Post #4394009
2026-08-05 13:01 UTC
Replies (6)
-
@Damselfly@blackqueer.life 2026-08-05 16:58
@gvwilson@mastodon.social I got through question 17, but found many of the questions written to vague to answer with assurance and the first 17 took me 30 minutes to describe/qualify my answers. It’s impossible for me to complete, unless we are focusing solely on web-based using agile/scrum methodology. Some questions even include Scrum in the question itself. But that ignores a lot of software development, and as someone who developed and managed teams building safety-critical systems under DO-178 with a regulated waterfall process and mixed sensor arrays and build environments, the survey itself excludes me. Like one question is specifically about Infrastructure as Code, but those frameworks don’t even exist in many of the systems I worked on. Maybe I’m just too autistic and/or not the software demographic you want. I mean here is my answer to question #1: You have not clearly defined the terms. Are we thinking about group performance across peopl or individual performance across time? Another relationship? Do we have normalization for experience? Code understanding? Etc? We have a quantitative value of improvement (15%), but a qualitative metric (“better”). Without defining better quantitatively, this cannot be measured. Is better equal to faster? “faster to write?” “Faster to run?” “Faster to compile?” “Faster to troubleshoot?” “Faster to refactor?” “Faster to retrieve a real-time signal from an aircraft transponder and decode it, calculating the travel time and aircraft position, and returning a guidance signal via antenna?” Or is better equal to more lines of code? Or fewer lines of code? Or better abstracted? Or more modular? The measurement and the metric are too unclear to answer. Perhaps a single junior developer, working more efficiently, can only improve by some arbitrary metric by 15%, while a senior developer experienced with the codebase and compiler could improve that metric by 100%. I feel like we need to think more closely about the real core of the question we are trying to answer and reframe this with more focus.
-
@Lee_Holmes@infosec.exchange 2026-08-05 20:28
@gvwilson@mastodon.social Answered - hope it helps some folks. There are quite a few "it depends" answers in this. If the goal is to compare responses with what is known as being empirically correct, most of these do not have an empirical answer.
-
@oisin@mastodon.ie 2026-08-05 13:30
@gvwilson@mastodon.social Done. There are some odd ones in there!
-
@bert_hubert@eupolicy.social 2026-08-05 13:41
@gvwilson@mastodon.social done!
-
@MarcParadise@hachyderm.io 2026-08-05 13:47
@gvwilson@mastodon.social I found a lot of the questions difficult to answer as posed, because it presents a lot of things as binary that are not in spite range of agree/disagree provided. it would be great to see the final results/analysis your students arrive at
-
@rhempel@cosocial.ca 2026-08-05 13:48
@gvwilson@mastodon.social Done - some of those questions are ... shall we say open to interpretation. This is what I worry about most with AI - developing software is a social activity. Every one of those questions is a great team or water cooler discussion topic. When developers are seen as isolated managers of faceless agents, how are conversations taking place about nuances in software development?