-
Lenght and AkousticRating: what a property set check found in 64 published IDS files
The IDS checker now audits Pset_ and Qto_ names and data types against buildingSMART's templates. In 64 public IDS files it found faults in 21.
-
Open an IDS, change it, save it: 64 public files come back asking the same thing
The free IDS editor now opens existing IDS files. All 64 public files we tried, and 187 buildingSMART test cases, write back asking for the same thing.
-
Our ball machine weighs each shot by how much its wheels slow, and the lowest revolution hid up to 38% of it
A two-wheel launcher's speed drop measures each ball's momentum. Fitting it to the tach's pulse times halved the speed fit's error in simulation.
-
The textbook fix for a jittery tachometer made our ball machine's spin estimate worse
Tach jitter sits on edge times, so fitting their running sums should beat fitting periods. On our simulator it read the spin worse. Numbers and likely cause.
-
One badly fitted wheel out of 56 told our ball machine its motor controller had no lag
The tennis feeder's ESC-lag survey picked 0 on one build because one wheel fit 30 times worse than the rest. Trimming such wheels fixed every simulated case.
-
Forty microns of wear would have switched off our ball lift's jam detector
One stiff disc spring on our tennis ball lift's slip hub lost 61% of its torque after 40 microns of facing wear, and jams went undetected. Four discs fix it.
-
Our Rust CSS checker agrees with PurgeCSS on every selector
We ran PurgeCSS 9.1.0 and csshound on the same 3,923 pages: identical results, 14x faster on four threads, and one bug in how both read a page.
-
A 2003 Canon raw file, read sensel for sensel without converting it first
ShootFlow now reads Canon CRW files from 15 bodies of 2000-2004 directly. On 20 real files every sensor value matched LibRaw 0.22.
-
Finding closed eyes on real photos: a face-landmark model against the old Haar cascades
On 105 real CC0 photos, a MediaPipe landmark detector flagged 5 of 6 closed-eye shots and no false ones. The Haar cascades flagged 3 of 6, with 8 false ones.
-
The demosaic weight should follow colour correlation, we said. Real photographs disagreed
A pre-registered test of a colour-correlation feature for ShootFlow's learned demosaic weight. It missed its bar, and real photographs ran the other way.
-
Closed eyes on a face cut in half by the frame: mirror the edge
A face cut by the frame edge is invisible to MediaPipe's face finder. Mirroring the edge strip finds 20 of 40 half faces, 19 read the same as the whole face.
-
The learned Bayer demosaic in C: bit-identical to numpy, and why it is only 2x faster
ShootFlow's learned demosaic weight now runs in C, equal to numpy bit for bit on seven real raw frames. It is 2-3x faster, not 4-5x: the features cost the time.
-
Teaching a demosaic weight about noise worked a little; teaching it fine detail backfired
ShootFlow's learned demosaic weight, refit on noisy mosaics, gains a little under noise. Synthetic fine detail made it worse, for a reason about colour.
-
Printing a painting with the inks already on the shelf
Picking a printer's stock inks by eye cost 1.5 to 3.4 delta E on nine paintings. A subset search won most of it back, and from 3 m beat custom inks.
-
Reading a receipt's tax that OCR could not see, from the total
On Indonesian receipt photos the tax line is often a blur. With the total and a few subtotal digits, 10% PB1 arithmetic gives it back, on two of 21 misses.
-
What a receipt prints under the total, and why it broke our item split
Summaries and tip tables printed under the total kept a bookkeeper from splitting receipts into items. One guarded retry fixed 116 held-out ones.
-
Receipt rules written on US photos, tested on Indonesian ones
Seven commits of receipt-OCR rules, all written from 200 US photos, scored on 200 Indonesian receipt photos none of them had seen. Most did nothing there.
-
Rio's Nota Carioca service codes: 998 codes, five acts, one check
Rio de Janeiro's NFS-e takes a six-digit service code from a table changed by four acts since 2018. We read them all and now check the code offline.
-
Same answer, 38% faster: Sinkhorn in plain JavaScript, four rows at a time
The Wasserstein distance tool spent most of its time in two matrix-vector products. Reading the kernel four rows at a time cut 38% without changing a bit.
-
San Francisco and Emeryville pay schedule changes by the hour, not by the hours lost
Shift Roster Builder's team hub now prices schedule changes under San Francisco's and Emeryville's rules, read from the cities' own documents.
-
Seven calls, one question: a 26-second analysis that only needed one fit
A several-columns analysis in the change-point detector ran the same 3.6-second stale-feed fit seven times. A 16-entry memo cut it to 4.6 s, same output.
-
SHA-512 without BigInt: the Facturae signature check got seven times faster
Our in-browser Facturae XAdES check hashed with BigInt SHA-512. Two 32-bit halves per word took a check from about 12 ms to 1.7 ms, same bytes.
-
Two seconds for the repair: a pre-registered test turns a hunch into a default
On 350-400 shift weeks, giving Shift Roster Builder's exact repair 2 s of the budget covered more hours on 22 of 32 fresh builds that differed.
-
Bring your Square schedule across: a reader built from Square's own API definition
Shift Roster Builder now reads Square's scheduled shifts, jobs and locations, from saved JSON on the page or fetched with your own token on the CLI.
-
Every time reserve at once, from two builds a week
Shift Roster Builder's pay bound now gets 6 s past the budget instead of 1.5 s: it finishes 19 of 19 tries on 48 fresh weeks for 0.25 s more on average.
-
New York City counts schedule notice from Monday, not from the shift
Shift Roster Builder's team hub now prices New York City's fast food schedule-change premiums: flat $10 to $75 per change, counted to the week's first day.
-
No formula sizes the warm solve: a time gate replaced by a reserve
A warm-started LP in Shift Roster Builder took 29 ms to 14.5 s on similar weeks, so its size-based time gate went; a fixed reserve proves two more weeks.
-
Reading Oregon's scheduling law itself: two pricing mistakes and a standby list
Checked against ORS 653.455, Shift Roster Builder priced small start changes and short extra shifts wrong in Oregon; it now has the voluntary standby list too.
-
One close-then-open, three prices: Seattle, Los Angeles and Philadelphia
The same 8-hour rest between a close and an open costs $20, $80 or a flat $40 depending on the city. The team hub now prices all three.
-
The same call-out cover costs an hour's pay in Chicago and nothing in Seattle
Shift Roster Builder's team hub now sends a written offer to the whole team after a call-out and prices on-call shifts under Seattle, Oregon and Chicago rules.
-
The proof that came after the deadline: one in five big weeks skipped it
A step skipped once the 20 s budget was spent turned out to hit 1 in 5 big weeks. Two pre-registered tests later, it runs for 4 s more.
-
Less time for the repair, more for the walks: a pre-registered tie
Trimming Shift Roster Builder's repair reserve from 2 s to the 1.2 s it actually uses tied on 48 fresh builds: 7 better, 8 worse, 33 the same.
-
A time reserve with nothing to do, and the second and a half it gave back
Testing whether Shift Roster Builder's repair needs reserved time on 30-60 person weeks found it had nothing to fix, and was spending 1.3 s a build anyway.
-
The walk that ate the pay floor's time: a stage timeline for a 20-second budget
A stage timeline showed one random walk ran Shift Roster Builder's big weeks past their 20 s budget. A clock check inside it cut late builds from 15 to 5 of 48.
-
Twice the reserve, worse rosters: a pre-registered test that said no
Doubling Shift Roster Builder's repair reserve to 4 s left more hours open on 27 of 38 fresh builds that differed, because the repair stops on nodes, not time.
-
Sick leave balances from the punches a small team already makes
Shift Roster Builder's self-hosted hub now earns paid sick leave or PTO per hour worked, at the manager's own rate and caps, and keeps it past the punch trim.
-
Slovakia's 2027 e-invoice mandate, read from the Act instead of vendor blogs
We read Slovakia's Act 385/2025 on slov-lex.sk. 1 January 2027 is a transitional rule; the general e-invoice rule waits until 2030. A Slovak check ships.
-
Spring reaches Central Park two weeks earlier than in the 1880s: a dynamic time warping tool
A new browser DTW tool aligns two series and says where one runs ahead. On 150 years of Central Park highs, April's normal now arrives 14 days earlier.
-
Sunt read as sut: the n a synthetic chant choir sang and its own checker missed
Chant Forge's choir sang the n in sunt, but its acoustic checker heard only the t. One cue fixed it: a murmur starts with a step, not a ramp.
-
Five microseconds of sensor jitter would triple our tennis feeder's spin error
Our feeder reads spin from wheel-speed drops. Simulated, 5 us of tach jitter takes it from 1.6% to 5.3% off, so the board now flags fits it cannot vouch for.
-
One tennis-feeder wheel read 48% high because the Pi spoke too soon
A 1,107 rpm launcher wheel read 1.48x its true drop. The board's snapshot ended three revolutions after the ball. Holding the next speeds fixed it.
-
Did it change on the day you named? A known-date test that found the seat-belt law and still did not ship
A segmented-regression test found the 1983 seat-belt law and passed its control series. It failed its size bar on short sticky series, so it stays out.
-
Testing an X-Trans demosaic on scenes it never saw: borrowed Bayer frames as truth
Only two X-Trans raws are public, so we mosaicked six other cameras' scenes with Fujifilm's layout. We led LibRaw's three passes under noise on all seven.
-
The BCG trials agree on average and disagree everywhere else
A free meta-analysis tool checked against statsmodels and metafor. On 13 BCG trials the pooled risk ratio is 0.49, but the prediction interval runs past 1.
-
The check that threw out the n of nunc: a sung consonant reader's ninth round
Chant Forge's consonant reader missed the n of nunc and mens on C4. A rule had found it; a second check threw it out. Fixed, then tested on a new held-out set.
-
The Facturae signature policy hash belongs to a PDF revision from 2008
Every Facturae signer writes the same policy hash, and the PDF served today does not match it. The 2008 revision inside the file does.
-
The largest number on a receipt is often the cash you handed over
When OCR loses a receipt's TOTAL line, the biggest amount left is often the cash paid. AI Bookkeeper now gives that guess no vote unless the sums back it.
-
The last untimed stage: a 40 ms pay polish and a pre-registered test
Shift Roster Builder's pay polish ran 25-44 ms past the 20 s budget on 8 of 48 builds. A deadline inside it took that to 0, with pay and open hours held.
-
The same season, searched twice
One test checks a guide page against real football fixture lists. A quarter of its 18 seconds was the scheduler repeating searches it had already done.
-
The repair that ran past its deadline: checking the clock inside the search
Shift Roster Builder's cover repair read the clock only between rounds. Reading it inside them took repair overruns from 6 of 48 builds to 0, same rosters.