Temporal Joins and Twines
(PDF Presentation)
This is a presentation about different types of temporal joins and the twine algorithms that can be used to improve them. A recording of the presentation can be viewed on YouTube.
Are there aliens on the dark side of the moon
(PDF Presentation)
This is a different justification for data warehousing based on trust networks. It starts with a question that most have a quick and confident answer to. But why is it so? And what has this got to do with data warehousing? A recording of the presentation can be viewed on YouTube. There is also a PowerPoint version available.
Bareclad, the Pains and Pleasures of Building a Database Engine
(PDF Presentation)
The presentation held at the Knowledge Gap 2022. In the presentation a new database engine, bareclad, is introduced. Bareclad is based on transitional modeling and is an in-memory hypergraph database, supporting temporal, conflicting, and uncertain information. Within bareclad, information can be searched similarly to relational, graph, document, and name-value-pair databases. It is written in the high performance memory secure programming language Rust. A recording of the presentation can be viewed on YouTube. There is also a PowerPoint version available.
It is High Time for Bottom Up
(PDF Presentation)
The presentation held at the Knowledge Gap 2021. In the presentation a comparison is made between ensemble modeling, industry models, and upper ontologies. From there it moves on to a bottom up approach to classification found in transitional modeling. Using it, the difference between classes and types can be explained through a precise definition. A recording of the presentation can be viewed on YouTube. There is also a Keynote version available.
The Orange Apple
(PDF Slide)
This is a question that was asked on LinkedIn in preparation for the Knowledge Gap 2021. It has stirred a lot of interest and the comments are plentiful. Very few has so far had any suggestions under which circumstances one way of modeling would preferable to the other. There is also a PowerPoint version available. The answer to the question will be given at the conference during my talk “It’s high time for bottom up”.
Lost in Translation
(PDF Presentation)
The presentation held at DVEM Rising 2020. Questioning what it would be like to use a database as a lo-fidelity recording device in order to capture the ongoings in reality. Particularly what this means for information going into and coming out of databases and what we can learn from this if we view databases in a grander scheme. It ends by introducing three core concepts; context, syntax, and evaluation, which all have to be considered in order to produce the very best models. There is also a Power Point version available and a video recording that you can watch.
Modeling Closer to Reality
(PDF Presentation)
The presentation given at the Knowledge Gap 2020. A discussion on what it entails to model closer to reality and the hard problems traditional modeling techniques struggle with. It serves as an introduction to transitional modeling, substantiating the need for a new way of thinking in order to capture reality better. There is also a keynote version available and a video recording that you can watch.
A lot of power in a little structure
(PDF Presentation)
A presentation explaining the structure of posits and meta-posits, such as assertions. All exemplified through a round of professional golf, in which differences of opinion arise, multiple realities exist, and consensus is reached. A keynote version is also available.
A bite size template
(PDF Presentation)
A presentation “template” for making really short (five minute) presentations, with guidelines of how to actually make something stick in so little time. There is also a Keynote and PowerPoint version.
Modeling is Upside Down
(PDF Presentation)
A presentation highlighting the discrepancy between the real and the modeled world. The underlying question is whether we have really come that far when it comes to modeling and if we are stuck sustaining a sphere of influence with little incentive for modernization. The hand out exercise is available as a PDF: “A Model“. Both originally created for the DVEM Rising conference in Stockholm, October 2019. A PowerPoint version can also be downloaded by clicking here.
Anchor for Vaulters
(PDF Presentation)
This presentation gives a very brief introduction to Anchor by contrasting it against a Data Vault model. It also highlights many differences and similarities between the two techniques. Being familiar with Data Vault is more or less a prerequisite, and the presentation was originally created for the DVEE Summit in Stockholm, October 2019. A keynote version can be downloaded by clicking here.
Philosophy of Modeling
(PDF Presentation)
This presentation aims to introduce the philosophical grounds upon which Transitional modeling stands. It further explains the basic notions of a posit and an assertion and how Anchor modeling can be derived from these, albeit with limited but important features. It also touches upon the subject of “Schemaful Databases” which is the direction we believe new database technologies will head in. A keynote version can be downloaded by clicking here. There is also a video of the recording that you can watch.
Temporal Dimensional Modeling
(PDF Preprint) — DOI: 10.13140/RG.2.2.30079.87200/1
This is a preprint, in which the shortcomings of current Slowly Changing Dimension types are discussed, along with a suggestion to “fix” dimensional modeling using temporal dimensions. Performance tests and results are included, along with the discovery of the twine, a novel approach to achieving high performance for queries that need to retrieve historically correct relationships.
Modeling Conflicting, Unreliable, and Varying Informaion
(PDF Preprint) — DOI: 10.13140/RG.2.2.34381.49121/1
This is a preprint, formalizing Transitional Modeling, a mathematical model upon which many other modeling techniques can rely, and special cases among them are concurrent-reliance-temporal Anchor Modeling, Data Vault, and the third normal form. In transitional modeling, the information, its model, its types, and its constraints are all expressed through two simple constructions; the posit and the assertion. This makes it possible for all of them to evolve over time and to be disagreed upon or unsure about.
Turning Horrible Joins into Wonderful Speedups
(PDF Presentation)
This presentation explains some features of modern query optimizers that help turn joins into speedups. Among those features are clustered and ordered indexes to avoid sorting, table elimination to exclude tables from the execution plan, and column statistics with cost based optimizing to find the optimal join order with respect to reducing the intermediate result sets. A keynote version can be downloaded by clicking here.
Posits and Assertions
(PDF Presentation)
A presentation outlining the formalization of Anchor Modeling, showing how uni-temporal Anchor is built around posits and how these are used in assertions in concurrent-reliance-temporal Anchor. The symmetry between negative and positive assertions is presented and how it can be used to transform assertions into canonical form. The axiom of non-contradiction, including the formula for ensuring it, is also found in the presentation. A keynote version can be downloaded by clicking here.
A Bitemporal Round of Golf
(PDF Presentation)
Golf is a great game and it is also very fitting for showcasing some of the features in Anchor, particularly bitemporality and reliability, but also concurrency, since you often play it together with others. This presentation shows the development of the score, the accumulated number of strokes, which is normally posited after each played hole and then recorded on a score card. It also shows how to find the score in bitemporal effect, given a point in bitemporal time that you want to travel to. A keynote version can be downloaded by clicking here.
The Confusion Solution
(PDF Presentation)
When working with information, confusion is sometimes unavoidable. To be more precise, when the process of identification cannot give unambiguous results, such confusion arises. But don’t worry, in Anchor you are allowed to be confused and push that problem into the future, to solve it when you find the missing pieces, while still retaining analytic capabilities. Simply store all the possible outcomes in advance, with different reliabilities, or store the most likely scenario and correct it later if it was wrong. A keynote version can be downloaded by clicking here.
Big Data Normalization for Massively Parallel Processing Databases
(PDF Preprint)
This paper was accepted to the 34th International Conference on Conceptual Modeling (ER2015) and presented at the hosted fourth International Workshop on Modeling and Management of Big Data (MoBiD). The paper introduces a novel technique, highly normalized Big Data using Anchor modeling, that provides a very efficient way to store information and utilize resources, thereby providing ad-hoc querying with high performance for the fi rst time in massively parallel processing databases. A case study of how this approach is used for a Data Warehouse at Avito over two years time, with estimates for and results of real data experiments carried out in HP Vertica, are also presented. You can also download an accompanying PDF presentation by clicking here.
Anchor Modeling Introduction
(PDF Presentation)
This presentation was first held at DMZ Europe in Hamburg, 2015. Please note that the file is rather large, at 38MB, due to numerous high resolution images. It serves as an introduction to Anchor modeling, focusing on what it is, rather than how it is done. For those also interested in the how, the Back to the Moment presentation further down on the page is suitable. A keynote version can be downloaded by clicking here.
Anchor Training Day Exercise
(PDF)
An exercise that is part of the Anchor Modeling Certification Course. The exercise intends to provide an introduction to the modeling concepts in and highlight some of the specifics of Anchor Modeling. The scenario is to build and later extend a database for a new regulatory authority that will monitor all stock market transactions.
The Iceberg Exercise
(PDF)
An exercise in Anchor Modeling that involves many of the features found in concurrent-reliance-temporal modeling. The exercise is to build a database for an iceberg tracking and drift prediction system. Its goal is to track and predict the movements of icebergs in order to avoid collisions with vessels traveling on or below the oceanic surface, and allow for better route planning in the presence of icebergs.
Back to the Moment Extended
(PDF Presentation)
What if you could travel in time and become someone else? A presentation of concurrent-temporal Anchor Modeling, that was first held at the Next Generation DWH Modeling event in Amsterdam, the Netherlands and later extended for Data Modeling Zone in Hannover, Germany and Baltimore, USA. It is also the presentation held at the Data Warehousing guest lecture at Stockholm University in 2013 and 2014.
Keynote version can be downloaded by clicking here.
(updated December 9th, 2014)
Concurrent-temporal Inflation Example
(PDF Presentation)
This is an example showing how inflation is estimated every year in two scenarios, the most probable and a worst-case development. In our terminology, we have a multi-bi-temporal model with two positors, and decreasing reliability as the prognosis stretches further into the future.
AM and DV comparison chart
(PDF Comparison Sheet)
This is a comparison sheet between Anchor Modeling and Data Vault, that highlights some of the most important differences between the two techniques.
Anchor Modeling with Bitemporal Data
(PDF Presentation)
The topic of this presentation is the temporal features of Anchor Modeling, and how temporal referential and entity integrity is elegantly handled through the use of immutability.
PowerPoint version can be downloaded clicking here.
Ankarmodellering Big Data Stockholm 2012
(PDF Presentation)
The presentation held at the Big Data Event at The Museum of Modern Arts in Stockholm, April 26th, 2012. Note that this presentation is in Swedish.
Anchor Modeling Ordina
(PDF Presentation)
The presentation held at Stockholm University and later at Ordina in the Netherlands 2011.
Keynote version can be downloaded by clicking here.
Time
(PDF Presentation)
The study day presentation held at the Anchor Modeling Week (HAN) 2011.
Keynote version can be downloaded by clicking here.
Anchor Modeling – A Technique for Information under Evolution (Extended version)
(PDF Presentation)
The presentation held at the Anchor Modeling Week (HAN) 2011.
Keynote version can be downloaded by clicking here.
Anchor Modeling – A Technique for Information under Evolution
(PDF Presentation)
The presentation held at the GSE Nordics 2011 Conference.
Anchor Modeling – Agile Information Modeling in Evolving Data Environments
(PDF Preprint Journal Article)
The article published in Data & Knowledge Engineering in 2010.
From Anchor Model to XML
(PDF Article Appendix)
Appendix to the DKE Anchor Modeling Paper.
From Anchor Model to Relational Database
(PDF Article Appendix)
Appendix to the DKE Anchor Modeling Paper.
Anchor Modeling: Naming Convention
(PDF Article Appendix)
Appendix to the DKE Anchor Modeling Paper.
Performance Testing in Anchor Databases
(ZIP Archive)
Test results from various performance tests, as well as scripts for reproducing the tests.
Used in the DKE Anchor Modeling Paper.
Analysis of Normal Forms
(PDF Article)
An article on the different table types in anchor databases and their relation to 6NF.
Referenced from the DKE Anchor Modeling Paper.
Three Concepts of Time in Anchor Modeling
(PDF Article)
An article describing the three different ways time is modeled in Anchor Modeling.
Anchor Modeling ER09 Presentation
(PDF Presentation)
The presentation held at the ER’09 Conference (where we won the best paper award).
Anchor Modeling TDWI Presentation
(PDF Presentation)
The presentation held at the TDWI 2007 European Conference.