SARA

1
SARA BNCweb-Specialised Concordancing Software

• Presentation by Susan Boser

2
SARA
3
What is SARA?

• SARA SGML Aware Retrieval Application
• Client/Server software tool to query the BNC
• Developed by Tony Dodd as part of the BNC Project
• Holds a central database of texts with SGML
  mark-up to be queried by clients

4
What does SARA?

• Enables the user to search through the BNC for
  examples of specific words, phrases, patterns of
  words etc.
• Examples can be sorted and displayed in different
  formats
• Searches can be limited to particular SGML
  contexts

• To particular kinds of texts
• Wild card or regular expression searching is
  possible
• Supports complex queries

5
5 Types of Queries

• Word Query
• Phrase Query
• POS Query

• Pattern Query
• SGML Query

6
Types of Queries 1
7
Types of Queries 2
8
Types of Queries 3
9
Types of Queries 4
10
Types of Queries 5
11
Complex QueriesQuery Builder

• Combines the different or same kind of the 5 main
  query options
• The parts are represented by nodes of various
  types
• Consists at least of two nodes
• Scope node defines the context (SGML element or
  Span(number of words))
• Content node may be linked together and edit
  type of query

12
CQL Query

• CQLCorpus Query Language
• Defines a query using own internal command
  language of SARA
• Atomic query
• A word, punctuation mark, or delimited string
  (e.g. jam, ?, Mrs.)
• A word-and-PoS pair (e.g. CANNN1)
• A phrase (e.g. not in your life)
• A pattern (e.g. colo?r)
• An SGML query (e.g. ltbodygt)
• Wildcard character _ (e.g. home _ center)

13
CQL 2

• Unary operators
• Case operator makes query case-sensitive
• Header _at_ operator makes query search within
  headers as well as bodies of texts
• Not ! Operator matches everything which is not a
  solution to the query (e.g. !cat dog finds
  occurrences of dog not preceded by cat)

14
CQL 3

• Binary operators
• Sequence blanks between two queries(e.g. cat
  dog)
• Disjunction operator matches cases which
  satisfy either query (e.g. cat dog)
• Join (order matters) and (order does not
  matter) operator match cases which satisfy both
  queries(e.g. cat dog)

15
Displaying the results

• Line mode, all occurences of the item displayed
  as a single line
• Page mode, each occurence displayed in full on
  the screen

16
Display Format

• PLAIN, displays only the words and punctuation of
  the results
• POS, displays the Part of Speech information for
  any word on the screen

• SGML, each result displayed with full SGML
  mark-up
• CUSTOM, displays each hit according to a
  user-defined format

17
BNCweb
18
What is BNCweb?

• User-friendly, web-based client program
• searches and retrieves for lexical, grammatical
  and textual data from BNC
• relies on SARA
• offers a wide range of additional features to
  SARA

19
Search Options

• 2 Main Options
• Standard Query
• Lemma Query

20
1 Standard Query

• Searches for words or phrases
• Searches are not sensitive for wordclass
  distinction (e.g. nouns, verbs, adjectives,
  adverbs, conjunctins etc.)
• Possibility to restrict searches to a subset of
  either written or spoken texts
• Subsets are defined by selecting relevant
  metatextual categories

21
2 Lemma Query

• Searches for words by additionally specifying the
  lemma type (nouns, verbs, adjectives etc.)
• Also possibility to restrict searches to a subset
  of written or spoken texts and its metatextual
  categories (compare Standard Query)

22
Other Functions

• Browse a file (by filename sentence number)
• Word lookup (produces alphabetically ordered
  lists of lexical items)
• Scan keyword/titles (retrieves a list of BNC text
  files on the basis of the classification
  contained in the "title" and "keyword" element of
  the file headers)
• Explore genre labels (retrieves a list of BNC
  text files according to genre classification
  criteria )
• Frequency lists

23
Display Options

• Sentence format, views concordance lines
• KWIC (Keyword-in-Context) format

24
SARA vs. BNCweb 1Features included in both

• Usual functions of
• thin delete
• sort save
• collocations
• Search words, phrases, patterns, POS
• Restrict search with the help of metatextual
  categories
• Bibliographical data available

25
SARA vs. BNCweb 2Advantages

• SARA
• Definition of complex queries with the help of
  Query Builder
• BNCweb
• User-friendly
• Works faster than SARA
• No restrictions on the number of search results
• Possibility of tag-sequence search (specify
  syntactic

• Structures within the 4 words
• Preceding and/or following the
• match of a given query)
• Offers detailed descriptive statistics for query
  results in the distribution option
• Gives both absolute numbers of hits and
  normalised frequency counts

26
SARA vs. BNCweb 3Disadvantages

• SARA
• Not very user-friendly (complex, complicated)
• Download restrictions (max 2000 hits)
• no facility for automatic generation of collocate
  lists or further statistical analysis
• No lemmatized index or lemmatizing component
• No possibility to define, save re-use
  subcorpora except by saving and

• re-using the queries which define them
• SARA client can address only the whole of the
  SARA index
• Cannot be used to search for patterns of POS
  codes without specifying the query by attaching a
  word

27
SARA vs. BNCweb 4Disadvantages

• BNCweb
• Can only be installed on the basis of a full
  installation of BNC text files, the software tool
  SARA (with its index files) and some additional
  UNIX-tools
• Only compatible for 2nd version of BNC (BNC World
  Edition)

• Complex query only possible if the user is aware
  of CQL

28
Sources

• Literature
• Aston, G./Burnard, L. (1998) The BNC Handbook.
  Exploring the British National Corpus with SARA.
  Edinburgh Edinburgh University Press.

• Internet Sources
• www.natcorp.ox.ac.uk/using/papers/burnard96a.htm
• http//homepage.mac.com/bncweb/manual/bncwebman-ho
  me.htm
• http//www.linguistlist.org/issues/13/13-2840.html
  
• http//www.natcorp.ox.ac.uk/sara/index.html