exploring the role of instant messaging in a global software development project
TRANSCRIPT
EXPLORING THE ROLE OF INSTANT MESSAGING IN A GLOBAL SOFTWARE DEVELOPMENT PROJECT Yvonne Dittrich [email protected] Rosalba Giuffrida [email protected]
IT#UNIVER
SITY#OF#CO
PENHA
GEN
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Problem Ethnographic Empirical Research WB project Role of Instant Messaging Findings Future Work
Outline
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Problem
Fundamental problem of GSD is that many of the mechanisms that function to coordinate the work in a co-located setting are absent or disrupted in a distributed project.
J. D. Herbsleb.
IT#UNIVER
SITY#OF#CO
PENHA
GEN
How Informal communication
shared through Instant Messaging complements
formal communication usually shared through structured tools and
documentation?
Research Question
Y. Dittrich, K. Rönkkö, Jeanette Eriksson, Christina Hansson, and Olle Lindeberg. Cooperative method development. Empirical Software Engineering, 13(3):231–260, December 2007.
Cooperative Method Development (CMD)
Evaluate introduced methods
Propose methods & guidelines
Empirical investigation
CMD Phase 1 CMD Phase 2
CMD Phase 3
Evaluate introduced methods
Propose methods & guidelines
Empirical investigation
CMD Phase 1 CMD Phase 2
CMD Phase 3
Cooperative Method Development (CMD)
Y. Dittrich, K. Rönkkö, Jeanette Eriksson, Christina Hansson, and Olle Lindeberg. Cooperative method development. Empirical Software Engineering, 13(3):231–260, December 2007.
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Ethnographic Empirical Research
! Data Collection: Semi-structured interviews, Document Analysis, Log Analysis, Workshops, Participant observation
! Data Analysis: Interaction analysis of IM logs, Categorization of chat episodes, Member checking
5 domain experts/software developers in Denmark 6 software developers + 1 tester in India 1 team member in Portland
Customers in Ethiopia
WB Project for water management in Nile Basin
Portland
Copenhagen
New Delhi
Addis Abeba
IT#UNIVER
SITY#OF#CO
PENHA
GEN
A typical work day
Time India Denmark
9 IST / 4:30 CET Start work
12:30 IST / 8 CET Ask questions Start working from home
13 IST / 8:30 CET lunch Go to office
14:30 IST / 10 CET Meeting
15:30 IST / 11 CET Intense collaboration
16:30 IST / 12 CET Lunch
17 IST / 12:30 CET Intense collaboration
18 IST / 13:30 CET Finish work
21:30 IST /17 CET Finish work
IT#UNIVER
SITY#OF#CO
PENHA
GEN
A typical work day
Time India Denmark
9 IST / 4:30 CET Start work
12:30 IST / 8 CET Ask questions Start working from home
13 IST / 8:30 CET lunch Go to office
14:30 IST / 10 CET Meeting
15:30 IST / 11 CET Intense collaboration
16:30 IST / 12 CET Lunch
17 IST / 12:30 CET Intense collaboration
18 IST / 13:30 CET Finish work
21:30 IST /17 CET Finish work
Communication Channels
Communication Channels
Communication Channels
Communication Channels
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Dimensions of the use of IM
Coordination Collaboration
Awareness Socialization
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Coordination Dimension
Many coordination chats start from the Indian side, e.g. to inform the task/incident changed status, and is assigned to the counterpart. [14:07] IND: 974 is yours now for verification (gap filler scrollbars) [14:09] DK: thanks [11:58] IND: 1008 [12:07] DK: all yours [12:57] IND: your now. [15:58] DK: 847 is yours
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Collaboration Dimension
! Technical discussions ! Easy-to-resolve issues that do not require much
discussion. ! Last for no more than 10 minutes; on average,
they last 7 minutes. ! Thereafter the topic switches to a call. ! If issues have a wider implication, the
discussion is moved to mail.
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Awareness Dimension
! Indian developers inform what they are working on in order to avoid conflict.
! Danish team members check what people are working on in order to plan.
! This information is provided by issue tracker, but often the official tool does not show what people are actually working on at the moment.
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Socialization Dimension
• Often present as a sub-text in chats mainly addressing another dimension.
• Some chats are purely addressing the social dimension taking up e.g. commuting problems or personal topics.
IND: Hi DK DK: hi IND IND: hows monday treating you? DK: better than usual. Working from home IND: oh we r on the same boat DK: good to hear :)
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Ecology of channels IM as a glue articulating other channels The Social Dimension is Important
Findings
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Ecology of Channels
! Analysis needs to focus on ! the ecology of channels. (in line with Turner et al, 2010 ! Social protocol and communication breakdowns.
(Damian et al. 2007) ! Communication and coordination mechanisms (Schmidt
Simone1996)
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Instant Messaging as Glue
! Unstructured ! Less obtrusive ! Awareness ! Nearly synchronous communication
Instant Messaging
IT#UNIVER
SITY#OF#CO
PENHA
GEN
The Social Dimension is Important
! Explicit social chats and sub-text in work related chats
! The team decided not to use the notification function of the issue tracker system to allow for more social interaction.
! Earlier research proposed video and virtual worlds to improve the social dimension of communication. IM seems to take that role today.
IT#UNIVER
SITY#OF#CO
PENHA
GEN
Future work
! Comparing how different teams combine heterogeneous artifacts and communication channels in GSD.
! Exploring the use of other kinds of Social Software, such as microblogging, social networks and content sharing sites.
! The current concepts from CSCW and GSD are not enough to understand the ecology of channels we have observed.
! Using this knowledge to improve communication and collaboration in GSD.
EXPLORING THE ROLE OF INSTANT MESSAGING IN A GLOBAL SOFTWARE DEVELOPMENT PROJECT Yvonne Dittrich [email protected] Rosalba Giuffrida [email protected]
IT#UNIVER
SITY#OF#CO
PENHA
GEN
THANK YOU :)
Questions?