re: nhbill
She has a system database that has among other things, a few dozen tables where she remembers data about data structures and user interfaces. When she syncs, she reads the system tables, primary keys, foreign keys, etc and “remembers” them in her system database. From this, she is able to create web applications on the fly that can have search, forms, child forms, etc. An administrator can then use the app and modify any of her metadata…like changing labels, widget characteristics, security, etc. Within an hour or two, one can change all the settings to put the polish on all the “virtual” pages in the “virtual” app on top of a typical DB of around 10-20 tables. One can apply icons, create relations in the data where they do not exist explicitly in the db, etc. There are about 50 meta properties on each entity and around 65 for each field. You can do search, add, edit, view, copy, delete, and reorder operations on anything with zero coding…and my favorite part, deal with lots of one-to-one and many-to-many relationships, in collapsible/expandable sections with a grid for each relation.
I needed all this because I have so many memory types to maintain. Now that I have it, it makes for a really easy way to build web business apps on top of dbs and apis…so I’m getting a bit distracted from robotics with that.
There is also a memory-driven rules engine so one can create complex rules in the U.I., like enabled a field based on other fields, and complex database db update rules…once again, without coding. Fun stuff. I spent my career building metadata driven frameworks for fortune 500…this is the latest in a long line of ancestors, I am getting a lot of interest in licensing/consulting opps with it.
Everything is modifiable in the app, while using the app at runtime…no compile step…just a constantly living, usable app that gets better with human input. When an underlying table changes, or new tables are added, Ava adapts to the new data, adding it into whatever pages are relevant.