Proposal: Social Networks in xP (SNiP) enterprise management process
Principal investigators: KatyaRogers, ZhonJohansen
Abstract
We model developer to developer, manager to developer, developer to manager, and manager to manager interactions using a paradigm of a canine social network, and borrow an idea of sensor-enabled collar communication devices to enhance team communications and decrease tedious managing tasks in a large XP group.
Introduction
Managing an XP group presents new challenges. While smaller groups can be self-organizing (JonathanHouse), due in large part to evolutionary social development of smaller-sized communities (Charity Begins...), larger groups present new challenges for a manager. Tracking individual achievements for annual reviews is obscured by pairing process, complex integration environments lead to frequently broken builds etc.
With the widespread appearance of social network infrastructures (Friendster, cellphone based networks) and maturing of related technologies, it becomes increasingly easy and affordable to adopt the ideas in other fields. We focus on a canine social network (SNIF) as a model for managing XP groups since interactions in a canine network have multiple similarities to an XP group dynamic: managers, like dogs owners, often have difficulties understanding language of their charges; programmers, just like dogs, are unlikely to manually track their time pairing (playing); and, finally, snacks (treats) are essential for productive work in both groups.
System Description
In SNIF networks dogs are equipped with sensor collars, and leashes containing data collectors. Leashes are docked to a station for centralized information transfer. In SNIP we propose to use unobtrusive bracelets - that can serve as fashion statements supporting XP process (to differentiate from a multitude of wrist bands that came in vogue recently we propose bracelets in light green color of JUnit Green Bar complete with "Team Extreme" logo). We use keyboards as docking stations and equip them with data gathering sensors (e.g., Bluetooth).
In this environment we are able to gather information about programmers' preferences for pairing, allowing manager to optimize the team by excluding programmers that are avoided by most; as well programmers' tendency to "drive" versus "navigate". The latter information could be used to promote "thinkers" to architect positions based on how little time they spend doing actual typing.
A related shock collar technology could be incorporated and used as a feedback device to communicate back the information about broken builds (Cruise Control Is Evil), decrease in code coverage, or not following "test-first" principle. This is a great improvement over current state of development, where feedback on the problems mentioned above is normally achieved only upon submitting the code to repository, and relies heavily on vigilance of fellow team members. These bracelets could be identified by their red color (of JUnit Red Bar) and "Extreme Shock" logo.
Security
Security should be given maximum consideration in the proposed system given the potential ramifications of violating access restrictions to data that could be used for promotions, or access to and inappropriate use of "Extreme Shock" bands by unauthorized personnel.
System Implementation
Although the technology is mainly of interest to large, enterprise-size groups, and thus have a potential as a commercial application, we propose to use an Open Source model of development in order to minimize security concerns.
Future extensions
If the initial response to the SNIP infrastructure is good, we plan to extend it further by adding blood pressure and heart rate monitor to the sensor bracelets. Acquiring this data would help us to detect an unfortunate situation of developing non-work-related feelings when pair programming with a member of opposite sex (XP2E Chapter?). We suggest, where facilities allow, connecting SNiP network "pairing alarm" with a sprinkler system fire relay. In this maximum efficiency configuration the sprinklers are used to "extinguish" inappropriate team dynamics. If workplace is not equipped for sprinkler use, preventive separation of affected team members, though much less efficient, could be substituted.
Conclusions
We showed that managing an XP group can be further simplified though fashionable sensor equipment and an automated data gathering network.
Acknowledgments
Authors are indebted to Ray Cox for helpful discussion, and, in particular, the idea of further automating managing process through integration with a sprinkler system. We also thank our many friends for proof-reading early versions of this proposal and supplying encouraging laughter. We hope their sense of humor is far better than their grammar skills.