This project has a severe case of creeping featuritis; the time has come to face this disease and make some decisions. A creep is a feature that was not initially scheduled into the project, but seems so natural, obvious, or desirable that we just can't resist slipping it in. All project schedules assume a number of petty creeps, but this can easily get out of control, and this appears to be happening with the Erasmotron. We need to evaluate these creeps and either slip the schedule or kill the creeps.
Here are some of the features that have been creeping into the project, features that I had not anticipated when I first scheduled it:
These first four have already been implemented
1. combinable facial elements (you can assemble a face out of components)
2. colorable facial elements (the ability to assign colors to wigs, etc)
3. stackable facial elements (you can layer clothing elements)
4. the screentest
These have not yet been implemented:
5. the background display, which helps storybuilders organize their thoughts.
6. more realistic facial hair, requiring a new drawing algorithm
7. multiple modes of editing. Currently the default editing mode is "replacement", with a limited capability for "outsertion". I have determined to make outsertion universal and add a third editing mode, "substitution", which allows the storybuilder to substitute one term in place of another without disturbing internal parameters.
8. The recapitulation feature of Le Morte D'Arthur, which allows the player to examine the story so far.
9. The advisory function in Le Morte D'Arthur, implemented with an avatar of the storybuilder, including an image inserted by the storybuilder. This is a biggie.
10. Extensible facial image files. (We can add more facial images after we've shipped the product without screwing everybody up.)
11. A proper on-line help system.
12. Additions to the stage data structure, including background assignments and abnormality indicators.
13. Embellishments in the facial expression editor: more features, higher resolution, better use of intelligent constraints such as forcing the eyelid to move as a unit.
14. A new system for "Expository Substories". This would replace the "Teach Me About and Tell Me About features in Le Morte D'Arthur. The storybuilder would define a group of expository substories that are meant to pop up randomly to help explain the context and backstory for the storyworld. The storytelling engine would construct a "confusion parameter" based on the median response time to menu options and the number of "DoNothing" options chosen; the greater this confusion parameter, the more often it presents expository substories.
15. List of who is on stage. This allows the player to know who's watching him.
These are only the creeps that I can recall right now; so many have popped up over the last few months that I have given up trying to keep track of them. I had thought that I could stay on top of all of them, but Laura has had so many good ideas that I haven't been as ruthless as I should be. Perhaps I just can't say no to a mother in distress.
Creeping featuritis is not necessarily a fatal disease; with proper care and behavioral adjustments, some projects have been known to go on indefinitely with creeping featuritis. The trick is to harmonize schedule objectives and feature objectives. I offer three options for consideration:
There is no question as to my ability to achieve the goals I set out to achieve and promised in June; all it takes is a reaffirmation that I should do so and the consequent willingness to make hard choices. This option prevents us from implementing most of the creeps in the above list.
Push the Macintosh/demo storyworld ship date back from April 1 to July 1. Maintain PC ship date of October 1. Hire PC programmer earlier (perhaps January) and interleave the PC programming with the Macintosh programming. This is a trickier project management task, one that I would prefer to avoid, but it makes up the schedule costs of the additional Macintosh programming. We would pay for the additional programmer-months with the excess money arising from Dave's reduced participation. The additional three months time would be used for earlier marketing efforts on my part, implementation of most of the creeps on the above list, and for Laura to do a better job with her demo storyworld.
This requires a complete renegotiation of budget and schedule. We should pursue this option only if Markle feels enthusiastic about the creep list and is eager for even more of the same. An example of a budgetary adjustment would come from the artwork. We have a goodly set of facial artwork; it has cost us about $10K so far. We could provide more artwork if we decided that it is worthwhile.
I am excited about the possibilities on the creep list and would like to implement them, but I am also embarassed by my continuing susceptibility to creeping featuritis. I prefer Option #2 but am uncomfortable with its implications for my competence as a project manager. Therefore, I can immediately proceed with Option #1, with no prejudice whatsoever; after all, it's what I promised to do.