Text Input in Indic Scripts
description
Transcript of Text Input in Indic Scripts
1
Text Input in Indic Scripts
Anirudha JoshiIndustrial Design Centre, IIT Bombay
2
3
4
5
Graduates (3.7%)
Metric (8%)
Middle school (9%)
Primary school (14%)
Illiterate
(35%)
Tele-density ShareUrban 148% 62%Rural 40% 38%
(Jan 2013, TRAI)
HSC / Diploma (4%)
Illiterate
(26%)
6
Sour
ce: T
he T
imes
of I
ndia
7
How many people in India speak English?
8
How many people in India prefer to speak in English?
9
Sour
ce: T
op 1
0 Pu
blica
tions
201
2 Q1
, Han
sa R
esea
rch,
MRU
S, IR
S
Dainik Jagran (Hin); 17%
Dainik Bhaskar (Hin); 15%
Hindustan (Hin); 12%
Malayala Manorama (Mal); 10%
Amar Ujala (Hin); 9%
The Times Of India (Eng); 8%
Lokmat (Mar); 8%
Daily Thanthi (Tam); 8%
Rajasthan Pa-trika (Hin); 7%
Mathrubhumi (Mal); 7%
1010
11
Sour
ce: A
n In
stal
latio
n in
IDC
Structure of the Devanagari Script
क ख ग घ ङ च छ ज झ ञ ट ठ ड ढ ण त थ द ध न प फ ब भ म
◌ ा� िा ा� ा� ा ा! ा" ा# ा$ ा% ा&
य र ल व श ष स ह ळ क्ष ज्ञ
क ख ग घ ङ
च छ ज झ ञ
GutturalsPalatalsLingualsDentalsLabials
Vowel modifiers
Semi-vowels
अ आ इ ई उ ऊ ए ऐ ओ औ अ% अ&
Vowels
ट ठ ड ढ ण त थ द ध न
प फ ब भ म
Cons
onan
tsVo
wels
12
Challenges in Text Input in Indian Languages
13
क ख ग घ ङ च छ ज झ ञ ट ठ ड ढ ण त थ द ध न प फ ब भ म
◌ ा� िा ा� ा� ा ा! ा" ा# ा$ ा% ा&
य र ल व श ष स ह ळ क्ष ज्ञ
GutturalsPalatalsLingualsDentalsLabials
Vowel modifiers
Semi-vowels
अ आ इ ई उ ऊ ए ऐ ओ औ अ% अ&
VowelsCo
nson
ants
Vowe
ls Large number of
characters– ~660 frequent glyphs
2+ keystrokes = glyph क + ा# = क# (C + V)
क + ा# + ा% = कों (C + V + V) Difference in pronunciation
and visual sequence क + िा = किा
Conjuncts (halant ा=between two consonants)ट + ा= + व = ट=व (C + C)
Varying similarity between conjuncts and consonantsस + ा= + त = स्तक + ा= + र = क्रर + ा= + क = क@क + ा= + ष = क्ष
Text Input in Indic Languages
QWERTY keyboard for Devanagari input–Devanagari needs 52 keys (13 vowels,
34 consonants, 4 conjuncts, 1 halant)–QWERTY has 26 un-shifted keys
Leads to cognitive load on users–Complex structure of Indic scripts (G >
K)–Large number of glyphs
Need much training and practice–30-50 hours to reach 25 wpm–Slow starts–Only professional typists put in the
effort QWERTY is not suitable14
15
16
17
Sour
ce: A
nshu
man
Kum
ar
Task Success /5 Users
18
Without helpp (one tailed)
Nokia (0.208)Samsung (0.195)Sony (0.001)
Totalp (one tailed)
Nokia (0.000)Samsung (0.010)Sony (0.000)
19
Disha
20
Swarachakra
21
Inscript
22
23
Aug 2013
Jun 2013
Oct 2013
Jan 2014