Converging on OOAD agreement

The Unified Modeling Language (UML) has generated considerable enthusiasm within the object-oriented software development industry. Based on the object-oriented analysis and design (OOAD) work of three of the world's most well-known methodologists -- Grady Booch, Ivar Jacobson and Jim Rumbaugh -- the UML is a combination of notation, semantics/syntax and a meta-model for a modeling language. The UML is, however, only a piece of the overall puzzle for activities in the OOAD world.

More and more interest has recently been raised in the area of OOAD for several reasons, including: the UML and the visibility of the methodologists involved; an OOAD standardization/convergence effort through the Object Management Group (OMG), Framingham, Mass.; an increased interest in business object modeling; and advancements in OO Computer Aided Software Engineering (CASE) tool support. In this update on the state of OOAD, we will review the recent history of OOAD, in particular the rise and dominance of Rational Software Corp., Santa Clara, Calif.; today's industry efforts; the future of OOAD as it looks at this time, especially in conjunction with other efforts like Shlaer-Mellor -- a methodology apparently less interested in converging with others; and parallel development trends in the OOAD software development arena such as the rise of business objects and OO CASE.

Rational convergence efforts

The OOAD community has seen many changes in the last two years. From a methodological perspective, one company has received tremendous visibility during these years for its move to bring together those many consider to be the leading methodologists in the industry. Just two years ago, Rational Software had a single chief scientist, Grady Booch. During the 10 or so years that Booch had then been with Rational, he had helped introduce object technology to the masses through a number of publications and a presentation style that made it both fun and educational to look at object technology.

In what many considered a bold move, Rational then hired Jim Rumbaugh. Rumbaugh, along with others, created and helped popularize the Object Modeling Technique (OMT), the most used and supported methodology around. Although Booch and Rumbaugh had collaborated and influenced each other's work, Rational's hiring Rumbaugh nonetheless stunned the OOAD community. Interestingly, even after Rumbaugh joined the company, he continued to write articles showing suggested extensions to OMT, which many then referred to informally as the "Rumbaugh" technique. Many wondered how Booch and Rumbaugh would work together and what direction the industry would take. Most assumed that a convergence of their work would take place.

A few months later, at a prominent object technology conference, Booch and Rumbaugh announced that they were working on a Unified Method that would be the evolutionary next step in methodologies and would most certainly be the "upgrade path" for users of both the Booch and Rumbaugh (or OMT) techniques. They invited other methodologists to participate in defining the Unified Method, which was to be a graphical notation and a process to help users practice proper software engineering techniques. Using a Rational supported mailing list, they created the Object Technology Users Group (OTUG) to track comments and requests, share ideas and get feedback from the industry on the Unified Method. They pledged that the Unified Method would be open to public scrutiny and used to converge an industry that appeared deep in religious fervor over methodologies.

Rational soon made another bold move -- they purchased Ivar Jacobson's Objectory -- which combined Booch, Rumbaugh and Jacobson all under one organizational roof. Once again, the industry wondered how the group would work together given the individual visibility each had within the OOAD community. Jokingly, the group came to be called the "three amigos." Rational now commanded an even greater visibility in the convergence of the OOAD methodologies effort.

The next OOAD convergence move by Rational, expected by the end of the third quarter of 1997, will be through three separate books, one from each of the three amigos. Booch, in the OTUG forum noted, "Jim, Ivar and I are working on three books: a reference manual, a user guide and a process book for the UML."

OMG convergence efforts

Around the same time that the OTUG was formed, activity began within the Object Management Group to determine whether a convergence of the OOAD industry was possible. A special interest group, called the OOAD SIG, was formed at an OMG sponsored meeting in San Jose, Calif., after many of the leading methodologists and tool and training vendors began to discuss what could be done to converge the industry. That first meeting showed the industry that not only were Booch and Rumbaugh interested in convergence, but also that the vast majority involved in methodologies saw benefit in formally working together. The group quickly elected Mary Loomis, Hewlett-Packard's Software Technical Labs director, as chairperson and James Odell, a consultant and educator, as well as co-author of the Martin-Odell methodology, as co-chair. By the end of the morning, the group was formally in place with an initial charter and on its way to determining what level of convergence they could attain.

Common Data Interchange Format

The Common Data Interchange Format (CDIF) standard would allow vendors to share information contained in their repositories with other tools, as shown above.

Fig. 1

Source: Viktor Ohnjec

During the OMG meeting, it was pointed out that some OOAD tool vendors had already tried to work together, specifically through a Common Data Interchange Format (CDIF), which had been suggested as an EIA (Electronic Industries Association) standard several years prior. The CDIF standard was to allow vendors to share information contained in their repositories with other tools, both to supplement the capabilities of any given tool in a more easily integrated format and to give users the ability to use different modeling tools if they so desired. Although not universally accepted, CDIF was in fact supported by certain tool vendors, and some suggested that reinventing the wheel in this area would be counter productive. Said Rational's Director of Object Technology, Jon Hopkins, "The OMG originally thought that a standard OOAD meta-model would be a good thing to standardize. That evolved into the need for a standard meta-object facility (in other words, a meta-meta model), a standard interchange format and a standard OOAD object model." (See Fig. 1.)

The other question was whether or not to attempt to create a single methodology as the OMG standard. Such a move would have put tool vendors and training organizations in the position where their features and services would be the main added value offered to clients, rather than their support for any one specific methodology. This goal was considered to be too challenging initially. Instead, the group took on the goal of driving a common meta-model for any methodology. A meta-model would allow the underlying definition of object technology concepts to be fundamentally the same, regardless of which specific notation was chosen, and would then still support a convergence in the industry while allowing the notation and process elements of a methodology to be considered later.

The task at hand

The three amigos, shortly after Jacobson's arrival, had narrowed the scope of the Unified Method by focusing on notation rather than process and renaming it the Unified Modeling Language or UML for short. "Standardizing on a notation and a meta-model first and then tackling process after that to, in essence, create a method was considered a better approach," said Rational's Hopkins. The UML would be used as part of Rational's official submission in response to a request for proposal (RFP) to what was now the OMG OOAD Task Force. This proposal would suggest a modeling language with a meta-model, a notation, syntax and semantics, and a standard exchange format, but would formally defer any process-related direction until some time in the future. The UML submission itself was also supported by other industry heavyweights like Microsoft; Oracle Corp., Redwood Shores, Calif.; Texas Instruments, Plano, Texas; HP; Unysis Corp., Blue Bell, Pa.; MCI-Systemhouse, Washington, D.C.; Intellicorp. Mountain View, Calif.;and Icon Computing, Austin, Texas. On the importance of UML to TI, Keith Short, director of Advanced Technology at Texas Instruments Software said, "UML notation will be at the heart of Texas Instruments Software's component and object modeling methods and product lines."

The OMG OOAD task force RFP has generated other responses in the industry as well. Erick Rivas, VP of Engineering at Oakbrook Terrace, Ill.-based Platinum Technology Inc.'s Clear Lake Labs located in Houston, led one particular submission based on the works of numerous methodologists, including Brian Henderson-Sellers, professor of computer science in object technology at Swinburne University of Technology, Victoria, Australia, Ian Graham, Don Firesmith and about a dozen others. Derek Coleman of Kings College, London, recently added his Fusion methodology to this effort. Fusion itself is a converged methodology that integrates the ideas of a number of methods to form a second generation object-oriented methodology. Once again, Fusion appears to be leading by example, as supporters of a new Team Fusion effort move to incorporate UML as the Fusion notation while retaining an evolved Fusion process.

The Platinum-led combined effort is formally named the Open Object Advisory Group (previously the Open Object Alliance) and has its own mailing list (called OPEN). This submission is much more process and document rich than the notation-centric UML, but is equally, if not more, complete in its meta-model aspects. The submission is currently referred to as the Open Modeling Language (OML). Brian Henderson-Sellers said, "OPEN is both an acronym (standing for object-oriented process, environment and notation), as well as an invitation and opportunity to participate in a truly open, worldwide development of a rapidly emerging OO method which can be loosely termed 'third generation.'" A book on OPEN is expected by February 1997 and should be available around the time this article is published.

Additional OOAD submissions to the OMG OOAD RFP came from IBM and ObjecTime, Kanata, Ontaria, Canada and another lead by Taskon A/S, Oslo, Norway. Each of the submissions presented at the January 13-17, 1997, OMG meeting in Tampa, Fla., included some reference to UML as a supported modeling language either by displaying models that use UML notation or by referencing UML in some other way. Chris Stone, president and chief executive of the OMG finds the whole RFP response interesting. "We didn't ask for the UML, we asked for interoperation on design models."

Where will convergence lead?

As with most OMG-related standardization efforts, lobbying by all submitting groups is expected. When a standard is in place, it will undoubtedly be some combination of the different submissions. "I expect the outcome to be a merger of what Platinum (the OML camp), what Rational (the UML camp) and what IBM are doing," said Rivas of Platinum. At worst, the effort would result in two accepted standards, most likely UML and OML, rather than the tens of methodologies available today. (See Fig. 4.)

If the result is a single standard, the current industry speculation is that the UML notation, syntax and semantics, and some of the UML meta-model will become part of the emerging standard. Similarly, it is expected that the OML process-related areas of management, testing and documentation, as well as most of its meta-model, would also end up integrated into any single emerging standard. "I expect some tweaking and tuning to the UML and some additional work in the CDIF area, but I don't expect many changes in the meta-model area," stated a confident Hopkins at Rational.

Speculation also is mounting that Rational may choose to incorporate the Objectory process into its UML while simultaneously retaining the process as a proprietary one, as it is now, and making that process available only to Rational supported customers. If this speculation holds true, the only public-domain, UML-based methodology would then be Derek Coleman's Fusion.

"Rational would in a parallel path submit a process in a separate submission to the UML effort," said Rational's Hopkins. "This makes the problem much more tractable." Rational may internally plan to converge its own process and the Objectory process, with the resulting process to be provided to clients in terms of intellectual property and tools and to keep the process proprietary. Booch, himself on the OTUG forum has stated, "Rational has a number of people working on process-related issues, carrying on the evolution of the process work that Ivar, Jim and I have done, as well as incorporating the work of others at Rational such as Philippe Kruchten and Walker Royce, and a number of the process people from the original Objectory."

On the same OTUG forum Jacobson added, "Grady Booch, Jim Rumbaugh and I have agreed to work on a generic process for software and business engineering ... we are writing a book on this subject." Jacobson does not believe that there is a need or a desire to standardize on process in the same way as for a modeling language. Time will tell, however, and who knows, perhaps Rational will make a home for Derek Coleman and the Fusion process as well.

The translation approach

The translation approach to modeling involves heavy analysis performed upfront and allows a design to be automatically generated.

Fig. 2

Source: Project Technologies

The elaborative approach

The elaborative approach to modeling consists of involved modeling notations to make for a seamless design.

Fig. 3

Source: Project Technologies

OMG standardization

Some experts speculate the efforts to standardize a notation language would result in two accepted standards, UML and OML.

Fig. 4

Source:Viktor Ohnjec

In a written report on how the OMG would reply to the submission process, Liz Barnett of the Giga Information Group Inc., Cambridge, Mass., suggested that the OMG would "push back the proposals to the two groups and ask for a compromise solution." Barnett expects two approaches to be blessed and that users would not see any benefits for years, while tool vendors change their products. The OMG, said Barnett, "has little standing or credibility in the OOAD world. Its endorsements will not impact user buying decisions."

Indeed, on the tool front of today's convergence efforts it is interesting to note that Rational's Rose is among the products to support the notation only later in the process, and that the company's training offerings remain focused on Booch or OMT rather than the newer UML. Other vendor's products, including Platinum's Paradigm Plus, Littleton, Colo.-based Advanced Software Technology Inc.'s Graphical Designer and several more, on the other hand, show their ability to quickly track changes in the UML and offer UML support shortly after a revision is posted. Training vendors show their offerings, although most tend to take Booch-, OMT- or Fusion-based courses and modify them with UML notation, syntax and semantics, while retaining whatever development process the original course suggests. The approach is valid, they argue, since it has worked in the past and since no official stance by the UML authors on process has been immediately available.

Translation versus elaboration

Some within the industry choose not to participate in the OMG effort because they consider their own efforts to be more advanced. "Imagine that the airplane manufacturing industry was trying to standardize on turboprop engines and I suddenly gave you a jet engine. How easy would it be to standardize if the two approaches were so different?" asked Steven Mellor of Project Technologies, Berkeley, Calif., and co-author of the Shlaer-Mellor methodology. "We believe that a simple notation is important, but in the seamless world of UML, that is not possible because of the elaborative approach they use." Mellor's plans include continuing to watch the OMG efforts, as he remains interested in the areas that can be standardized so that he can simply adopt those areas as they stabilize. He suggested that Shlaer-Mellor will not be "arbitrarily different just for the fun of it, but we see no way in the near term to reconcile between each other." In the interim, Mellor said he had no intention of slowing down either the UML or OMG efforts, but would instead continue to work in his own area to improve. "We see the future focusing on a translation-based approach to constructing designs," he said.

Shlaer-Mellor takes a translation or "zippered" approach to modeling, whereby heavy analysis is performed up front and once it is fairly complete, a design can be automatically generated. (See Fig. 2.) Its users produce their code by building an application-independent definition of the system design that they refer to as a software architecture and then building a set of rules that define precisely how to embed the application into the architectural components. This approach differs from the elaborative (or seamless) approach that UML and most other methodologies use, Mellor suggested. (See Fig. 3.) "You can use a simulator or a verifier to execute your models and verify that they have the right behavior. You can get more predictable and manageable project schedules because there is a defined process, and you can split your subject matter in a way that allows tighter iterative loops because you don't have to go as deep into the analysis/design loop as you do with other approaches," Mellor said of his approach. He also believes that the UML is only a modeling language and has no defined process at present.

In regard to standardization, Mellor suggests that after the OMG effort is complete, the industry will look for the next step to take and that such an effort will lead toward a Shlaer-Mellor supported approach. "We don't expect to see any unification of these two camps in the near future because there are very different technical drivers. Elaboration methods want a rich modeling notation to make the transition from analysis to design smooth, easy and seamless. Translation methods want simple notations to ease the job of defining the translator," said Mellor.

Rational's Hopkins suggests that the industry will ultimately decide, and that at present, both in terms of mind share and in the fact that Shlaer-Mellor stands alone in its position, Shlaer-Mellor will be left behind. Said Hopkins, "I think Steve actually said it best, 'There's my (Shlaer-Mellor) camp and there's everybody else.'" Hopkins said that Rational and its supporters are very comfortable using the iterative and incremental way of developing. He does concede that translation is possible, "if you have full and unflawed knowledge of the domain, but that doesn't happen very often." Platinum's Rivas, on the other hand, feels that there may be less of a gap to bridge than Mellor suggests, but that as a meta-CASE tool vendor, his company's Paradigm-Plus tool will continue to support Shlaer-Mellor both today and as long as there is market demand.

OO, BPR and CASE convergence

Parallel trends are also involved in the convergence efforts in the OOAD industry, including the introduction of business objects. Although inconsistently defined and described, there is nonetheless interest in finding a way to componentize aspects of an industry into a reusable bundle. The OMG has already sponsored a special interest group that focuses on business object modeling, called BOMSIG, and is attempting to drive increased awareness and eventual convergence in the business object area. The activity is not isolated, however. The Business Object Facility (BOF) is another area that focuses on business objects, although at present BOF focus appears to be on how to define a common, core, enterprise object architecture with very little business object or methodology specifics yet in place.

A MOF (Meta Object Facility) group also exists and in the OMG Tampa, meeting the BOF, MOF and OOAD groups openly stated that their efforts must be complementary.

To support a business process engineering effort, several authors have introduced methodologies or extensions to existing methodologies to show how object technology modeling techniques could be used. Among these methodologists are David Taylor of Enterprise Engines Inc., San Mateo, Calif., whose previous book on managing object technology helped to introduce many to objects; Ed Swanstrom of Arthur D. Little, Cambridge, Mass., who took the foundation of Peter Senge's organizational change and learning from Senge's text, "The Fifth Discipline"and extended it in an object-oriented fashion; and Ivar Jacobson, who offered object-oriented business engineering.

Jacobson's use cases were introduced to OOAD from his Objectory methodology, with support from the company's tools and services. He documented parts of the Objectory methodology in his book, "Object-Oriented Software Engineering." Use cases offer a semi-formal way to capture functional user-level requirements. Together with objects, use cases were then shown by Jacobson to be equally appropriate to business engineering (as he referred to business object modeling) in his book, "The Object Advantage." A new trend to support business process engineering with object modeling is becoming apparent and tool vendors are updating their tools and/or forming partnerships to support this movement. Rational included objectory business process modeling extensions as a business process related aspect to their UML submission.

One such partnership resulted when OO CASE tool vendors joined the BPM rush. "Business Process Modeling is very key to us and it is one of the areas where we are putting a lot into research and development, since we feel that it is very important for people to be able to model their workflow independent of any software that automates that business," said Rivas of Platinum. Both Rivas and Platinum support the Object-Oriented Change and Learning (OOCL) methodology by Ed Swanstrom and Peter Senge. In supporting OOCL, Rivas hopes to offer an open approach to modeling business processes that tools supporting a proprietary methodology cannot offer.

The Fusion methodology

Developed by Derek Coleman, Fusion calls itself a "second generation" methodology, because it openly acknowledges the benefits of utilizing different methodologies and combining them with a more formal process. The methodology acknowledges the development process and should be watched. Fusion was increasing in popularity at about the time Rumbaugh joined Rational. When Fusion entered the mainstream, it seemed to validate the idea of converging complementary parts of different methodologies. In fact, the original Fusion combined some parts of Booch, Rumbaugh and Jacobson, as well as others. There is no question of the validity of the convergence approach for anyone using Fusion. (See Fig. 5.)

The Fusion methodology

The Fusion methodology openly acknowledges the benefits of utilizing different methodologies and combining them in a formal process.

Fig. 5

Source: Viktor Ohnjec


The Fusion Method has since become Team Fusion, jointly being developed by HP, LBMS, Houston and Derek Coleman. "The Team Fusion effort plans to extend the current Fusion method and to introduce UML as its notation," said Coleman. This will bring much needed process through Fusion to those that want to use UML, but cannot wait for the three amigos. A Fusion book is expected by the 1997 OOPSLA conference (October 5th through the 9th), about the same time as the Rational-supported UML books are expected.

The leading contenders in the convergence efforts of the OMG OOAD are Rational's UML and Platinum's OML-based submission. While the UML consists of a meta-model and a notation, it currently contains no process or full life-cycle support. Rational's Hopkins said, "The UML is not a meta-model for capturing all known methods, it is a specific modeling language with a specific meta-model that supports it." He further noted that this is the primary distinction between what the UML camp is doing and what Henderson-Sellers and Rivas (from the OML camp) want to do.

The UML submission is not trying to create a meta-model that will handle all the different existing tools, they are creating a model language that will support the interchange of models between different tools that support the same modeling language. "That's an important distinction between the two activities," said Hopkins.

According to Henderson-Sellers, OML, or the OPEN Modeling Language, consists of a full life cycle, process-centered methodology with emphases on reuse, quality, organizational issues (including people), project management, a meta-model and notation. OML appears to have the same scope as the UML.

Both the OMG OOAD effort and the UML effort at Rational are highly visible and open. Rational's Web site and the OTUG mailing list are used to distribute information and solicit feedback. OOAD CASE tool vendors capitalize on this openness by supporting whatever revision of UML is published and using their support of UML as a differentiator between themselves and competing products.

It is certainly an exciting time to be watching the OOAD industry. You can expect to see greater focus in the areas mentioned above, as well as new focus in currently untouched areas like OOAD support for legacy relational databases. The next one to two years will bring some form of convergence and standardization to the industry, and tool vendors are expected to have more and more integrated product offerings for both the project-centric analysis and design efforts, along with the more enterprise-oriented business process engineering activities that so many more development organizations are choosing to use.

For Further Reading

Booch, Grady. Object-Oriented Analysis & Design with Applications. Redwood City, Calif.: Benjamin/Cummings Publishing Company Inc., 1994.

Booch, Grady. Object Solutions, Managing the Object-Oriented Approach. Menlo Park, Calif.: Addison Wesley, 1996.

Coleman, Derek, et al. Object Oriented Development, The Fusion Method. Englewood Cliffs, N.J.: Prentice-Hall, 1994.

Jacobson, Ivar, et al. The Object Advantage. Menlo Park, Calif.: Addison-Wesley Publishing Co., 1995.

Jacobson, Ivar, et al. Object Oriented Software Engineering, A Use Case Driven Approach. Menlo Park, Calif.: Addison-Wesley Publishing Co., 1992.

Malan, Ruth, et al. Object-Oriented Development at Work, Fusion in the Real World. Upper Saddle River, N.J.: Prentice Hall, 1996.

Rumbaugh, James, et al. Object-Oriented Modeling and Design. Englewood Cliffs, N.J.: Prentice-Hall, 1991.

Senge, Peter. The Fifth Discipline, The Art and Practice of Learning Organization. New York: Currency Doubleday, 1990.

Shlaer, S., Mellor, S. Object-Oriented Systems Analysis, Modeling the World in Data. Englewood Cliffs, N.J.: Prentice-Hall, 1988.

Shlaer, S., Mellor, S. Object Lifecycles: Modeling the World in States. Englewood Cliffs, N.J.: Prentice-Hall, 1992.

Taylor, David. Business Engineering with Object Technology. New York: John Wiley & Sons, Inc., 1995.

Taylor, David. Object Oriented Technology: A Manager's Guide. Menlo Park, Calif.: Addison Wesley, 1990.

"Unified Method for Object-Oriented Development," Documentation Set, Version 0.8. Santa Clara, Calif.: Rational Software Corp., Grady Booch and James Rumbaugh, October 1995.

"The Unified Modeling Language for Object-Oriented Development," Documentation Set, Version 0.9 Addendum. Santa Clara, Calif.: Rational Software Corp., Grady Booch, James Rumbaugh and Ivar Jacobson, July 1996.

Taking inventory of OOAD Tools

Object-oriented application technology is one of the world's fastest growing technology areas. International Data Corporation (IDC), a research firm based in Framingham, Mass., predicts by 1999, companies worldwide will spend more than $4 billion annually on object-oriented application development tools. According to IDC, the market for object-oriented analysis and design (OOAD) tools grew to $127.5 million in 1995, up 43.7% from 1994.

A list and brief explanation of selected object-oriented application development tools follows:

ROSE: Rational Software Corp.

Santa Clara, Calif.-based Rational's goal is to add several the features of the acquired Objectory methodology into Rose over a two-year period. "From a process prospective, the plan is to continue supporting the Objectory process and roll it into the Rational process -- since both approaches are similar to begin with," said Jon Hopkins, Rational's director of object technology. Rational incorporated Use Cases, added a Windows 95-compliant interface and added both a scripting environment and an API with a customizable developers in the new Rose version. The new Rose finally adds support for the UML notation, which some competitors had already been supporting. Rose 4.0 also adds a component focus by repositioning Rose's object-oriented features and adding general component services.

The product was not available for review at the time this article was written.

PARADIGM PLUS: Platinum Technology Inc.

Oakbrook Terrace, Ill.-based Platinum's Paradigm Plus, Version 3.5 integrated UML with its business process modeling, physical database modeling and object modeling capabilities. The new Paradigm Plus version provides a bridge between business processes, applications and database resources. It is repository-based, supports multiple methodologies and provides true round-trip engineering support for a number of popular languages.

Paradigm Plus supports access to relational database systems (RDBMS). Mapping object-oriented analysis and design techniques to and from RDBMS tables can be error-prone and time-consuming, so Paradigm Plus automates the process, providing a range of options for mapping from an object to an entity-relationship-type physical database model. Paradigm Plus can generate or reverse engineer schema for many database environments.

The meta-CASE technology in Paradigm Plus 3.5 can automate most leading object-oriented methods including OMT, Fusion, Booch, OOCL, Martin/Odell, Shlaer/Mellor, Coad/Yourdon, UML V0.9, and can extend support to Jacobson's Use Case modeling. The meta-CASE technology makes it easy to add new notations as they become available. Paradigm Plus was the first OOAD tool to support UML in May 1996.

Round-trip engineering for C++, Visual Basic, Smalltalk, Powerbuilder, Platinum ObjectPro and Forté enables iterative development and synchronization of designs with implementations. Customizable-code generation is provided for Java, Delphi and Ada. A distributed object repository supports teams of concurrent users, object-level locking, versioning and access rights.

Paradigm Plus is a piece of the larger Application Lifecycle Solutions (ALS) integrated product set and of Platinum's component warehouse philosophy. The single largest difference between Platinum and other OOAD tool vendors is the ALS ability to establish a reuse repository offering shareable components through web-based search tools, common browsers and standard download techniques to acquire

FRAMEWORK: Process Technology Inc. (Ptech)

Process Technology Inc., Westborough, Mass., provides tools for modeling and automating the capture of business processes and information. FrameWork uses a graphical interface accessible by all users from boardroom to computer room to divide the scope of modeling into:

  • Activity models that capture high-level business strategy and goals;
  • Object models that represent structural and functional detail; and
  • Process models that reflect the dynamics of the business.

Framework can capture the underlying semantics of a business and validate entries in order to assure that models are consistent and logical. In addition, tPtech models can be extended infinitely, enabling users to customize the tool to a preferred methodology. The business processes and information captured in the models are stored in a single repository, the Ptech KnowledgeBase, which can then be accessed by any user within a corporation. Benefits include an ability to capture business expertise and to efficiently and cost-effectively disseminate best practices throughout an organization.

FrameWork shares a formal foundation with the Martin/Odell methodology and Ptech promises it will offer full support for the Unified Modeling Language (UML) by March, 1997. Ptech has stated its intent to submit to the Object Management Group a proposed specification of business process related extensions to the UML in order to establish a process modeling standard.


Aonix, San Francisco, inherited and continues to offer Software Through Pictures (StP) product following parent company Thomson CSF's acquisition of Interactive Development Environments Inc. last year. Aonix is a joining of Thomson firms IDE and Thomson Software Products.

Aonix sells a multiuser multidimensional model repository based on the StP analysis, design and code generation toolsets. The multidimensional modeling architecture allows support for cross-functional model-driven development processes, thus unifying project managers (and their project plans), business analysts (and their process analyses and designs), application developers (and their objects and legacy structured code), data modelers (and their RDBMS), QA (and their test suites) -- and multiple methodologies, tools and deployment technologies.


Southfield, Mich.-based Proforma's BPM focused tool can create a variety of models and display them to several methodological approaches including UML. ProVision Workbench (PVW) supports business interaction modeling, workflow modeling, use case modeling and hierarchy modeling for BPR efforts. For business object analysis, modeling can be performed in any object, subtype, interaction, state or method modeler role with rich support tools available throughout. Integration with C++ and IBM's VisualAge products allow forward and reverse engineering. Results can be published in the HTML format.

SELECT ENTERPRISE V5.1: Select Software Tools Inc.

The Select Enterprise toolset from Select Software., Santa Ana, Calif., joins three key components -- modeling, architecture and process.

The modeling techniques are based on 'best of breed' standards. Business Process Modeling techniques include process hierarchy and process thread models as defined by the 'Catalyst' method from Computer Sciences Corp. Object-oriented modeling techniques include use cases and object interaction models based on Ivar Jacobson's OOSE, and object and dynamic models based on James Rumbaugh's OMT.

Process modeling and actor modeling are supported. Process models can be represented as process hierarchies, which break high-level business processes into elementary business processes.

Distributed architectures are modeled based on a Select Enterprise four-schema architecture. Modeled interface and business classes are mapped to PowerBuilder visual and non-visual objects respectively. Storage (persistent) classes are mapped to relational tables for generation to DDL. A schema generator is included.

Development lifecycle support is focused on iterative, incremental RAAD (Rapid Architecture Application Development) development. The Select Enterprise repository can manage and store modeled objects in a multiuser environment. An object has one occurrence in the repository ensuring that each iteration updates that occurrence. Round-trip engineering supports iterative and incremental code generation enabling models and code to be kept in step. The toolset supports Forté, Visual Basic, C++, PowerBuilder, Centura, Smalltalk, NatStar and Delphi. A Java version will ship by March, 1997. Version 5.1 also includes middleware integration through Corba IDL and ActiveX generation.

Through an included component, Select Enterprise Process Mentor SELECT offers a pragmatic approach to software development.

GRAPHICAL DESIGNER: Advanced Software Technology Inc.

Littleton, Colo.-based Advanced Software Technology's Graphical Designer 2.0 provides powerful tools that support software design and re-engineering. Version 2.0 adds the Graphical Designer product family that can support developers in an integrated fashion with other installed tools throughout the software lifecycle. The Graphical Designer family includes three products: GDDraw; GDMethodBuilder; and GDPro. GDDraw provides powerful engineering drawing facilities, including user-defined drawing palettes. GDMethodBuilder can extend all aspects of, or build entirely new design methodologies. GDPro provides support for a variety of software design methodologies, including code generation and reverse-engineering.

GDPro is a design and re-engineering toolkit that supports methodology based design and the reverse-engineering and re-engineering of software systems. Supported methodologies include: use cases; UML; Rumbaugh OMT; and Booch. The toolkit supports "seamless" round-trip engineering and can modify both source code and designs and have both automatically remain synchronized. Re-usable components can be stored in GDPro through the integrated repository.

WITH CLASS: MicroGold, Inc.

MicroGold Inc., Edison, N.J., supports some UML .8 features in With Class Version 3.1. UML 1.0 support is promised for version 4.0, slated to ship in March 1997. With Class v3.1 supports Coad-Yourdon, Shlaer-Mellor, OMT and Booch methodologies. With Class supports process modeling via either operation models and object interaction diagrams. With Class has a scripting language that can generate code or reports for any OO language. It includes scripts for Visual C++, Java, Delphi, Visual Basic, Borland C++, Smalltalk, Eiffel, Ada (version 4.0 will add Delphi). With Class can convert C++, Java, Delphi and Visual Basic into class diagrams.

OBJECTMAKER: Mark V Systems Ltd.

ObjectMaker support for OML and UML has recently been announced. ObjectMaker allows modeling of requirements and design in more than 30 OO and non-OO analysis and design notations and allows for the mixing and matching of notations to build a single application model. ObjectMaker supports detailed design, code generation and document generation for construction, delivery and maintenance phases. Reverse engineering is also supported.


Although not an object-oriented Case tool, the toolset from Visio Corp., Seattle, has found tremendous interest because of their ease of use and vast symbols library. Visio Professional includes more than 1,000 SmartShapes symbols. Visio Business includes more than 1,000 additional SmartShapes symbols for business process analysis diagrams. Wizards and templates are included for all common methodologies, which can include UML, Fusion, Booch, OMT and Shlaer-Mellor.

S-CASE: MultiQuest Corp.

S-Case from MultiQuest, Schaumburg, Ill., supports both UML and Booch. Booch support is implemented on top of the UML metamodel allowing an easy transition from Booch to UML. S-CASE offers forward and reverse engineering capabilities for C++ code. Code generators for C, Delphi and Java can be downloaded from the MultiQuest Web site. S-CASE supports multiple team member development through basic check-in/check-out facilities that give control of changes to the developer checking out a component. Programmatic access to the tool is available through a scripting language. The tool does not support BPM.