CS 502 : Computing Methods for Digital Libraries Spring 2001 -- Monday, Wednsday, Friday 02:30 - 03:20 PM Professor Herbert Van de Sompel 3 Credit Hours
		syllabus	assignments	code
home		administration	notices	background reading

Assignments

All assignments are to be submitted via e-mail to herbertv@cs.cornell.edu . The subject line of the e-mail MUST read: CS502, Assignment #n

Assignment #1, due 02/17 12:00 PM

Group assignement: Write a paper on the topic "Is the Web (as a whole) a Digital Library? "

For this assignement, you will be organized into groups of approximately 10 students. Each group delivers 1 co-authored paper. Get yourself organized in those groups; create a mailing list to discuss this question; get together in person for a discussion; create a draft document to be discussed; deliver a co-authored paper of maximum 6 pages. The groups are listed here.

1 person of each group e-mails the resulting paper (HTML format) to herbertv@cs.cornell.edu by 02/17 12:00 PM, with subject line: CS502, Assignment #1.

Grading will be based on structure of the paper, on the ideas/arguments that it introduces and whether these are relevant/convincing/original, not on conduct of the english language.

Tips:

• Use both perspectives on digital libraries, as introduced in lecture 1
• Relate to the typical functions of libraries [4S’s] (see lecture 1)
• Find inspiration in the readings for Week 1
• Find inspiration in lecture 1, lecture 2, lecture 3, and the lecture by John Saylor.
• Find inspiration in a comparative exploration of the Web and the Cornell Library Gateway.
• Include your own perspectives on libraries and the Web. Don't be afraid to be outspoken!
• The answer might not be black-or-white: maybe some aspects of the Web would qualify to be DL, some might not?
• Include references (hyperlink them!);
• Give appropriate credits!

Assignment #2, due 03/31 12:00 PM

Group assignement: Creation of composite digital objects

For this assignement, you will be organized into groups of approximately 6 students. Here are the groups.

Delivery: 1 person of each group sends an e-mail containing 2 PURL's to herbertv@cs.cornell.edu by 03/31 12:00 PM, with subject line: CS502, Assignment #2.

Starting point: Choose two pages with printed text, one english, one spanish. Make sure that the pages are different in nature.

Goal: Create two composite digital objects that contain various digital renderings of this source material. Use only computer-based techniques to create the renderings, i.e. never leave the digital domain once the printed material has been digitized.

Think Kahn-Willensky when structuring a rendering into a digital object and when structuring the various renderings into a composite digital object. Each composite digital object itself must be an HTML page accessible via a PURL [http://www.purl.org/]. The HTML page contains:

• information about the creators of the object
• descriptive metadata about the source material
• an entry per digital rendering of the source material. each entry has:
  • a PURL to access the rendering
  • technical metadata: information documenting the process that led to the rendering, information re the media format etc...
  • the data

Each composite digital object needs to contain AT LEAST the following renderings:

• scanned image (different formats)
• OCR-ed version (different formats) [http://www.tmsinc.com/prizmcorporate/]

But, be CREATIVE and try to think of other possible renderings of this source material.

• automatic translations [http://babelfish.altavista.digital.com/translate.dyn]
• turn text into phonemes [http://tcts.fpms.ac.be/synthesis/mbrola.html]
• turn phonemes into audio files [http://tcts.fpms.ac.be/synthesis/mbrola.html]
• ...

When moving from one generation of digital rendering to another, do NOT correct the errors in the output of a digitization/conversion process. Document the problems that are being introduced by the digitization/conversion as part of the technical metadata.

Assignment #3, due 05/05 12:00 PM

For this assignement, you will be organized into groups of approximately 6 students. Here are the groups.

Goal: The creation of a formal mechanism that allows for the verification of the validity of Open Archives Metadata Harvesting (OAMH) protocol requests.

The current specifications of the OAMH protocol state that an HTTP Status-Code 400 should be returned when the syntax of the protocol request is illegal, i.e.:

• Illegal verbs are used in the request;
• Illegal arguments are used in the request. This means:
  • The argument itself is illegal as such;
  • The argument is illegal in conjunction with a certain verb;
  • Illegal values are used for valid arguments, i.e. the values used do not comply to the rules that are specified in the protocol. For instance: incorrectly formated dates, metadataPrefixes that are not comformant to the specfied syntax for metadataprefixes, etc. This does NOT refer to the validity of requests in the context of a given repository. For instance: issuing a protocol request with a set- argument against a repository that has no set-structure is NOT illegal. However, issuing a protocol request with a value for set that set-argument does not comply to the specifications for set-values is illegal.

From an exploration of the repositories that have implemented the OAMH protocol, it can be seen that the interpretation of the above rules is very diverse, inconsistent and not even compliant with the specifications. Therefore, it would be helpful to have a more formal definition for valid protocol requests. Such a definition should formally specify what the the family of protocol requests is that must retrun a valid OAMH protocol reply. With such a definition in place, repositories can consistenly reply:

• with an OAI protocol reply (XML document) when the protocol request matches the formal definition;
• with a Status-Code 400 when a protocol request does not match the formal definition.

This assignment is about establishing and testing such a formal definition for OAMH protocol requests by means of an XML Schema:

• Task 1 in this assignment is to explore registered OAI repositories and to identify inconsistencies in their handling of the specifications re returning Status-Code 400 for illegal OAMH protocol requests. Part of this task is to report on such inconsistencies, by means of assembling at least 6 OAMH URLs that result in invalid error handling.
• Task 2 in this assignment is to create an XML Schema that defines the family of valid protocol requests in the most stringent possible manner. Use this version of the XML Schema specifications. The validity of this Schema itself must be tested using this XSV Schema validator.
• Task 3 is to convert 6 legal and 6 illegal (seeTask 1) protocol requests into XML instance documents (see example below). This can be done manually, or by means of the creation of a program, by the creation of a web-interface and CGI program, ...
• Task 4 is to manually test the comformance of the XML instance documents that result from Task 3 with the XML Schema that results from task 2. This validation must be done using this version of the XSV Schema validator.

Example, the request:

http://an.oa.org/request?verb=ListRecords&from=12-01-1999&until=14-01-1999&set=theset&metadataPrefix=oai_dc

could be rendered into the following XML document::

<?xml version="1.0" encoding="UTF-8" ?>
<oai-request namespace_stuff schemalocation_stuff>
<ListRecords>
<from>12-01-1999</from>
<until>14-01-1999</until>
<metadataPrefix>oai_dc</metadataPrefix>
<set>theset</set>
</ListRecords>
</oai-request>

The conformance of such an instance document with the XML Schema specified by "namespace_stuff schemalocation_stuff" (which is the result of Task 2) can be tested using the XSV Schema validator.

Delivery: 1 person of each group sends an e-mail containing a URL to herbertv@cs.cornell.edu by 05/05 12:00 PM, with subject line: CS502, Assignment #3. Accessible from that URL must be:

• the XML Schema (.xsd file)
• 6 URLs that are valid OAMH protocol request (use real requests targeting registered data providers) followed by:
  • the XML representation of those URLs (.xml file)
  • the XSV output of the validation of these XML representations (should validate)
• 6 URLs that are invalid OAMH protocol requests (use real requests targeting registered data providers) followed by:
  • the XML representation of those URLs (.xml file)
  • the XSV output of the validation of these XML representations (should not validate)
• maybe a web-form that allows input of URLs and transforms those URLs into XML documents?
• maybe an executable or a script that allows input of URLs and transforms those URLs into XML documents?

Herbert Van de Sompel