Discussion Forum

Introducing TypeDB, TypeQL and Vaticle

Dear beloved Grakn Community,

As some of you may have heard during the keynote of Grakn Orbit, we’ve announced a significant change to the Grakn brand.

Keynote: Grakn 2.0: a Database Evolution | Haikal Pribadi, CEO Vaticle @ Orbit 2021 - YouTube

When we started rewriting the Grakn from the ground up last year, we knew we were going to see a dramatic evolution of the architecture. By the end of it, we were also pleased to see a dramatic evolution of the database performance. What we did not expect, though, was a renewed perspective on how we see our work.

In rewriting Grakn, the knowledge representation system was transformed into a more primary component that had an overarching impact on the entire system - it became our “type system” . Understanding the significance of our type system’s role in our codebase made us better understand the meaning of many invaluable functionalities of Grakn that was previously underrated. This helped us realise the principal components that uniquely defines Grakn as a database.

  1. Expressivity: the type system in our database enables programmers to think higher-level – at a conceptual level, or first normal form – as opposed to join tables, columns, vertices, edges and properties.

  2. Safety: the type system allows our data validation goes beyond basic value type checking. It includes logical verification of inferrable patterns, and it detects meaningless and invalid queries – resulting in data you can trust.

  3. Inference: the type system encodes the domain for logical interpretation of the reasoner. It enables type-inference and rule-inference that creates logical abstractions of data.

  4. Optimisation: the type system provides our query planner with more type information to perform thrifty/non-greedy optimisation globally using “Mathematical Optimisation” algorithms.

The fundamental problem addressed by a type system is to ensure that programs have meaning. The more expressions and restrictions imposed by the type system, the more “strongly-typed” a program is. Seeing that all four of our defining characteristics were all enabled by our type system, it became clear to us that Grakn should be defined as a “strongly-typed database” .

It’s not that Grakn is no longer an effective knowledge graph, or that Grakn is no longer inherently capable of storing graph data – both facts remain true. The realisation that Grakn is effectively a strongly-typed database does not exclude all of its previously known values. It does, however, capture a much wider scope of usefulness that transcends across many domains of software engineering.

Inevitably, we had to question whether “Grakn” was still an appropriate name for a strongly-typed database. We felt that a more general-purpose name that could capture the broad applicability of the database is probably what the technology deserves. And so we decided to rename Grakn and Graql to something that uniquely represents the technology: TypeDB and TypeQL.

The new names for the technology, unfortunately, renders the company name outdated and misaligned. So we embarked on the journey to find a name to replace “Grakn Labs” - a name we love and hold dear to our hearts for a very long time.

Pondering upon the characteristics that describe our community, we found and fell in love with the name “Vaticle” - a name that is based on the word “vatic” , which means “describing or predicting the future” .

I hope my explanation above gives clarity to our motivation behind the name change, and I hope that everyone shares our excitement on the new trajectory that this will take us. And I hope you continue to fall in love with Vaticle, TypeDB, and TypeQL.

PS: Over the coming weeks, you will see many of our internet accounts and platforms transition over to the new names.

Onwards and Upwards!