technological education level of Internet users represent our global population. This over 80% of our population of non users could have very different expectations about privacy. This population has not had the opportunity to create blogs and other possible portals of personal information the current group of the Internet has created. If the untaped Internet population has deep privacy concerns, a future semantic web as a large human social network looks grim. However, if this population looks at public online databases as little more than convenient telephone books and event calendars, “Joe Sixpack” might see having information someone can generally find anyway if they look hard enough as a positive influence and convince in their life.
     As the current Internet uses face the question of should each person have a personal identity only visible to those they choose, such as in health care information, the remaining population might be waiting for their chance to become visible to the world. In days past most people wanted their names published in the telephone books and their names and numbers available for long distance operators to give to those requesting the information. A small percentage of telephone users opted not to be listed for privacy reasons, but by and large most people were comfortable with being listed. Even with health care information, different populations would wear medical identification bracelets so health care providers would instantly know about a possible life threating problem as opposed to the general population feeling all health care information should be very private.
     I would propose a study to gather information about what private information non current Internet users would feel comfortable having put online. This information could assist with developing standards of what metadata would be included on a personal web site just as we have standards about different tags web sites are expected to have to be browser readable. This could lead to a separate standard used when browsers display web pages annotated with a document type definition stating the site was a homepage.
     My target audience would be non-Internet users. I would give information explaining how we use current public information such as phone books, libraries, and newspapers to gather information about ourselves and others. I would explain some of the privacy concerns current Internet users have developed and some of the possible reasons they have them. I would then ask them if they had an online presence such as a personal homepage, what information would they feel comfortable with other people online having easy access to. These questions would relate to information on drivers liences, addresses, phone numbers, family relationships, places of employment, history of education and employments, friends, neighbors, etc. With each question I would give a real life example of how this information might be very useful or very harmful.
     As an example, asking if someone would want to be semantically related to their neighbors. If you didn't enjoy visitors coming to your home you might not want your neighbors to know you were three houses away from them or you could be concerned with some safety reason if you belonged to a venerable population. The example of how many were against the Facebook mini feeds could be illustrated. However, if you didn't have any close family and you didn't show up for work, this information could be a way for your employer to possibly save your life by knowing who to ask to go over and check on you. This information could also be used to alert people within a certain distance from the residence of a neighbor that the neighbor was convicted of a crime. This conviction would already be public information and could help decrease crime if everyone knew all their semantically connected neighbors, family, friends, employers, etc. would be notified if they were convicted of a criminal offense.
     I don't think we have any idea what personal information the average person would want put on the web in a semantically designed metadata structure. This information would help guide our direction and efforts designing the Internet for users yet to join the approximate 1/5 of the population currently using the Internet and how semantic can our demands realistically be within the domain of personal information.

To learn about metadata etiquitte visit Metiquette.com