NOTE ON SEARCHING
The WTW archive uses XML-compliant DynaWeb software to convert TEI-conformant SGML- encoded texts to HTML for delivery over the Web. There are three modes of access to the texts:
IMPORTANT NOTE: Whichever access mode is used, the software will count the hits and highlight each one in red. With the first two types, the word highlighted is obviously linked to the search term (eg: "children" retrieves "children," or "children," "daughters," etc). But sometimes the word highlighted is not obviously linked. When analytical markup is used, the appropriate tag is attached to the first significant word in the sentence/paragraph--see the section on analytical categories in our Note on Encoding--which may not be a noun connoting the search term (eg: in one instance the word "fact" is highlighted in response to the "gender-female" search because it is the first significant word in a sentence that relates to, but does not contain the words, "female gender"). However, the reason for the highlighting becomes clear as the user reads and absorbs the content of the sentence.
Basic Full-text Search
This approach is used to retrieve all instances in the text(s) of the search term or phrase being used.
Example:
A search on "children" will retrieve all instances of the word "children" anywhere in the document header(s) or text(s).
To test this search, access the server and use the middle white search bar.
NOTE: This approach may be used with all our texts, at the collection or individual book level.
Expanded (Thesaurus-based) Search
This approach is used to retrieve not only instances of the search term, but also instances of related words in the text, as defined by inbuilt DynaWeb thesauri. (The thesauri in question are derived from Houghton-Mifflin thesauri, and are not changed in any way by WTW).
Example: A search on "children" will retrieve not only instances of the word "children" but also instances of words like "daughter," "descendant" anywhere in the document header(s) or text(s).
To test this search, access the server, use the middle white search bar, and check "Expanded Search".
NOTE: This approach may be used with all our texts, at the collection or individual book level.
Analytical (Metadata) Search
This approach is used to retrieve additional passages that reflect the content of the search but do not necessarily contain the search terms used. The software looks for analytical metadata tags created using SGML (Standard Generalized Markup Language) that are attached to the texts by our encoders. These tags represent SAMPLE research categories identified by our advisory board. (We provide these categories to demonstrate how researchers can easily create customized analytical categories using SGML; their primary purpose is to serve as models that researchers who wish to download our texts can adapt to their own purposes).
WTW analytical categories are clustered under four main themes--two "subjective" and two "objective"--as follows:
WTW ANALYTICAL CATEGORIES
Ethnicity
Used sparingly for references to ethnic groups, and
ONLY in cases where the discussion is extensive or marked.
Gender Marking
Used only for overt or marked references to gender.
Subcategories include:
Female; Male; Other.
Example: A search on "Transportation-Road" retrieves passages encoded with the ID "trans-road." (For a list of tagging IDs and typical instances of use, consult the section on analytical categories in our Note on Encoding). The passages may not contain the words "transportation" or "road," but discuss some aspect of road transportation.
To test this search, access the server, click on "Project Categories," and click on any choice.
NOTE 1: This approach will NOT at present find relevant passages in *all* our texts (some texts are still being encoded). Even so, several analytical searches can now be performed using the "Project Categories" menu.
NOTE 2: An additional "Course Categories" menu has just been introduced. This will contain categories developed in connection with individual courses, and taking this approach, users will retrieve passages only in selected texts, those encoded for the courses in question (whereas the user who takes the Project Categories approach will eventually be able to search those categories across all texts in the database).