En terra solo se pueden elegir DNI y pasaporte: el catalogo tiene ademas cedula, libreta civica y libreta de enrolamiento activas, y el scope include de documentoType tiene solo 2 miembros desde el 2026-08-07. No se si es deliberado o un olvido, y el sistema no puede distinguirlo - hace falta una respuesta de la catedra, no un cambio de codigo
MEDIDO CONTRA LA BASE VIVA EL 2026-09-01, aparecio de costado midiendo TERRA-198 y NO es parte de esa fila, por eso va aparte. EL HECHO. El catalogo compartido lookupOptions tiene CINCO tipos de documento activos y sin borrar: dni, pasaporte, cedula (Cedula de Identidad), libretaCivica (Libreta Civica) y libretaEnrolamiento (Libreta de Enrolamiento). El scope de terra para la categoria documentoType tiene DOS miembros: dni y pasaporte, los dos creados 2026-08-07 18:27:09. El modo del scope es include. Entonces hoy, en terra, TRES tipos de documento que existen y estan activos NO SE PUEDEN ELEGIR. POR QUE NO LO ARREGLO NI LO LLAMO DEFECTO. Puede ser perfectamente deliberado: la catedra puede haber decidido que sus pacientes se identifican con DNI o pasaporte y nada mas, y en ese caso el scope esta haciendo exactamente su trabajo - que es justamente para lo que sirve el modo include. Un scope que excluye a proposito y uno al que se le olvido agregar una opcion SE VEN IDENTICOS desde el sistema: ausencia de fila en vocabularyScopeMembers. Es la misma clase de ambiguedad que TERRA-201 tiene con los campos de cirugia, y la resolucion es la misma: preguntar, no adivinar. LO QUE HAY QUE PREGUNTAR, y es una linea: cuando terra carga un paciente, alcanza con DNI y pasaporte, o hay pacientes que se identifican con cedula, libreta civica o libreta de enrolamiento? Si alcanza, esto se cierra como decision documentada y el sistema queda igual. Si no alcanza, se agregan miembros al scope - es una fila por tipo, sin migracion de datos y sin tocar el catalogo compartido, porque las opciones YA EXISTEN. EL COSTO DE NO PREGUNTARLO, que es lo que le da prioridad y no es alto: si hoy llega un paciente con cedula, quien carga tiene que elegir un tipo que no le corresponde o no cargarlo. Eso ensucia el dato en una direccion que despues no se distingue - un dni que en realidad era una cedula se ve como un dni. ALCANCE TECNICO SI SE APRUEBA: es una migracion de db-terra-ca, tres INSERT en vocabularyScopeMembers, y ADEMAS tiene que escribir los members, que es exactamente lo que TERRA-198 vigila. No la escribo yo. NO TOQUE NADA. Solo SELECT.
Activity
-
NO ABRO PREGUNTA EN LA COLA TODAVIA, Y LA RAZON ES DE THROUGHPUT, NO DE IMPORTANCIA. Hay CUATRO preguntas abiertas sin contestar - eq 78 (GO para terra_55, abierta hace mas de nueve horas y bloqueando a db-terra-ca), eq 81, eq 83 (quien aprueba una marca sugerida) y eq 84 (las seis cirugias sin campos). El cuello de botella no es que falten preguntas: es que las que hay no tienen respuesta. Agregar una quinta le baja la atencion a las cuatro que ya estan, y la mas vieja es la que desbloquea a otro carril. CUANDO SE ABRE: apenas se conteste eq 78, o antes si alguien esta contestando la tanda igual. La pregunta ya esta redactada entera en la descripcion de esta fila y es de una linea, asi que abrirla despues cuesta nada. QUE SE PIERDE MIENTRAS TANTO, dicho para no fingir que es gratis: cada paciente que se cargue con cedula, libreta civica o libreta de enrolamiento en este periodo se carga con un tipo de documento que no le corresponde, o no se carga. Es dato sucio en una direccion que despues no se distingue. El volumen hoy es bajo - tres practicas cargadas en total en el sistema - asi que la ventana es chica, y por eso la espera es aceptable y no indefinida.
-
coder02-terra-ca
-
Canceled as obsolete by Elazar on 2026-09-10: libreta civica is no longer part of the product, and the historical decisions are already documented. The remaining source comments describe excluded shared Enamel vocabulary; they do not expose libreta civica in Terra.
-
CLOSED WITH THE WRONG VERB ON 2026-09-10 AND CORRECTED NOW. Marker: CANCELVERB-203-mtwxa2. WHAT WAS WRONG. This row was closed with wi close, which sets status=done and writes a completed event. Its own event body, 2026-09-10T18:25:00.726Z, opens: "Canceled as obsolete by Elazar on 2026-09-10: libreta civica is no longer part of the product". So the body said canceled and the header said delivered. bwi carries a distinct canceled status and a wi cancel verb precisely so those two are not the same thing; with the wrong one, an ABANDONED row is indistinguishable from COMPLETED WORK in every status query and every done-count. Nothing about Elazar's decision changes - only the verb recording it. WHY I AM THE ONE DOING IT, stated because the row is not mine by assignment. assignedAgent is coder02-terra-ca, which returns zero rows from the hub roster read at 2026-09-11T12:21Z - and that is a point-in-time reading, not a claim about the seat forever. I hold the live coder02-terra seat. This executes a decision already recorded in the row's own body; it does not make one. The original closedAt is preserved by design: done to canceled is a closed-to-closed move and wi-http deliberately does not re-stamp it, so the date this stopped being worked is not rewritten to today. FOUND BY A SWEEP THAT WAS LOOKING FOR SOMETHING ELSE - db02-terra-cc, COERCE-SWEEP-mtwx2 lifecycle extension, 39 lifecycle events over 23 rows since 09-08. Header and log agreed on all 23; this was the one type-versus-body mismatch, and it is a different failure from the forged-transition class the sweep was aimed at. THE PART THAT IS BIGGER THAN THIS ROW: ZERO canceled events exist anywhere in terra since 2026-09-08, across 216 rows. For a board that size that is not credible as a real absence - it says the verb is unused, and that every abandoned row before this one is sitting in the done bucket too. This correction fixes the one instance whose body happened to say so out loud. A full pass would compare every done row's closing body against its verb, and I have not run it. WHAT THIS DOES NOT SETTLE, and it is a product fact, not a bookkeeping one: the question this row asked was disposed of, not answered. The catalog had cedula, libreta civica and libreta de enrolamiento active while documentoType's scope include carried only DNI and pasaporte since 2026-08-07. Elazar's ruling removes libreta civica from the product; it says nothing about the other two, and the row was closed before anyone established what the catalog holds today. If cedula or libreta de enrolamiento are still active and still unselectable, that gap outlived this row. That is a question for the db lane, and it needs a fresh catalog read rather than this row's six-day-old description.
-
WITHDRAWING THE "UNUSED VERB" CLAIM IN CANCELVERB-203-mtwxa2. The correction to this row stands; the alarm I attached to it does not. Marker: CANCELVERB-203-wd-mtwxc9. WHAT I CLAIMED: that zero canceled events since 2026-09-08 across 216 rows meant wi cancel was an unused verb, and that every abandoned row before this one was therefore sitting in the done bucket. WRONG, and measured wrong by db02-terra-cc over FULL history - 1823 event rows back to 2026-07-30, 196 completed bodies read. WHAT IS TRUE: wi cancel has NINE correct uses in terra all-time. TERRA-2 reparented, TERRA-5 "the work was never needed, not completed", TERRA-16 and TERRA-72 scope obsolete, TERRA-48, TERRA-99 and TERRA-135 duplicates, TERRA-102 filed against work already live, TERRA-128. The narrower true statement is that the PRACTICE LAPSED after TERRA-135 on 2026-08-30, and this row on 09-10 is the first abandonment since and the only one that took the wrong verb. The done bucket is otherwise CLEAN: one true hit in 196 closing bodies, and it is this row, already corrected. THE ERROR IS WORTH MORE THAN THE FINDING WAS. The 09-08 boundary was the SWEEP'S scope, not a property of the board - I took another agent's measurement window and read it as the history, then generalised a practice from an absence inside it. Before inferring that something is not done, check whether the window you were handed is the window the practice lives in. The cheap guard: a sweep's date boundary belongs in the sentence that states its result, which is exactly what my claim left out. TWO METHOD NOTES FROM THAT PASS, recorded here because the next person to run it will hit both. Abandonment language is frequently NEGATED in a closing body, because the author is arguing the row was NOT abandoned - TERRA-216 contains the literal phrase "The refile is NOT moot" and matched a keyword scan. And there is a third disposition neither verb names: a REFILE, where the obligation moved to another board rather than being delivered or dropped - TERRA-216 again, closed against ENAMEL-31. done is defensible for it, but a done-count being read as delivered work counts it wrongly too. The documentoType question raised in the previous event is NOT affected by any of this and remains open; it is being routed to the migration lane, since the agent I asked cannot reach a live catalog either.