5. Mallados y modelos matemáticos seleccionados
5.1. Dominios y mallados utilizados
D
EVELOPMENTHave you ever worked on a project that wasn’t going according to plan? Perhaps the plan was unrealistic to begin with, or maybe the plan was dis- rupted by a series of unanticipated difficulties. How did the team present project status or task completion? Was everyone open, honest, and realistic in delivering the bad news, or were efforts made to put the most positive (and maybe unrealistic) spin on the situation? If you’re like me, you’ve prob- ably worked on projects where bad news is suppressed or downplayed, and the project is presented as being “on track.”
Historically the bearers of bad news are punished, so we naturally tend to avoid being the messenger with bad news. One good friend of mine tells of a large, high-exploration-factor software project he was part of. The grem- lins were ever present on this project: Things that could go wrong did, risks became realities, and the project was not going according to plan. After a series of weekly status meetings in which the software engineers truthfully reported the problems they were having, a senior director in engineering became frustrated and pronounced that he “no longer wanted to hear any bad news. Only good news was to be reported!” This unfortunate Dilber- tesque story reflects a behavior that occurs far too often on large projects. The suppression of bad news is bound to catch up with us sooner or later. The later bad news is discovered, the greater the negative impact and the cost of course correction.
ptg6843605 SELF-ORGANIZATION REQUIRES GLASS-HOUSE DEVELOPMENT 135
Healthy Agile data warehousing teams make extra efforts to operate in a glass house. That is, they strive to make it easy for anyone who is interested in the project to gain an accurate and honest insight into the project. Fun- damentally this glass-house nature is in the team’s DNA. Good Agile teams adopt a variety of behaviors that promote this glass-house development environment—even when it is difficult and appears to be unappreciated. Additionally, there are some specific practices that Agile data warehousing teams should incorporate.
Foremost, the Agile data warehousing team that operates in a glass house uses visual controls liberally. Visual controls include all of the butcher paper, index cards, flip charts, sticky notes, and other things that we post on our team room walls to promote collaboration. When I walk into an Agile team room, I expect to see core values and working agreements posted on the wall, the current iteration plan, the prioritized backlog of user stories, the overall release plan, a collection of known technical debt (each item on its own index card), a bug-tracking chart, a risk management chart, and any other information that should be radiated to the project community. These are primarily for use by the team during development, but anyone visiting the team room can easily review the visual controls, ask questions about what they convey, and point out areas of concern. Additionally, disciplined Agile teams use automated dashboards populated with metrics culled auto- matically from development and testing tools directly. In essence, this is BI for use by BI developers.
The glass-house environment also includes a demo of the working BI sys- tem that includes all of the features that have been developed and accepted to date. This system is always up and available for anyone who wishes to see the work that has been completed. Nothing conveys the status of a data warehousing project more accurately and honestly than the working system. The demo system is probably based on a static snapshot of operational data and is not necessarily running in production, but it is available for others to explore. This practice is supported nicely by the preproduction or dem- onstration sandbox that is discussed in Chapter 7, “Test-Driven Data Ware- house Development.”
In addition to the demo system, the glass-house environment should pro- vide access to work in progress. Features that are in development, features that are finished and awaiting review, and any experimental or exploratory work should be available frequently. Chapter 9, “Project Automation,” dis- cusses the mechanics of establishing an integration sandbox that is routinely
ptg6843605
updated to include new features and enhancements. Establishing a continu- ous integration sandbox also enables others to see that tests are passing and the build is successful.
Project wikis are a powerful mechanism for making supporting artifacts and discussions available to others. Many teams use a wiki to present design documents, data models, domain models, use-case diagrams, business analysis documentation, and other project residue.
Finally, feature showcases and executive showcases provide further visibility into project status and any issues that might disrupt the plan. Many Agile teams also produce weekly status reports that convey accomplishments, agreements, issues/risks, and other useful information.
Agile Analytics Practice: Big Visual Controls
Having team plans, progress, decisions, as well as project and itera- tion status visible on the walls will boost team performance by keeping everyone aligned and communicating with others. When visual controls become outdated or obsolete, be sure to take them down or update them.