Witt Project Web Site

This page contains additional resources related to the manuscript "Automatically Categorizing Software Technologies" submitted to TSE by Nassif, Treude, Robillard

Search Term Replacement

Replacements for search terms. This table refers to the heuristic described in Section IV.A, "Performing a General Wikipedia Search"
Original Sequence	Replacement
ms	microsoft
vb.net	visual basic .net
vb	visual basic
vc	microsoft visual c++
[single letter]	[single letter] programming

Acronym Expansion

Expansions for Acronyms. This table refers to the heuristic described in Section VI.A "3) Acronyms"
Acronym	Replacement
osx	os x
os	operating system
ide	integrated development environment
dbms	database management system
rdbms	relational database management system
api	application programming interface
ui	user interface
gui	graphical user interface
sdk	software development kit

Tag Examples from Each Set

To get a sense of what the actual data is like, here are 50 tags from each of the 3 evaluation sets. Those tags have been chosen ramdomly from each set, each tag having a similar probability to be selected to all of the other tags from the same set. The order of the tags in the lists is also completely random.

Popular Tags (50 out of 301 tags)

ios
object
joomla
parsing
numpy
dictionary
image
java
sqlite
java-ee
vim
xcode
backbone.js
cookies
android-listview
data
search
ipad
selenium
database-design
mysql
data-structures
ruby-on-rails-3
html5
hash
c#
uiview
zend-framework
session
authentication
events
powershell
perl
excel-vba
git
heroku
postgresql
input
grails
sql-server-2005
excel
twitter
loops
android-layout
sql-server
jsp
django
security
cocoa
c++11

Common Tags (50 out of 10,381)

alphabet
pthreads
textscan
catiledlayer
oracle-xe
m2eclipse
argumentexception
hudson-plugins
postgresql-9.2
alert
datacontract
clgeocoder
conditional
modx
opa
ironpython
jquery-accordion
dynamic-forms
cloudera
dropwizard
image-resizing
constraint-programming
python-module
google-places
onactivityresult
javadoc
svc
xalan
modernizr
mjpeg
rijndaelmanaged
android-alertdialog
marshalling
google-nativeclient
autodesk
associative
jquery-post
url-rewrite-module
calling-convention
httpclient
syscall
sicp
progress
file-management
bufferedwriter
pyglet
w3c-validation
asp.net-mvc-areas
complete
command-line-arguments

Rare Tags (50 out of 27,208)

ikimagebrowsercell
greybox
eventargs
cddvd
r-forge
misfire-instruction
olingo
processors
uitabcontroller
wgl
tjws
artistic-license
ubuntu-unity
flashwindowex
contiguous
netburner
offscreentabs
fraud-prevention
relstorage
optionbutton
error-suppression
ora-06502
django-tests
winmerge
bento
node-blade
autoreconf
css-validator
2d-engine
iobserver
xpath-api
westwind
com0com
blaze
openid-selector
expect.js
capstone
postman
syn
datamaps
avm2
mit-scheme
method-group
either
mix
late-static-binding
opennetcf.ioc
orangehrm
detailtextlabel
udt

Coding Guide

Given a phrase and a list of candidate hypernyms, for each candidate in the list, give it one of three values:

Correct: The candidate is truly a more general concept encompassing the input phrase. The phase “is a” hypernym, AND the candidate is not a general concept (see next point).
General: The candidate is correct according to the definition above, but is a general concept that could apply to a majority of the phrases. Examples include “library”, “object”, “structure”, “technique”, “software”, “framework”, “website”. General concept with a general attribute, such as “open-source” or “Google”, should also be considered as “General”.
Incorrect: The candidate is not a hypernym of the input phrase.
Alternate: The candidate is a correct hypernym but for a sense unrelated to programming. For example, “company”.

Special Cases:

Treat a plural candidate phrase as if it were in the singular. “Math libraries” is the same as “Math library”.
Treat CamelCase words and words separated with underscore as split. “MathLibrary” is the same as “Math Library”
Ignore case
Nonsensical and ungrammatical candidates are incorrect.
If candidate is the original phrase followed by a words (in parentheses or not), ignore the original phrase.
When a term includes the “by” preposition, flag as incorrect
When the candidate is the same as the phrase, flag as incorrect
If the phrase includes a date, flag as incorrect
Simple expansions of acronyms are incorrect
If in doubt, err on the side of incorrect

Terms Considered too General:

act
activity
add in
addon
aggregator
api
app
appliance
application
area
arrangement
artifact
calculation
checker
class
client
collection
column
command
component
concept
constituent
container
control
controller
convention
creator
cross
data collection
decoration
depiction
description
design
device
diagram
directed graph
document
effort
element
engine
entity
environment
event
exception
executable
experience
expression
extension
feature
field
file
format
framework
freeware
function
game
gateway
gathering
generator
hardware
idea
idiom
information appliance
infrastructure
integrated environment
interface
internet appliance
interpreter
investigation
label
language
library
liquid
location
measurement
mechanism
method
methodology
mode
model
module
name
network
new port
object
operation
package
paradigm
part
phrase
platform
platform as a service
plugin
portion
presentation
principle
problem
process
product
program
project
protocol
release
representation
resource
response
sequence
server
service
skin
software
software environment
software factory
specification
standard
strategy
string
structure
subclass
subroutine
subsystem
symbol
system
table
technique
technology
template
term
text
textual matter
tool
tool suite
tooling
toolkit
toolset
type
unit
utility
utilization
value
verbal description
view
viewer
visual communication
website
widget
window
wrapper

50 Largest Categories Produced by Witt

To understand the added value of Witt_C/CA over Witt_H, we provide the 50 largest categories returned by both variants, together with one member of each category to help understand potentially ambiguous terms. These members are provided only as example of the corresponding category. They may also be in other categories.

Witt_H
Category	Size	Tag Example
open source	298	torch
company	254	avaya
process	184	amalgamation
software	146	powertab
programming language	145	groovy-2.3
multi-paradigm programming language	120	perl5.12
web application framework	117	grails-2.0.4
functional programming language	113	haskell
integrated development environment	104	xcode4.1
imperative programming language	96	algol
library	90	rodf
object-oriented programming language	86	python-3.1
free software	84	ikiwiki
website	82	justin.tv
computer program	71	virus
tool	70	grako
framework	62	aurora.js
operating system	62	windows-3.1
procedural programming language	57	scheme48
method	56	linear-search
c	54	char32-t
open source web application framework	54	django-1.4
software framework	53	lime
application programming interface	52	opengl-2.0
general-purpose programming language	51	ruby-2.1
javascript library	51	planetary.js
function	51	identity-operator
ide	48	visual-c++-2013
reflective programming language	48	r
api	48	webcl
technique	47	record-locking
file format	46	.war
class	46	uiwebview
term	45	truthiness
data structure	44	binary-tree
way	40	orientation
unix-like os	39	android
mobile operating system	39	android
web browser	37	internet-explorer
open-source	36	git
graphical user interface	35	windows-8
command	35	kill
scripting programming language	34	javascript
compiler	34	gcc4
content management system	33	joomla
web server	33	httpserver
program	32	compiler
jquery plugin	32	gamequery
algorithm	31	binary-search
system	31	64bit

Witt_C/CA
Category	Size	Tag Example	Attributes
library	1104	quantlib	numerical, free, open source, finance, software
tool	889	zpt	html/xml, generation
framework	871	spring-2.5	open source, platform, application, development
system	715	sourcegear	commercial, proprietary, revision control
programming language	642	python-3.3	general purpose, imperative, high level, multi paradigm, reflective, procedural, functional, object oriented
company	555	iar	swedish, computer, software
language	486	xslt-1.0	transformation, xml
process	397	magnify
program	380	md5sum	computer
platform	332	azure-platform	durable, highly scalable, cloud based, storage
service	306	online-storage	internet, hosting
operating system	293	ubuntu-12.10	linux, debian based, server
class	272	qtablewidget	qt
plugin	261	jcarousel	jquery, open source
integrated development environment	260	phpstorm	php, commercial, cross platform
application programming interface	252	photon	realtime, multiplayer, cross platform
function	245	sha	hash, cryptographic
protocl	236	kermit	computer, transfer, transfer/management, data
engine	209	yahoo-search	web, search
web application framework	209	laravel-3	php, open source, free
method	208	accessor
interface	199	tkinter	python, standard
free software	196	mcedit
package	189	r-raster	r, data analysis
technique	184	regression	collection of, statistical
c	179	crtp
project	178	koneki	incubator, eclipse
website	174	yahoo-search
environment	174	dymola	simulation
format	172	opentype	computer font
file format	170	.doc	document
algorithm	167	sieve-of-eratosthenes	simple, ancient
extension	166	semantic-mediawiki	mediawiki, open source, free
technology	159	3g-network	third generation of, telecommunication, mobile
utility	157	diff	comparison
structure	156	cpu-cache	hardware
term	153	dynamic-html	umbrella
editor	150	wymeditor	open source, html, text, wysiwym
module	139	pyserial	python
object	138	cube	three dimensional, solid
database management system	136	sql-server-2014	relational
database	134	sql-ce	relational, compact
file	132	assemblyinfo	stardard
device	121	touchpad	pointing
product	121	winfax	software, microsoft, windows based
compiler	120	g++4.8
name	116	retina-display	brand
client	116	curl	ftp, http
way	107	idioms
element	105	noscript	html