This page is going to be used to create the paper we are going submit to XPUniverse
- Deadline: April 1, 2002 Topic: history up to and including the first few days of Carlos. Plus lessons learned Length: 5-6 pages
Authors: ZhonJohansen, JeffGrover
- Our Historical Start
It all started in the early in the year nineteen hundred and nighty-nine when TheMethodologist introduced TheXpWiki to TheAgentOfChange. She said, "This is very weird, you might like it." She was right he did think it could help out in his current company (SomeBigFancyName).
When did JeffGrover (TheCoConspirator) pair with MarkWells(TheProgrammingManager) to do the big refactor? A: During most of the summer of '99. Reason: Two major pieces of server-side software did not perform well under even a light load (a few dozen clients). When complete, thousands were supported. Individually, both lacked the know-how (and courage) to accomplish this task. Wasn't test-first, no automated safety nets... but it worked out with a rapid (daily) QA cycle.
When did ZhonJohansen start writing the acceptance tests.
ZhonJohansen attended the first XpImersion offer by ObjectMentor in December of 1999 and brought back many ideas.
Some things we were doing wrong were: signing up for tasks as "a pair", ignoring the role of acceptance tests.
JoeShull (TheOpenMindedManager) allowed this GrassRootsMovement to take shape. He didn't try to curtail our informal training. He was supportive in our attmpts at testing, refactoring, standups and IterationPlanning. He even watched our standups without trying to use them as a personal agenda. Later exploits with TheClosedMindedManager would make us long for this kind of support.
- Problems XP could solve
"Customer-driven" in a bad way: the programmers were being yelled at by (TheManyTheVocalAndYetDisconnected) customers, as well as starting a new two week top priority project every two days. Top notch developers were leaving in droves because of this process. We had SustainingTeam (TheFireFighters) comprised of ten people on many of our projects (which TheXPConverts would do anything to avoid). Priority projects were completed and then never shipped because of difficulties in (TheMadMouseClickers)QA (understaffed and manual) and (TheFireChiefs) product managment (overworked and on late projects).
One of our products had just had its worst release ever. That release was almost 3 years late. It was released distainfully buggy. During most of the early adoption of XP at Symantec, we were trying to recover from this painfully public blunder.
Problems We couldn't solve as a GrassRootsMovement
- Support from our Customer
- Our Customer not being a real customer
- Customer being over-worked and not having time to write stories and answer questions.
- Managment's desire to use the standup meeting to collect status on projects
- Re-design the building layout so everyone could collaborate in a common work area
- "Enforce" pair progamming, coding standards, etc.
- Here is what we did accomplish with XP
- Persuade management to give us a small common work area
- Wrote new remote installer (never released)
Refactor of the KeyRing
- Mark and Jeff's server refactor
- Sustaining was reduced from ten to two people per iteration
- Developers were no longer being interupted durring that iterations
AcceptainceTests started to gain favor with developers and QA. Alas not the Customer
- Developers were having a better time. Sean said, "Pairing with Zhon on the acceptance tests was some of the most enjoyable time he had at Axent"
UnitTests were supporting parts of the code. These parts were very good code.
- The planning and looking back allowed each of the next three releases to complete on time.
- Finally our story ends with some pedantic moralizing
Summing or our GrassRoots attempt... Its success and failures.. The telling the the cool story of Carlos (TheIronFist?) and PSP (ThePeopleServingProcessProcess?).
And then our saga is continued by another: see AlexandraWeberMorales' article http://www.sdmagazine.com/documents/s=2279/sdm0201a/0201a.htm