Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


On the Perceived Interdependence and Information Sharing Inhibitions of Enterprise Software Engineers

Research Paper Presentation for 2012 ACM Conference on Computer Supported Cooperative Work

Alicia M. Grubb

on 4 July 2017

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of On the Perceived Interdependence and Information Sharing Inhibitions of Enterprise Software Engineers

Department of Computer Science, University of Toronto amgrubb@cs.toronto.edu
Supervised by Steve Easterbrook

Microsoft Research

2012 ACM Conference on
Computer Supported Cooperative Work
Monday, February 13, 2012 Surveyed 3000 engineers at Microsoft in summer 2010.
Total Sample 898 (33% response rate).

Analyzed questions of survey that related to interdependence and information sharing.
Survey questions derived from literature and interviews.
Sampled Project Managers (PM), Developers (Dev), and Testers (Test)
iDepend – the respondent perceives that he depends on work done by others outside his team.

oDepend – the respondent perceives that individuals from outside his team depend on him for his work. Senders vary how widely they share information based on artifact type. How widely send share information varies by role.

Testers less comfortable sharing than Developers and PMs. Information Sharing - Altruism
How would preceived dependencies affect how widely the share information?

Expected Results: If lots of people depend on you, then you want to share more informaiton.

Actual Results: ...
Those who say they depend on the work of people outside their team (iDepend) share to a wider audience than those who say they do not have outside dependents. Respondents were apprehensive about sharing even when asked.
33.2% of respondents reported having been asked for work-related information that they were not allowed to share.
25.1% of respondents reported being asked for information that they did not feel comfortable sharing.

5 reasons not to share:
the sensitivity of the request artifact
uncertainty as to whether the information should be shared
the information was too preliminary
mistrust of the requestor (general)
requester would not understand artifact Newsfeeds, like source code checkin emails and automated bug edit notifications, undermine the senders' abilities to control the distribution of their information.

Senders need help to send the right message to the right recipients.

A semi-automated approach could:
help identify the audience (and make it obvious that there is one)
help tailor the message to each audience Enterprise networking is not social networking
Engineers within enterprise settings are statistically different from one another.
PM have more dependencies than Dev and Test
Testers were shy - but it made sense
Altruistic sharing of information more widely if you depended on others is not yet understood
Automated communication tools should enhance sender's control over message distribution. On the Perceived Interdependence and Information Sharing Inhibitions of Enterprise Software Engineers Are coordination patterns uniform? If not, why?

What impact does non-uniformity have on successful communication? Allowed Comfortable Non-Sensitive Sensitive F (2, 788) = 7.43, p = 0.001 F (2, 790) = 8.07, p = 0.000 F (2, 785) = 6.65, p = 0.002 B D C E A Coordination Dependency Networks
code dependencies and people dependencies
actual coordination and communication patterns Conway 1968, Herbsleb and Grinter 1999,
de Souza et. al. 2004, Cataldo et. al. 2006 Information Sharing - Role Information Sharing - Altruism Motivation Methodology Allowed Comfortable Non-Sensitive Sensitive t(769) = 2.14, p < 0.05 t(770) = 2.25, p < 0.05 t(765) = 2.23, p < 0.05 t(802) = 2.87, p < 0.01 Sharing Concerns Lessons Learned Possible Implications for Senders Artifacts (and their means). Increased values indicate wider sharing. Information Sharing Andrew Begel Alicia M. Grubb Work Group Range Work Dependency G I F H
Nonsensitive information is shared more widely than sensitive information. Information Sharing - Sensitive Features Allowed Comfortable Non-Sensitive vs. Sensitive t(739) = 16.73, p < 0.0001 t(750) = 17.94, p < 0.0001 Most Widely Shared Most Narrowly Shared PM Dev Release Schedule 4.26
Specifications 4.24
Feature Schedule 4.18 Test Release Schedule 4.24
Specifications 4.24
Binaries 4.18 Release Schedule 3.79
Bug Fixes 3.77
Bug Reports 3.69 Source Code 2.77
Checkin Msg. 2.81
Test Suites 2.95 Meeting Notes 2.71
Test Suites 2.98
Source Code 3.00 Meeting Notes 2.53
Source Code 2.69
Checkin Msg. 2.71 Perception of Dependency varies by role, and years of experience.

Developers b = -1.09
Tester b = -1.30
Manager b = +0.45
Years-Exp. b = +0.08

Developers b = -1.01
Tester b = -1.32
Manager b = +0.49
Years-Exp. b = +0.08 Introduction
Coordination in Software Enginering Scenario:
Fred (application team) asked Bob (developer on the library team) for the ability to change the color of Bob's login dialog box
Bob said sure!
But a week later, Bob's manager Sara told him to drop it, because it was of low priority to 'her' library team
Bob forgot to tell Fred
So, Fred was suprised when the new version of the library was missing his feature
Fred had to delay his ship date, costing his team big money and wasted time We think better communication will lead to better software. z = -3.83, p < 0.000 (PM Ref.)
z = -4.70, p < 0.000 (PM Ref.)
z = 1.73, p < 0.083
z = 3.09, p < 0.002

z = -3.52, p < 0.000 (PM Ref.)
z = -4.75, p < 0.000 (PM Ref.)
z = 1.85, p < 0.064
z = 3.07, p < 0.002
Full transcript