Singapore cofounder trial reviews: branch tips versus common-ancestor changes

The problem

A founder reviewing a mutually agreed software trial sees a base-only addition missing from the trial snapshot and needs to decide whether the displayed comparison actually establishes a trial removal.

Choose the comparison that answers the trial question

Before reviewing a possible technical cofounder’s bounded software trial, distinguish two questions: how do the current branch snapshots differ, and what changed on the trial branch since it diverged? Those questions can have different answers. A base-only addition can appear absent in a comparison with the trial’s current snapshot even though the trial never contained or removed it. Identify the starting point before discussing that apparent deletion with the other person. Use the cofounder guide to agree the scope and working arrangements, then identify which comparison answers the question about that work.

• Agree the bounded trial and working arrangement →

Read GitHub’s two comparison definitions at their stated scope

GitHub’s Branches documentation says its pull requests show a three-dot diff: the most recent common commit, or merge base, is compared with the topic branch’s most recent version. A two-dot diff compares the latest base and topic states directly. GitHub explains that updating the base can change a two-dot diff while the topic stays unchanged. Its Comparing commits guidance also describes selecting endpoints and comparing specific commit identifiers. These are documented comparison behaviours, not an assessment of your repository. Record which view and exact versions you actually have; a branch name without its relevant version can leave the starting point unanswered.

• GitHub Docs: Branches and three-dot versus two-dot comparisons ↗ · Source checked 21 October 2026
• GitHub Docs: Comparing commits and identifying comparison endpoints ↗ · Source checked 21 October 2026

Complete a six-field comparison card

Use six entries: question to answer; exact base and trial endpoint identities; supplied common ancestor and lineage, or unknown; comparison definition and direction; content difference under that definition; and separate review, test or merge observation still needed. Keep the source of each entry visible. A screenshot can supply displayed content without supplying ancestry. If the question is the trial’s change since divergence, request the relevant common-ancestor-to-trial comparison. If the question is how two current snapshots differ, retain that direct comparison under its own meaning. Use this suggested card with permitted material; it does not assign repository access or require a real project to be displayed at an event.

Trace a fictional base-only addition and trial-only addition

Every file, version and lineage here is invented; no repository or Git command was used. Common version C contains only start.txt. The supplied history says base version P descends from C and adds base-note-L.txt, while trial version T separately descends from C and adds trial-note-K.txt. There are no other changes in this toy history. Comparing P to T shows L absent relative to P and K present relative to P. Comparing the common C to T shows only K added. For “what did this trial branch change since divergence?”, choose the C-to-T comparison. The P-to-T absence of L does not establish that the trial removed L: the stipulated trial history never contained it. The direct comparison still answers the separate current-snapshot difference question.

Hold the conclusion when the starting point is missing

A separate fictional record supplies only pictures of P and T with those same contents, but no ancestry or comparison definition. You can describe their supplied content difference; you cannot reconstruct the trial’s introduced change from that alone. Mark the common ancestor and branch-change conclusion unknown, then ask for the missing version and lineage record. GitHub also warns that diff limits or hidden files can prevent changes being displayed. Check what the actual view includes rather than treating a short displayed list as complete project coverage. A missing display, an absent item in a snapshot and a change relative to an established ancestor are different observations.

• GitHub Docs: Branches and three-dot versus two-dot comparisons ↗ · Source checked 21 October 2026

Discuss the branch change without grading the person from it

Use the comparison to ask about the agreed task, the reasoning behind a particular proposed change and the evidence still required. A branch can contain work supplied by several people. This card does not establish who authored it, whether it works, how a merge would turn out or whether the person is a suitable cofounder. Keep those questions with their own evidence and actual agreement. Use the code-review test-boundary guide to identify behavioural checks, and the approval guide to check review and integration permission separately. A useful trial conversation can clarify which change is under discussion while leaving capability, continuing responsibility and the relationship decision open.

• Check the behaviour that still needs testing →
• Separate review approval and merge permission →

Find a session that explains both views

Use Boardzip’s free GitHub search for collaboration teaching or a permitted review exercise that compares branch tips with a common-ancestor view. Confirm that the actual programme explains both comparisons, accepts your starting knowledge and offers a participant question or supplied practice task. Check preparation, accepted material, admission and charges. A founder mixer or GitHub topic match alone does not establish this explanation. You could ask, “Can this session use a wholly invented diverged-branch example to explain which comparison answers the trial-change question?” Keep private project material outside a public exercise unless its actual sharing arrangement permits it. If no suitable listing is established, retain the six-field question for an appropriate later opportunity; search supplies neither a reviewer nor a cofounder match.