Synon Founder Breathes New Life Into Lazy Software
August 26, 2013 Alex Woodie
Simon Williams, who created the Synon/2E and Obsydian 4GL development tools that were so popular during the AS/400’s heyday, is back at the helm of Lazy Software, the company he founded in 1998 to develop and market a revolutionary associative database. Williams rejoined Lazy Software in 2010 after the company and product foundered about for several years, and now he is shipping a new release of the database, Sentences 4.0, and gearing up to tackle big data problems.
Lazy Software got off to a big start in 1998, the same year that Synon sold itself to Sterling Software, which in turn would be acquired by Computer Associates two years later. The Synon/2E and Obsydian products today live on–and apparently are actually doing pretty well–under CA as CA 2e and CA Plex.
Williams’ big ideas around the associative data model are deceptively simple, and yet complex. The short explanation is that the associative data model holds key advantages over the traditional relational model that is predominant in today’s enterprise systems.
Associative Data Model
For example, consider that a typical enterprise software package requires tens of thousands of relational database tables, which are required to deliver quick application-level access to all possible permutations of customer names, numbers, products, product numbers, etc. It’s the developer’s responsibility to keep all the tables and their relationships straight, and that complexity makes it difficult, time-consuming, and expensive to build and maintain applications on top of all those tables.
By contrast, the associative model–as explained by Lazy’s Associative Model of Data, which you can read at www.associativemodel.com–eliminates the need for the programmer to know anything about the structure of the database. Under the associative model, all data is stored in a single, massive associative database table that has just four columns–entities, associations, items, and links. And this is sufficient to describe any and all kinds of data and relationships, the company says.
The associative approach not only dramatically simplifies the structure of the database, but enables the developer to work at a higher level of abstraction and to develop reusable programs that can operate on every type of data without modification, Lazy Software says. What’s more it eliminates the need to hard code rules into programs. The data and the metadata exist side by side, and the data model is self-described, self-evident.
Sentences in Practice
Williams and company put this associative model into practice with Sentences, a suite of database management tools built around an associative database engine that was first launched in 2000. The software was designed to essentially build an associative database “on the fly” as business logic is coded above it, in Java, RPG, COBOL, C++, or other languages.
In 2001, Lazy Software attracted $9 million in outside investment, and shipped Sentences version 2.0. In 2002, the company shipped Sentences version 3.0, which was adopted by several big organizations, such as Argos, Alcan, Electrolux, Forgecom, Johnson Matthey, Lloyds TSB, Man Group, Prime Selection, and Reuters. See Lazy’s website for case studies.
Then the economy went into recession, and in 2003 Sentences was effectively mothballed. For the next seven years, there was no development on the product. But then Williams detected a possible opening in 2010, “when the market’s new-found interest in all things data-related prompted to me to re-start development,” Williams tells IT Jungle via email.
That was about the time that new and exotic types of databases started to make big news. QlikTech, for example, uses an in-memory associative database with its QlikView business intelligence product, which has been successful and propelled the company to a big IPO. IBM also resurrected interest in the associative database model with Cognos Express, which is built off the Applix TM1 product. Then there are the NoSQL databases, such as MongoDB; the columnar databases, such as Quiterian (now part of Actuate and the BIRT Analytics group), and others.
With the launch of Sentences 4.0, Lazy Software is taking a slightly different tack with its associative database, and is now touting the product as “a web-enabled, multi-user data discovery, federation, and visualization environment for complex databases.”
To that end, the product with version 4.0 has been “comprehensively retooled for its new role in data discovery and federation, with an emphasis on its ability to retrieve, correlate, and present data from multiple data sources simultaneously.”
With Sentences 4.0, users will be able to pull data from multiple relational databases–say DB2/400, SQL Server, and Oracle–and store it in an associative structure, where it can be queried by users. That will have real-world applicability in organizations that struggle to integrate and analyze data in pursuit of that all-important “360 degree view” of things.
“Sentences automatically infers an associative schema from a relational schema,” Williams says via email. “It can extract and present data from multiple RDBMSs and other data sources simultaneously, automatically infer equivalence relationships between instances of the same real-world thing in different databases, and aggregate their attributes.”
Lazy Software announced version 4 a year ago, and is actively selling it as the company refines the data integration capabilities. The company has several prospects, Williams says, including with a large international bank and a UK government department. “No successes to talk about yet but enthusiastic reactions from our prospects. They like our ability to integrate multiple complex relational and other data sources and to query the integrated result,” he says.
Version 4.0 brings several other new features, including a new security model; several new business rules engines; enhancements to its queries engine; a new “sandbox” mode to facilitate experimenting; a new transaction visibility mode; and a server process scheduler.
Big Data Future
Sentences is not a big data machine quite yet, although Lazy Software does have a roadmap to get to big data. The limiting factor in Sentences’ ability to be a big data engine is the fact that it executes a single process on a single node.
“We need to get to multiple processes on multiple nodes,” Williams says. “For big data we need to develop a slave server process that can run on multiple nodes simultaneously.” Eventually, it will be able to support large sets of transactional data with ACID concurrency, he says.
With that said, Williams supports the notion that today’s “big data” isn’t really any different than other data types, and that differences are mostly semantic. “Soon all data will potentially be big data,” he says. Sentences is already being used with all types of data, from relational data and spreadsheets, to documents, emails, and social media posts, Williams says.
Sentences is available in a multi-user Enterprise Edition for commercial use, and a single-user Personal Edition for non-commercial use. The software runs in the Java environment, and requires a Java-based Web server. Lazy Software says it supports its software on Windows, Linux, and Unix, but if you look hard enough in the user manual, you’ll find instructions for running it on the IBM i platform.