Vision and Scope Document

for
Spell Checker for
KnowledgeTrac
Version 1.0 approved
Prepared ! "ric "ng#uist
Decemer $% &00'
Copyright 1999 by Karl E. Wiegers. Permission is granted to use, modify, and distribute this document.
Vision and cope for pell Chec!er for Kno"ledge#rac Page ii
Table of Contents
Table of Contents.................................................................................................................ii
Revision History..................................................................................................................ii
1. Business Requirements..................................................................................................1
1.1. Background............................................................................................................1
1.2. Business Opportunity.............................................................................................1
1.. Business Ob!ectives and "uccess Criteria..............................................................1
1.#. Customer or $arket %eeds.....................................................................................1
1.&. Business Risks........................................................................................................2
2. 'ision of t(e "olution....................................................................................................2
2.1. 'ision "tatement.....................................................................................................2
2.2. $a!or )eatures........................................................................................................2
2.. *ssumptions and +ependencies.............................................................................
. "cope and ,imitations...................................................................................................
.1. "cope of -nitial Release..........................................................................................
.2. "cope of "ubsequent Releases...............................................................................
.. ,imitations and ./clusions....................................................................................#
#. Business Conte/t...........................................................................................................#
#.1. "take(older 0rofiles...............................................................................................#
#.2. 0ro!ect 0riorities.....................................................................................................#
#.. Operating .nvironment..........................................................................................&
Revision History
Name Date Reason For Changes Version
.ric .ngquist 1212312# -nitial Revision 1.2
$ark Biddlecom 2112412& Refined iteration 1 requirements after discussion
5it( stake(olders.
1.2
Vision and cope for pell Chec!er for Kno"ledge#rac Page 1
1. Business Requirements
The business requirements provide the foundation and reference for all detailed requirements
development
1.1. Background
The Spell Checker for KnowledgeTrac is intended to increase the efficacy of the KnowledgeTrac
search functionality by evaluating search terms submitted by the user, offering suggested
replacements for misspelled terms, and to assist the user in finding search results relevant to
their topic of interest.
1.2. Business Opportunity
This product will be valuable to the project sponsors because it adds value to eisting
KnowledgeTrac search technology and makes the eisting search capacities of KnowledgeTrac
more useful to its end users.
1.3. Business O!ecti"es and #uccess Criteria
!o quantifiable success criteria have been identified, but a complete solution will be effective in
assisting the end user to find search results useful to them without significantly compromising
the performance of the KnowledgeTrac search itself. Suggested alternate search terms will be
consistently intuitive and related to the original search words or phrases submitted. The project
sponsors will derive value from this project by improving the usefulness of its KnowledgeTrac
product to its end users.
1.$. Customer or %arket Needs
KnowledgeTrac satisfies the need of an internet user to perform directory"based searches to
obtain useful information as efficiently as possible. Spell Checker will provide a #shortcut$ for
users who have entered ineact or vague search terms, and offer suggestions of search terms
that would be useful to them.
Specifically, the completed system will satisfy the following requirements%
&. The system will run as a service or a 'eb service.
(. The system will load dictionary of terms at system start and sit and wait for an incoming
request, then return the best user definable number of matches. The calling process will
then decide how to handle the results.
). The system will have the ability to easily add words to a dictionary.
*. The system will have the ability to handle submissions via an application or a 'eb page.
+. The system will more heavily weight the best match toward a business name.
,. 'hen adding words, the system will ignore duplicates.
-. The system will have the ability to learn the #best match$ based upon past user selections
.. The system will associate a commonly misspelled word with a #best match$ until a history
of selections has been built.
1.&. Business Risks
!o major business risks have been identified.
Vision and cope for pell Chec!er for Kno"ledge#rac Page $
2. Vision o' the #o(ution
This section establishes a long"term vision for the system to be built to address the business
objectives. This vision will provide the contet for making decisions throughout the course of the
product development life cycle.
2.1. Vision #tatement
Spell Checker for KnowledgeTrac is a software component designed to assist users of
TC!$s KnowledgeTrac software in refining search terms to increase their success in finding web
and telephone directory results relevant to their search terms. /ur software will stand alone
from the larger KnowledgeTrac suite, analy0e search terms, and compare them to eisting
dictionaries containing 1nglish words, business names, and domain specific terminology to
suggest potentially more successful search terms. These dictionaries must support dynamic
addition of new phrases and terms. Spell Checker$s functionality will be transparent to the end
user and will silently analy0e search terms, only offering alternate suggestions if the search
terms are clearly unlikely to return useful search results.
Time permitting, Spell Checker will support recognition of common word2term
misspellings and suggestion of related topical phrases to the terms submitted. /ther
enhancements may be identified as the project progresses.
2.2. %a!or Features
3unctional 4equirements%
5ook up search terms in a dictionary
Suggest replacements for misspelled terms 6closest match7
o Support acronyms
8e able to enter acronyms as search strings 6results will include
epanded spelling7
8e able to enter epanded spelling 6results will include acronyms7
o Support company2contact names
o Support 1nglish words
9efault language can be modified by changing loaded dictionaries
:dd new terms to dictionary
;rocess phrases 6as opposed to single words7
Support multiple dictionaries
o <eneric 1nglish dictionary 6by default7
o 9omain specific terminology
o 8usiness names 2 contact info
4esults should favor non"generic dictionaries
!on"functional 4equirements%
/bject"oriented design to be implemented with =8.!1T
'ritten as a 'eb Service project using .!1T technologies
:daptability
o >ust support ability to work with different data stores 6e.g., S?5 98 or tet file7
o >ust support he addition of new components 6e.g., learning and synonym
finding7
;erformance
o :nalysis of a search string cannot take longer than one second.
o 1pected hardware% >ultiprocessors at (.@<A0 eachB (<8 4:>
>aintainability
o >ust be able to hand off code to TC! engineers
o Code must be readable and understandable
Vision and cope for pell Chec!er for Kno"ledge#rac Page %
2.3. )ssumptions and Dependencies
The following technologies have been identified as technical constraints. The system must be
compatible with%
'indows Server (@@@
>S CCS +.@
>S S?5 Server (@@@
The system is to be developed as a web service using =8.!1T.
3. #cope and *imitations
The project scope defines the concept and range of the proposed solution.
3.1. #cope o' +nitia( Re(ease
:n initial release has been planned to serve as a proof"of"concept for the final deliverable
product. This initial release 64&7 will consist of a working demonstration of the framework on
which the remainder of the project will be built. 4& will include%
Cteration &% 3ebruary &
st
(@@+
Complete system design for system iterations &")
Cnstructions for installation and integration with TC! client software
4esearch
o :nalysis of historic search strings and business names from TC!
o 9ictionaries 6common words7
o 'ord search algorithms
Tokeni0er
5etter frequency
9egree"of"difference calculus
;honetic analysis
Software
o Can check for the presence of a word in the dictionary using a quick binary
searchB will support the addition of other arbitrary algorithms and rules
governing their application
o 'orks from a base dictionary only, but can support an arbitrary number of
dictionaries
o 'ill load dictionaries into process memory to minimi0e processor time
devoted to C2/
9atabase integration
Testing
3.2. #cope o' #usequent Re(eases
Cteration (% 3ebruary &.
th
(@@+
Suggest replacements for words not in the dictionary
:ddition of a new search algorithm 6as yet undecided7 to provide more intelligent
searches
Dsing multiple dictionaries
Vision and cope for pell Chec!er for Kno"ledge#rac Page &
Dnit Testing for all written code
Cteration )% >arch (&
st
(@@+
9ynamically add words2phrases to the dictionary
Support phrase searching
:ddition of further search algorithms
<DC Configuration tool
Cteration * 6time permitting7%
5earning
Cteration + 6time permitting7%
Suggest related words2phrases
3.3. *imitations and ,-c(usions
!one identified.TC! will take care of getting their :S; search system to work with our web
service. 'e will deliver a working web service.
$. Business Conte-t
This section summari0es some of the business issues around the project, including profiles of
major customer categories, assumptions that went into the project concept, and the
management priorities for the project.
$.1. #takeho(der .ro'i(es
Stakeholders are individuals, groups, or organi0ations that are actively involved in a project, are
affected by its outcome, or can influence its outcome. The stakeholder profiles identify the
customers for this product and other stakeholders, and states their major interests in the
product.
Stakeholder
Major
Value Attitudes Major Interests Constraints
TCN
executives
increased
revenue
See product as
avenue to increase
in market share
richer feature set than
competitors; time to
market
N/A
TCN
engineers
Integration
with existing
software
Desire a product
that is intuitive,
maintaina!e, and
customi"a!e
#ood documentation,
no undue comp!exit$,
and re!iai!it$ of
services
%ust e
deve!oped with
techno!ogies
engineers are
fami!iar with
&IT Senior
'ro(ect Team
)ducationa!
va!ue
Consider pro(ect an
opportunit$ to app!$
academic and co*op
expertise
Comp!etion of
o(ectives using
!earned software
engineering discip!ines
Time +,-*week
senior pro(ect
duration.
&IT staff
advisor
/armon$
etween
pro(ect team
and sponsors
&eceptive as
resource for pro(ect
success
Time!$ comp!etion of
pro(ect and meeting of
a!! sponsor o(ectives
N/A
TCN 0ti!it$ of &e1uire ease of use Software must e 0npredictai!it$
Vision and cope for pell Chec!er for Kno"ledge#rac Page '
customers Spe!!
Checked
searches
and maximi"ation of
usefu!ness
simp!e to use and
provide easi!$
perceived search va!ue
of search input
$.2. .ro!ect .riorities
Dimension Driver
(state objective)
Constraint
(state limits)
Degree of Freedom
(state alloable range)
Schedu!e &e!ease 2 to e
avai!a!e $ ,/2/-3,
future re!eases as
descried in section 45,
6eatures 2--7 of ase
functiona!it$ must e
imp!emented in &2;
other functiona!it$ to
e further refined
Adaptai!it$ Spe!! Checker must e
modu!ar and support
addition of
supp!ementa!
dataase formats and
functiona!it$
%aintainai!it$ Spe!! Checker must e
deve!oped to e easi!$
modified and
maintained $ TCN
staff
'erformance Spe!! Checker must not de!a$
existing search functiona!it$
$ more than severa! seconds
Staff Team si"e is 8
deve!opers
Time Schedu!e s!ippage of more
than 39 hours re1uires
schedu!e revision
$.3. Operating ,n"ironment
This system will operate on a 'indows Server (@@@ machine.