ta/vg151_funny_shit: h1 grading

1 files changed, 30 insertions, 0 deletions

diff --git a/docs/ta/vg151_funny_shit.md b/docs/ta/vg151_funny_shit.md
index 6c17530..f4ba878 100644
--- a/docs/ta/vg151_funny_shit.md
+++ b/docs/ta/vg151_funny_shit.md
@@ -132,3 +132,33 @@ Murphy's law worked. Apart from this VS atrocity, students were pushing:
 - pdf
 - symlinks to the pdf
 - a CC BY 4.0 International license
+
+## 2022-10-01 to 2022-10-02, Canvas (and everywhere else)
+
+Right past midnight, as the calendar flips to 2022-10-01, I began grading
+all freshmen's homework 1 with a Python script inherited from past TAs.
+The script got a few hiccups, because:
+
+- I misconfigured the # number of OJ test cases
+- OJ server constantly crashes
+- Students' first name on Gitea is capitalized but not on Canvas
+- Off-by-one bug produces non-existent group, hgroup-00
+
+It took me more than an hour to fix them. At 01:47, I finally released
+freshmen's first homework grade. It was disastrous. By tradition we gave
+them one day to fix mistakes they've made, and the next midnight I re-ran
+the script. Signs of improvement are showing, but there still remains four
+groups who got a full deduction (-2.5).
+
+Naturally, my inbox was flooded with complaints and questions (around 25),
+especially from these four groups. When I woke up, I decided to
+investigate.
+
+- Group A: Named their README file `README .md`
+- Group B: Named their README file `README(ex1).md`
+- Group C: Pushed to master but forgot to release
+- Group D: Not enough teammates to approve PR
+
+Considering A and B made understandable freshman mistakes, C had
+a technical problem as explained by team member, and D actually did
+contact us beforehand, I negotiated a partial refund with Manuel.