User-Agent Requirements for CBD
Basic User Agent
Assumptions
- no network connection
- AND inadequate resources to run personal httpd
Minimal Functionality
- select dictionary articles
- view dictionary article outlines with type-example references
- view reference context
Fulfillable by
- html instantiation of CBD
- accessed by any web-browser
Advanced User Agent
Assumptions
- satisfactory network link
[definition of satisfactory depends on user]
- OR adequate resources to run personal httpd
Minimal Functionality
- All Basic UA functions
- optionally import full reference lists dynamically,
including uncategorized references
- re-sort reference lists by genre/time/place/orth
Fulfillable by
- xml instantiation of CBD
- web connection or user-local httpd with CGI capability
- XML-aware browser with CSS2, JS and XSL support
[Mozilla will meet this requirement]
- CGI-based interaction with query engine
- EITHER optimized XML Query language
[XPath not enough (running time issues for large datasets)]
- OR external query systems
[SQL/home-grown]
Creation/Exploration User Agent
Assumptions
- client JVM runnable on local machine
- satisfactory network connection and server JVM on remote machine
- OR server JVM also runnable on local machine
Minimal Functionality
- All Advanced UA functions
- perform queries on corpora outside context of specific articles
- sort and structure query results, mark for deletion or ignoring
- save and reload query results locally
- create new articles outside of search result context
- create new articles based on ongoing structuring and sorting of
references within search result context
- invisibly manage relationship between references and their
locations in the corpora
- permit inclusion of notes pointing to pertinent secondary
literature on specific references or on the word or meanings
of sections
- permit querying corpora for inclusion of specific texts or
fragments by common referents (e.g., museum-number of manuscripts;
principle publication)
- facilitate locating and referencing bibliographical items from
secondary literature
- invisibly manage relationship between selected bibliographical
items and their location in the bibliography list
- save and reload articles locally
- submit articles for inclusion in CBD
Desirable Functionality
- interoperability with corpus-editing tools
[e.g., click on reference to update corpus-reference for quick
fixing of typos etc.]
- EITHER support of scholarly authoring within UA
- OR easy exportation of (structured/sorted) result-sets for importation
to text-editors/word-processors [RTF export generated by XSL?]
Fulfillable by
- PDOM instantiation of CBD
- (Java-based) PDOM manager/server
[Ozone for now; PDOM manager and server could be two separate
programs; local-only access could be done by binding Java
client directly to PDOM]
- purpose-built Java client (using Swing etc., as much as
possible to reduce implementation burden)
- integrated interaction with query engine
[embedded XQL; system calls to query programs; JDBC]