Pranjal Sailwal

recovering from log

  1. seq 0000recovering from log

Play

A drivable index of the work. Every figure you collect is a real claim from the ledger.

Receipts collected 0 of 9

The ground starts when you scroll to it.

The nine landmarks

This is what driving into each ring opens, as plain text. Tap a stamp to see how the figure is known.

Data and systems

Engines, sync and guards: the work where the question is what stays true when things go wrong.

  1. NitroDB

    What actually keeps a database's data safe through a crash?

    68unit and integration tests
    60,000writes in a recorded run of the real engine, drawn as its own SSTable files on each level, from the first flush to after compaction
    42concurrency outcomes checked, 0 failed

    Open the NitroDB caseNot collected yet

  2. SyncState

    Can people keep writing together when the network drops?

    226characters in the text two Yjs replicas merged to after 40 recorded steps, drawn as two threads that part and then braid. Yjs is the library SyncState builds on
    YjsCRDT library behind the merging (integrated, not written from scratch)

    Open the SyncState caseNot collected yet

  3. Appwrite Migration Guard

    Did the migration really move everything, and can it say why not?

    11 of 11planted changes caught in a live pass, no false positives
    8commands, all working against live Appwrite Cloud

    Open the Appwrite Migration Guard caseNot collected yet

  4. SecureHost Control Suite

    Can one place decide what reaches the network, the camera and the mic, and keep a record?

    4components: core library, service, GUI, CLI
    3default policy rules
    DPAPIpolicy storage encrypted with Windows data protection

    Open the SecureHost Control Suite caseNot collected yet

Product and growth

Funnels, receipts and a city of repositories: the work where the question is what the numbers really say.

  1. Ledger vault

    Can you check every figure on this site yourself?

    48claims sealed in a hash chain: each seal is the SHA-256 of the one before it and the claim itself, and your browser re-checks the whole chain

    Open the receiptsNot collected yet

  2. Growth Intelligence Lab

    Why did the cost per conversion jump four times over?

    98.1%of the cost increase happened after the click
    19.2xmore spend in the larger campaign
    44tests

    Open the Growth Intelligence Lab caseNot collected yet

  3. Repository city

    What has been built in public, and how much of it is real code?

    27public repositories drawn as towers, taller for more code, coloured by language, with a bead on those pushed in the month before the latest push

    Open the workNot collected yet

Research

Sensors and light: the work where the question is how much to trust a measurement.

  1. AetherScan

    How much can you trust a single air-quality reading?

    11data sources, each with licence, rate limit and how provenance is kept
    366days of Dehradun air quality from a file in my own files (PM2.5 from 5.4 to 155.3), drawn as a skyline that is tall in winter and low in the monsoon. The file states no source
    12backend test files, each named test_*.py

    Open the AetherScan caseNot collected yet

  2. Spectral Monte Carlo Path Tracer

    Is this render right, or does it only look right?

    1.81Msamples a second, checked against an analytical answer
    BVHbounding-volume hierarchy for ray traversal

    Open the Spectral Monte Carlo Path Tracer caseNot collected yet

Found a number that doesn't add up? Write to me.