Greetings, all,
I've just installed the latest build on the public server. This release includes an alpha release of Chapter Five of the User Manual, entitled "Software Project Telemetry: Understanding the dynamics of software development". You can find a link to this chapter on the public server help page:
<http://hackystat.ics.hawaii.edu/hackystat/controller?Page=help&Subpage=toc>
The good news is that this chapter provides a fairly detailed introduction not only to the "philosophy" of software telemetry (along with screen dumps to keep things grounded), but also to the pragmatics of defining Streams, Charts, and Reports. The even better news is that the writing of this chapter exposed a substantial number of design improvements to Telemetry, most of which Cedric has already implemented and is in place in this release. For example, the gnarly "template" angle bracket syntax is gone, replaced by a simpler parameter facility.
Chapter Five ends with a section called "A Telemetry Compendium", which is almost worth the price of admission by itself. That section shows an initial dozen or so little "programs" written in the Telemetry Definition Language along with the charts they produce. I will be adding more examples to this compendium over time and expect it to eventually contain a hundred or so examples.
Please take a look at this Chapter, and let me know if you have any comments or suggestions.
Here are some things we already know we need to work on:
- The TDL does not yet use a global namespace. Cedric will work on this.
- Cedric and I have discovered that there are many inconsistencies in the way the various Reduction Functions have been designed. To see this, go to the Preferences page, and run the Telemetry Management command using "Reduction Function" as the parameter, and check out the "Parameters" column. The good news is that the work of the past week has clarified some "patterns" for Reduction Function parameter specification. I am going to do a redesign of the parameters specs for the current set of Reduction Functions, and Cedric has volunteered to fix them.
- Some of the screen dumps are not sync'd with the latest release. They'll catch up by the time of the next stable release.
As soon as I get the Reduction Function parameter redesign done (tomorrow), I will dive into the evolutionary sensor data type stuff.
Cheers, Philip
