A New Platform to Battle Software Bloat?
Written By: Predrag Jakovljevic
Published On: October 7 2005
In early in May, a few weeks before Lawson announced its merger with Intentia, Lawson unveiled a new standards-based business applications platform at the Conference and User Exchange (CUE). Designed to increase overall application quality and improve the product lifecycle experience for current and future Lawson clients, the application encompasses construction, deployment, implementation, integration, support, and upgrades.
Codenamed Landmark, the new technology environment has been under development by a special team spearheaded by Richard Lawson, founder of Lawson. It has reportedly taken more than three years to mitigate the source of application complexity—the massive amounts of computer code, called "software bloat", which has brought business software innovation to a standstill. For more on this particular one and other problems plaguing current enterprise applications, see What's Wrong With Enterprise Applications, and What Are Vendors Doing About It?.
Part Two of the Can Java Perk Legacy Enterprise Resource Planning Systems? series.
The new model aims at dramatically reducing the necessary source coding and will ideally result in virtually error-free, consistent Java code. Lawson also intends to have Landmark-developed applications and Web services rolled out gradually and have them share the same data repository as clients' existing Lawson applications. This means that the existing customers would need to only perform an upgrade, not a migration or new implementation.
Based on a service-oriented architecture (SOA, for more information and background, see Understanding SOA, Web Services, BPM, BPEL, and More), Landmark is designed to enable Lawson and its clients to relatively quickly and easily modify and customize business processes to accommodate specific business or technology needs. The significant reduction in code and errors and this move to an entirely open standards technology is also believed to help considerably reduce implementation times and lower total cost of ownership (TCO) in the future. Landmark will also complement Intentia's Movex suite of Java-based, enterprise scale applications that also conform to industry standards for integration, interoperability, and SOA. Landmark supports key standards such as Web Services Description Language (WSDL), extensible markup language (XML), and Simple Object Access Protocol (SOAP).
With the new technology environment, Lawson will offer a backbone of Java 2 Enterprise Edition-compliant (J2EE) Web services for SOAs running on industry-standard platforms, such as IBM WebSphere. At the same time, it should also enable Lawson to more rapidly develop new services addressing industry or functional-specific business processes without the usual lengthy product development cycle, while enabling clients to incorporate Lawson applications and components into new composite applications. Lawson states that the first Landmark-generated applications are in development and can be expected some time in 2006.
Lawson Strategy Background
Landmark is Lawson's new technology platform that will eventually replace the current "environment" that allows Lawson applications to run across multiple databases and operating systems (OS). To put into perspective, although Lawson initially provided custom mainframe software for Burroughs and IBM installations, the company has continually recognized and anticipated new computing trends by broadening its platform support and adding products. For example, it added IBM System/38 in 1981, IBM AS/400 (now IBM iSeries) in 1988, UNIX in 1990, and Microsoft Windows NT in 1998. The company fully embraced open systems technology in 1985 by moving all development to a UNIX-based computer-aided software engineering (CASE) tool.
In 1993, Lawson was one of the first software application vendors to shift its market focus by delivering client/server applications. Lawson's cross-platform, open-database, component technology was incorporated into its flagship product, which was called the LAWSON INSIGHT II Business Management System in the 1990s, and lawson.insight in the early 2000s. Lawson customers can operate on a number of operating systems (such as UNIX, Windows NT and iSeries), database systems (such as IBM DB2, Microsoft SQL Server and Oracle) and hardware platforms (such as Hewlett Packard [HP], IBM, Sun, and Windows/Intel). Further, customers can access the applications through various end user devices, including personal computers (PC), wireless devices, and personal digital assistants (PDA).
Also Lawson has leapt eagerly onto the intranet and Internet bandwagon, building Web-based functionality into all its products before most of its competitors—including the larger and noisier ones. Its current Lawson 8 Series application suite features fully Web-based user interfaces (UI), while its historical iSeries user base now accounts for a minor part of revenues (up to 20 percent), partly because of the early provision of the upgrade path to UNIX. Lawson has long embraced layered, flexible, open, Web-based, and XML-accessible product architecture, with the idea of support forever changing technology standards. A good example was its early delivery of visionary, Web-addressable, and componentized products. Its Web-addressable features uses server-based application logic and data structure that can be referenced and executed via a universal resource locator (URL. It componentized products use Active Object Repository that exhibit an open architecture and support a wide range of platforms (using Business Component Integrator [BCI]).
The Lawson product architecture has since further evolved and is currently composed of the Web Services component (that includes Lawson Portal and Internet Object Services (IOS), which are key components for deploying Lawson applications via the Internet) and of the Application Server components that include the business logic applications, active object repository, database layer and environment (workflow, security, and BCI component integrator). For more information, see Lawson Software-IPO and Several Acquisitions After.
Lawson Current Strategy
Landmark will allow Lawson's business domain experts to specify applications in a high-level, domain-specific language (DSL), which would then generate Java program code. Lawson's approach is based on the "pattern language technology", which has been a hot topic among some software engineering communities, namely among practitioners of object-oriented (OO) design and development. Landmark could thus bring some advantages to both Lawson and its customers when the first applications based on the new development framework start to rollout next year.
Eventually, Lawson's developers will do all their programming in the Landmark development environment and the associated Lawson Pattern Language (LPL), which, when combined with the IBM's Eclipse application development framework, will generate Enterprise JavaBeans (EJB). The program code that it generates will reportedly run on any J2EE-compliant application server, such as the open source Tomcat or Jboss, or proprietary servers such as those of IBM, BEA Systems, or Oracle. Software objects and EJBs created by Landmark will be enabled for Web services, allowing them to interoperate in a SOA environment.
Consequently, Lawson customers doing customization and maintenance work will soon work with LPL instead of the various 3GL languages, such as RPG and COBOL, that are generated with the proprietary fourth-generation language (4GL) used by Lawson today. To put this into context, 4GLs are programming languages, mostly used to access databases, which are closer to human languages than typical high-level programming languages.
The other four generations of computer languages are first generation (machine language), second generation (assembly language), third generation languages (3GL, high-level programming languages, such as C, C++, Pascal, and Java), and fifth generation languages (5GL), which are mainly used for artificial intelligence (AI) and neural networks. DSLs such as LPL might also be regarded as 5GLs, given they enable developers to modify specific computer programs, such as Lawson business applications, in this case. At the same time, it works at a very high level and writes very dense code. Other examples of DSLs include RosettaNet, tools Bell Labs developed for modifying telecommunication switches, and even Microsoft Excel. Accordingly, DSLs are hailed as the future of business application development by these proponents.
The advantages of Landmark should come along similar lines for Lawson and its customers. Namely, instead of 30,000 lines of an RPG or COBOL code for a given application, the Landmark version of the code will expectedly be only 2,000 lines of LPL, because it will reuse various patterns. Lawson envisions a steady 15-to-20 times reduction in the total volume of code required under Landmark and LPL, which should equate with up to 95 percent fewer lines of code. This translates into the fact that any time one can reduce the amount of coding, one should also expect to reduce the error rate by a similar amount.
For example, instead of writing an allocations application five times, and perhaps doing it five different ways for the various components and versions of Lawson's enterprise resource planning (ERP) suite, Landmark's pattern-based development should allow Lawson to write the cost allocations logic just once, and then reuse it as needed. At runtime, when one of any number of ERP components needs to perform that particular function, it would then call that single piece of code, which should, in turn, boost the consistency of Lawson code, and enable faster upgrades, better versioning and patching, and other software life cycle benefits.
Lawson believes that this approach will allow it to focus on its domain expertise and on its "drill around" audit trails, and effective dating "signature features". For instance, the drill around feature (which is not dependent on a pre-defined path, but rather is driven by the unique transaction and the attributes that the user is investigating, regardless of the application) allows users to extend their knowledge search in a point-and-click manner at every level of the application, including reports for every data element within the database. To that end, the Active Object Repository enables any authorized user to look through and around transaction data by using the drill around feature, based on the user's needs and analysis criteria.
Further, the quality of traditionally questionable software should increase because Landmark's code generation approach will supposedly reduce the number of lines of code tremendously. Reportedly, there will be a twenty times or so reduction in lines of code for Lawson's Vendor Management module, or over forty times reduction in Lawson's Employee Management module. In terms of the customer rollout, Lawson is emphasizing a co-existence approach. All new Lawson modules will be written under Landmark from now on and will coexist and interoperate with existing Lawson modules, which will be "re-factored" in Landmark. Consequently, installed base clients can move incrementally to the written versions of Landmark.
While nobody who is familiar with the pitfalls of enterprise applications will dispute leaner, cleaner, and more consistent code, the question remains whether Lawson's current customers (some of whom have been running applications on the venerable iSeries server for twenty-five years) will be able to move to Landmark and learn the new LPL language without undergoing a major, ERP-rewrite and virtually abandoning everything they have built and their knowledge of RPG.
Incidentally, in addition to the complementary functional and geographic coverage of the Intentia and Lawson offerings, Lawson was likely attracted by Intentia's expertise in Java, and that it had created a code generation tool and a related LPL language. On the other hand, after six years of development and estimated $100 million (USD) in research and development (R&D) investments, Intentia has already ported its RPG suite to Java and created a Movex product suite that spans many platforms and is not bound tightly to the iSeries.
While Lawson has, in the past few years, renewed its commitment to the iSeries through some newer releases of its iSeries-based product, it is now moving in a direction that Intentia has already moved through. The expertise that Intentia has gained in moving the Movex suite to Java should therefore come in.
In any case, Lawson will provide its iSeries customers with tools and a metadata layer for bringing their RPG code into the Landmark integrated development environment (IDE). The move to Landmark should be regarded more like an upgrade than a full-fledged migration. RPG programmers will reportedly not have to learn Java, which does not mesh well with the traditional RPG, to work on Landmark applications. Instead they will have to learn LPL, that will be very declarative and feature an intuitive, easy-to-learn syntax. But, the programming language should not be the real issue with Landmark, since anybody who understands a company's business processes will remain valuable to an organization deploying Landmark applications, whether these may be the RPG programmers who have kept the applications up-to-date and running for nearly two decades or some other business analysts.
In terms of the EJB generation, Lawson acknowledges that Java applications typically require more hardware, including faster processors with more memory, than applications written in other popular languages. This may especially be the case with iSeries, where RPG is optimized for very efficient database access, and where a WebSphere deployment usually requires a hardware upgrade. The hope, however, is that, when Landmark is commercially available in a few years time, new technological developments will render Java more efficient. Perhaps by 2008, when computers will be running on multi-core processors with available on-chip memory stacks, and multi-threaded and 64-bit applications are put into practice, Java will have made tremendous performance gains, possibly also with the help of special Java appliances. Last but not least, maybe these performance gains will come from the expected sizeable reduction in lines of code.
At this stage, Landmark is still a mere vision, although after three years in the lab, it is certainly more than a statement of direction. The first Landmark application should be introduced in 2006 as a sourcing application for government procurement offices. At that point, the market should have a good indication whether Lawson is able to deliver. In any event, there will be no "big bang" migration to Landmark, and Lawson plans to make the transition gradually, over the next few years. On the other hand, as seen in the case of the iSeries product, Lawson is not waiting around to start writing in Landmark given it has already set the Java wheels into motion.
This concludes Part Two of a multipart note.
Part One detailed Intentia's current Java-based strategy.