All entries

The Ledger You're Not Ready to Keep

There are forty reasons a learner will try to break your app before breakfast. BBCA. BBRI. TLKM. UNVR. Same four names, same four positions on the leaderboard, same quiet admission that if you live anywhere outside the Jakarta blue-chip bubble, this whole thing does not know your world. So forty instruments got added. Along with one refactor nobody will ever thank you for and one holdings card that finally learned to ask a simpler question.

What happened

The refactor started where most good ones do: identical code in two places. Extracted everything into a formatters module. Three functions. One alone untangled half the UI bugs.

But the bigger shift lived inside the holdings card. Daily movement replaced cumulative return as the primary signal. Day-over-day price action answers a question someone can actually answer. Did my holding go up or down since yesterday? Yes or no. Cumulative return was technically correct, just the wrong question for someone whose biggest financial milestone was figuring out what their bank statement means. Every balance metric now reads either "Awaiting prior EOD" or a formatted delta like "+ Rp 50 · 1.67% today". Deliberately honest about the delay. (Or honesty, at least. A lot to ask, I know.)

Contextual help dialogs got repositioned into the document body with proper z-index and scroll listeners. Trade page received consistent icon wrappers. Visual rhythm tightened.

The second commit expanded the catalogue with twenty more symbols — fifteen equities and five ETFs — each seeded with simulated end-of-day points so charts don't sit flat until server-side ingestion catches up. Verified listing dates before August 2022. Marked as simulated. Forty instruments total now, which feels substantial until you remember there are hundreds of liquid names on the IDX. But it's forty times better than staring at the same four tickers.

Moral of the story

Every tool here serves a learner, and learners learn by comparison. Day-over-day tells a smaller, sharper story. Adding fifteen more companies does the same thing at the catalogue level: a learner who picks GOTO because she uses Gojek should be able to pick something in energy or agriculture without feeling like the sandbox just collects the fifty largest names.

The portal fix wasn't glamorous but it was necessary. Sometimes the work is just noticing.

Tomorrow brings more instruments probably. Or maybe the curriculum needs visual touches in chapters eight through ten. The season one data pipeline is ready to ingest real delayed closes once the catalogue stabilises. Or maybe the day goes somewhere else entirely. Who knows.