Project History as a Group Memory: Learning From the Past
Bibliographic record
Abstract
New members of software development teams must come up-to-speed on a large amount of information before becoming productive, even if they have previous software development experience. Often, this knowledge is gained through mentoring: an experienced colleague monitors the newcomer's progress on his or her first assigned tasks, and provides feedback and advice. The mentor is the person the newcomer turns to for help when stuck; these interactions are typically informal and lightweight, such as quick questions asked over the cubicle divider or at the water cooler. However, these light-weight channels are not always available in virtual teams, where the members of the team are not collocated. Moreover, workers are less likely to help their non-collocated colleagues, making it even harder for a newcomer to come up to speed on a project. The thesis of this dissertation is based on the idea that the collection of all artifacts created in the course of development of a software system implicitly forms a group memory—a repository of information that a work group can use to benefit from its past experience to respond more effectively to the present needs. I call this implicitly-formed group memory a project memory and make three claims: (1) that newcomer software developers can use information from the project memory about past modifications completed on the project to help them effectively perform modification tasks to the system; (2) that the project memory can be built largely automatically, requiring minimal adjustments in work practices of software developers; and (3) that the automatically-built group memory can recommend artifacts useful to the current modification task. To validate the claims of this thesis, I have developed a project memory model and associated tool, called Hipikat, that recommends relevant artifacts from the memory during a software modification task. This dissertation describes the memory model, the implementation of Hipikat, and its use in a series of case studies to validate the thesis claims.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.035 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.005 | 0.010 |
| Scholarly communication | 0.013 | 0.028 |
| Open science | 0.002 | 0.012 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".