Welcome to Certabo forum

Both posts in italian and in english are accepted on the forum to help everybody to provide contribution irrespective to geographical location. Object of the forum is to provide technical information on various items and help developer and users to personalise software for Certabo chessboards. User explicitly accepts the conditions of use of this forum here reported.

A brief introduction of user is always appreciated in section Welcome to….

Benvenuti nel forum di Certabo

Sul forum possono essere postati articoli sia in lingua italiana che inglese per permettere a tutti di dare il proprio contributo da diversi luoghi geografici. Il forum ha come scopo fornire informazioni tecniche in generale e aiutare gli sviluppatori ed utenti a personalizzare software per le scacchiere Certabo. L’utente accetta esplicitamente le condizioni del presente forum consultabili qui.

Un breve saluto del nuovo utente è sempre gradito nella sezione Benvenuti a….

Please or Register to create posts and topics.

calibration file

I did think about the calibration file these days, and my first idea would be to move from pickle to json. That would make it human readable and independent from specific pickle versions (e.g. lower pickle versions can't read files from later ones). On startup, an existing pickle file could be converted to a json automatically. So it would be transparent for users.

My second thought was if it wouldn't make sense to have a single calibration file for all instances. As far as I understood, multiple calibration files were added to support multiple boards being attached to a single computer at the same time. But in the end there shouldn't be any collisions with the 40bit RFID chip IDs. So a "global" calibration file would allow to learn in all sets once, and then use it with any board, irregardless on what port it is being attached to. We'd need to lock it somehow during calibration, so that we don't end up throwing away data when two processes run calibration at the same time.

Quote from hari on February 28, 2020, 3:35 pm

I did think about the calibration file these days, and my first idea would be to move from pickle to json. That would make it human readable and independent from specific pickle versions (e.g. lower pickle versions can't read files from later ones). On startup, an existing pickle file could be converted to a json automatically. So it would be transparent for users.

My second thought was if it wouldn't make sense to have a single calibration file for all instances. As far as I understood, multiple calibration files were added to support multiple boards being attached to a single computer at the same time. But in the end there shouldn't be any collisions with the 40bit RFID chip IDs. So a "global" calibration file would allow to learn in all sets once, and then use it with any board, irregardless on what port it is being attached to. We'd need to lock it somehow during calibration, so that we don't end up throwing away data when two processes run calibration at the same time.

Hello Hari

yes your suggestion is good we use already .json easy to edit to pass parameters to engines, so can be good improvement to have even more flexibility.

As for the second point yes that is the reason as during tournament several boards are normally attached to same pc via usb hub we found at least for the end user, easier to run calibration on different set. But the suggestion would be good if for example we create single calibration file which can moved from a device to another easily also  if i get correctly your point. Well you can add one set per time or two at same time using the 3456 rows for the additional set

All the best

Pietro