Networked Computer Science Technical Reference Library

OAI-NCSTRL Search TIPS

Limitations Of OAI-NCSTRL Search Service

The current OAI-NCSTRL search engine has a few limitations:

1. Browsing by cluster (i.e., a search with no keywords, just cluster values) no longer returns all members of a cluster.  Only a partial list is returned. This is caused by different authority file (subject, type, language, publication date) used by data providers.   Please be aware that browsing gives incomplete results.

We apologize for these limitations and we are working to resolve them by refinng the search engine.

Help on Searching

Search for an document by seeking word matches on ALL bibliographic fields or on SPECIFIC bibliographic fields.   You can specify how your search results will be sorted.

Searching ALL bibliographic fields

The top of the search form presents a single text entry box. Enter keywords in this box (e.g., digital library) to search the author, title, and abstract fields of the bibliographic entries for all documents in the OAI-NCSTRL collection. The search returns documents whose bibliographic entries contain all the words you entered in either the author, title, type, language,,archive,subject, accession date, discovery date or abstract fields.  If you wish to enter a phrase, put it in quotes.

There is a pulldown menu to specify how your search results will be grouped and sorted .
 

Sorting of results

There is a pulldown menu for either type of searching that allows you to specify the sorting of search results.  Search results can be sorted by  rank, by accession date, by archive.  If there is a "tie" (e.g. there are many documents with the same archive), the secondary sort is by Oracle Intermedia search rank.
 

Grouping of results

    There is a pulldown menu for choosing the manner in which the results are to be grouped. Search results may be grouped according to Archive, accession year, ,subject.

Rules for bibliographic keyword matching

Words that you enter in the three bibliographic keyword fields (Author, Title, Abstract) are matched to bibliographic entries according to the following rules:

  • Each word that you enter matches any word in the respective field that begins with respective word. For example, the word "comp" matches "computer", "computation", "comprehensive", etc.
  • You may choose either the logical connector AND or OR between the bibliographic keyword fields by using one of the radio buttons below the fields. For example, if you select the or button and the word "robot" in the Title field and "robotics" in the Abstract field, the search will return documents that have either "robot" in their titles or "robotics" in their abstracts. If you select the and button, the search will return only those documents that have "robot" in their titles and "robotics" in their abstracts.
  • You may also use logical connectors "and" and "or" within fields. For example, if you enter "robotics or vision" in the abstract field, the search will return documents that have the word "robotics" or "vision" in their abstracts. If you enter "robotics and vision" in the abstract field, the search will return documents that have both the word "robotics" and "vision" in their abstracts. Multiple words that are not separated by "and" are assumed to be "and" separated. For example, if you enter "computer vision" in the abstract field, the search will return documents that have both the words "computer" and "vision" in their abstracts. Finally, you may use parentheses to group words. For example, if you enter "Gries or (Teitelbaum and Field)" in the author field, the search will return documents authored by "Gries" or by "Teitelbaum" and "Field".

Rules for advanced search

In this you may specify the author,the title or some keywords appearing in the abstract. The search form below the three text entry fields for bibliographic keyword entry, has two radio buttons to specify whether the field match specifications entered should be "anded" or "ored" together.
To specify search criteria, enter keywords in one or more of the text entry boxes beside the author, title and abstract labels (see the rules for bibliographic keyword matching). A keyword may be a quoted phrase (e.g., "digital library") to specify that the words should be contiguous.

  • Author - The author's first or last name or names .
  • Title - The words in the title of the document.
  • Abstract - The words in the abstract of the document.

and select the "AND" or "OR" radio button to indicate how you want the different field search specifications to be combined.

Please note that not all documents in the collection have abstracts. If you want your search request to include documents without abstracts, enter your search criteria under Title as well as Abstract.

The filters provide a means by which  a user can specify some details of the document he is seeking.
The filters are

  • Archive - The entier digital library may be divided into archives depending on the archive from which the document has been obtained. The different archive types are descibed by their respective home pages whose links are provided at the footer of the search form.
  • Archives Set -  Archives may have sub collections in them this pull down menu lets the user select the collection he wishes to search. by default it is "All".
  • Subject - The user can also select the subject of the document he is searching.
  • Type - This lets the user specify what type of document he is searching. For example if the document is a patent or a doctoral thesis.
  • DateStamp - A datestamp is the date of creation, deletion, or latest date of modification of an item, the effect of which is a change in the metadata of a record disseminated from that item.
  • Discovery Date -

Display Option

    They are the same as mentioned above.
 

OAI-NCSTRL clusters

    A Cluster is a collection of documents all of them sharing the same value for a particular attribute. For example all the documents in the archive type xxx, have the same attribute of the source digital library being that of XXX. This type of organization of documents improves the searching capabilities of the digital library. This allows users to specify the exact type of document they are looking for. For example a user may wisht to view documents by a specific author which has been published in a specific archive. This can be done by specifying the necessary filters.
 

                               Cluster information for Digital libraries

The entier document database has been indexed along the following indices
 

  • ID                -     ID of the document (the equivalent of the handle in NCSTRL.
  • TITLE          -     Title of the document.
  • AUTHOR     -     Authors / authors of the document.
  • SUBJECT -     This is divided into subject and sub subject fields eg. physics/high energy physics.
  • ARCHIVE -     This specifies which is the original digital library from which the document has been obtained eg.XXX
  • SET     -     This specifies the collection from which teh document hails.
  • ABSTRACT          -     The keywords of all the abstracts in the digital library are indexed.


Working of the Advanced search

When the user types in the search string and a few of the filter arguments, the search engine searches the respective indices for the matches and retrieves the "sets" of  document Id of the matches for that index and then  it does an intersection of these "sets" and displays its results based on type of grouping (Eg. group by Archive type) that is expected by the user.


 


 
 
 

Further help

Need more help? Feel free to e-mail us at info@ncstrl.org