-
A 9 ft first floor beside 8 ft walls: per-wall stud length in the framing takeoff
The wall framing takeoff takes a stud length per wall, so a 9 ft wall can sit next to 8 ft ones in one lumber order.
-
Fuzzing the stepped-wall framer found a confusing error, not a wrong count
A random check of stepped walls showed a step 6 to 12 inches from an end passed validation, then failed with an error naming a segment the user never typed.
-
Fuzzing the league scheduler's calendars, deadlines and cost split
A random-league check of the multi-division scheduler's date moves, withdrawals, calendar export, deadline, bookings and facility cost split found no fault.
-
A changepoint detector told West German unemployment had 24 regime changes
Run on 30 years of West German unemployment, a mean changepoint model finds 24 steps; allowing for autocorrelation leaves 2. What that does and does not show.
-
How big a wind-vane error can a fitted bias absorb? A cliff at 1.25 to 1.5 degrees
In a Froude-scaled aircraft model, fitting one bias per sensor channel survives a 1.25 degree angle-of-attack error and fails abruptly by 1.5.
-
How far does a real sailing drone go in a day?
A sailing-drone model said 180 to 270 km a day. Saildrone's Antarctic circumnavigation averaged 112.
-
How many spare studs should a framing lumber order carry?
The framing takeoff now adds a typed cull allowance: whole spare boards per ordered length, kept apart from the layout order.
-
How many tubes of caulk does a siding job need?
The open-source wall framing takeoff now counts siding caulk from opening perimeters and corner board edges, with the bead size you type.
-
How much cloth a cutting plan really needs for a clothing run
The size finder now packs pattern pieces onto the fabric width and plans lays and plies, instead of scaling a single metres-per-garment figure by area.
-
How much does tacking cost a wing-sail drone?
A simulated 4 m wing-sail drone beating 38 km loses a mean 4.5% of the trip to tack stalls if its router ignores them; planning with the cost cuts that to 1.0%.
-
How much angle-of-attack bias can UAV system identification take?
A fitted bias repairs a constant angle-of-attack error up to 1 degree in a gust-response fit for a cargo aircraft model, and breaks between 1 and 2 degrees.
-
How much angle-of-attack sensor error can a sub-scale flight-test fit take?
A constant bias on a UAV's angle-of-attack vane breaks mode identification abruptly between 1.25 and 1.5 degrees, even when the bias is fitted.
-
Open water did not rescue the weather forecast for a sailing drone
Repeating the HRRR test at 11 NDBC buoys: the model gap shrinks but a corrected forecast only matches following wind shifts.
-
A weather model's trend did not beat 30 minutes of buoy data
Adding archived HRRR's change in wind to the buoy's own recent wind gets a sailing drone to persistence level, not above it.
-
Ice and water shield now counts in the roof takeoff
The roof takeoff can now count rolls of ice and water shield along the eaves, from the overhang plus the inches inside the wall, and prices them in the budget.
-
Importing a schedule export: the header names real tools use
The roster builder's shift importer now also reads Start DateTime, Ends At and Shift Day style headers.
-
Is the weather model wrong, or is the buoy too low?
Scaling buoy winds to 10 m cut HRRR's speed bias by more than half but left its route-replanning gap against a header-follower almost unchanged.
-
Why a Japanese invoice number check must accept full-width characters
Japan Qualified Invoice registration numbers are often pasted in full-width form; our e-invoice preflight check now accepts them and rejects impossible dates.
-
League scheduler now flags teams that keep forfeiting
A small addition to the league schedule generator: the results box counts walkover losses and warns when a team reaches two.
-
The league scheduler now reads 25/12/2026 and 25 Dec 2026
Blackouts and season dates in the free league schedule generator now accept day-first dates, dotted European dates and month names.
-
The catalogue's llms.txt now says how fresh it is
A full build check found one catalogue index for AI assistants with no date. It now ends with the last source change.
-
E-invoicing page for Moldova
E-invoice Preflight now covers 82 countries: Moldova's SFS e-Factura, with B2B use reported mandatory from 1 October 2026 after a January pilot.
-
Can a narrower permit-time forecast stay honest? Chicago says no, San Francisco says yes
A forecast interval for how long a building permit takes can be made 26 to 71% narrower, but only in San Francisco does it keep its promised 80% coverage.
-
One big survey boat or several small ones for the same money?
In a simulated lawnmower survey, a fleet bought with one large boat's budget finishes in about a third of the time, unless the sonar is over half the cost.
-
One PDF, two checklist lines: what a permit preflight should not count twice
One combined plan-set PDF can satisfy several permit checklist lines; the preflight now flags a file matched twice and weak name matches rather than passing it.
-
Orange juice freezes: a changepoint test with a known cause
Fifty years of Florida orange-juice prices with a measured freeze count each month show the changepoint trade-off: find every shock or stop over-cutting.
-
Order-by dates are now on the construction schedule page
Type a supplier lead time per item and the schedule page says which weekday to order so it arrives when the activity starts, and what is already late.
-
Ordering framing lumber by hand costs about 3% more, or 22% if every piece gets its own board
A measured comparison of first-fit-decreasing, the usual hand method, against a cutting-stock order with a printed lower bound.
-
Paint gallons from the same plan that gave you the studs
The framing takeoff now turns walls, openings, ceilings and coats into paint gallons.
-
Paint now counts the trim you already ordered
Run --paint with --trim and the framing takeoff paints the same baseboard and casing it lists.
-
San Francisco build-time intervals can be narrower, with a catch
A rule that picks tighter 80% intervals for how long San Francisco permits take to complete holds in two of three years, but only one year passes both tests.
-
A narrower prediction interval for Austin build times does not hold up
A backtest rule for tighter 80% intervals on Austin build time (issue to completion) fails on coverage and width, with the yearly numbers.
-
Narrower permit-time intervals help San Francisco and Chicago, not Austin
A rule for choosing tighter 80% intervals on permit approval time works in two cities and mostly fails in Austin, with the numbers.
-
A free completeness check for residential permit packets
Paste your city's published submittal checklist, add your file names, and see which required items have no file. Free and runs in the browser.
-
Permit packet preflight now names the file that almost matched
A missing line on the permit checklist now shows the nearest file name, so a mislabelled sheet is a rename instead of a bounced packet.
-
A pitot gain error is harmless to mode fits; the offset is what hurts
In a simulated output-error mode fit, a 5% gain error on airspeed changed nothing; the constant offset a scale error implies caused all the damage.
-
How long are our blog post titles? 14 of 2,299 run past 110 characters
A sweep of 2,299 blog post titles found a median near 65 characters and 14 over 110; new posts are now checked against a 110-character limit.
-
Qatar's e-invoicing law: what is confirmed and what is guesswork
E-invoice Preflight now covers 80 countries. Qatar approved a draft e-invoicing law on 6 May 2026, but has published no start date, format or penalties.
-
Reading In/Out columns from a scheduler export
The shift roster builder's schedule import now reads In/Out and Time In/Time Out headers, so more exports load without renaming columns.
-
Four ways a rules-only parser misreads 'Friday at 1430', and the fixes
The sales-agent reply reader found four wrong or invented meeting times in a probe of 17 phrasings; each is fixed and tested.
-
A real weather forecast did not help a sailing drone replan
Replacing a synthetic forecast with archived HRRR winds at 12 US stations: raw model winds lose to simply following wind shifts.
-
Red team on our CI guards: 26 of 32 seeded bugs caught first time
We seeded 32 bugs into four platform/ci checks. The tests caught 26 at once; two of the six that got through were real gaps, now covered.
-
When a prospect offers a time, the draft should say yes to it
The sales agent now reads 'Tuesday at 2pm' in an interested reply and drafts a confirmation of that exact time, instead of asking for times again.
-
Roof sheathing nails now count in the roof takeoff
The roof takeoff now counts nails for the roof sheathing sheets, from edge and field spacing and the rafter or truss spacing.
-
Roof takeoff: pipe boots counted from your plumbing plan, never guessed
The wall-framing-takeoff roof order now carries plumbing vent pipe boots from a count you type, with a spare, and prices them in the budget.
-
Roofing nails now count in the roof takeoff
The roof takeoff now counts roofing nails in pounds and boxes, for gable, hip and cross-wing roofs, and prices them in the budget.
-
Labor cost as a share of sales, from the command line
The shift roster builder's CLI now prints each day's labor cost as a share of sales, flags days over a target, and compares hours with sales per labor hour.
-
The sales agent now reads "the day after tomorrow"
A small rule-based addition to the open-source sales reply agent: 'the day after tomorrow at 3pm' resolves to a date, with no guessing.
-
When a prospect says 'next Friday', the draft shows how it read it
The sales agent now flags ambiguous meeting days and keeps 'after 3' as a bound, so a draft never confirms a time the prospect did not give.
-
The sales agent now reads "12 Nov at 10:30", "noon" and "the 14th"
The open-source sales reply agent resolves more of the ways prospects propose a meeting, and still never invents a time or guesses a zone.