This Is A Custom Widget

This Sliding Bar can be switched on or off in theme options, and can take any widget you throw at it or even fill it with your custom HTML Code. Its perfect for grabbing the attention of your viewers. Choose between 1, 2, 3 or 4 columns, set the background color, widget divider color, activate transparency, a top border or fully disable it on desktop and mobile.

This Is A Custom Widget

This Sliding Bar can be switched on or off in theme options, and can take any widget you throw at it or even fill it with your custom HTML Code. Its perfect for grabbing the attention of your viewers. Choose between 1, 2, 3 or 4 columns, set the background color, widget divider color, activate transparency, a top border or fully disable it on desktop and mobile.

DMN Tools – State of the Market May 2016

//DMN Tools – State of the Market May 2016

DMN Tools – State of the Market May 2016

It’s still early days for DMN, but a number of tools are available or have been publicly demonstrated.  Here is my impression of where we are at the moment.

DecisionsFirst Modeler

James Taylor’s free DecisionsFirst Modeler was one of the first DMN tools available.  It allows a business user to draw DRDs and enter a wealth of business context behind each decision node.  Instead of providing a DMN native decision logic editor for decision tables and literal expressions, it links to “implementations” of DMN decision nodes created in Drools, ODM, and other commercial business rule engine products.  Users fluent in those specific rule languages (and presumably licensed to use those tools) can then view and edit the decision logic via the DecisionsFirst portal.  DecisionsFirst Modeler provides a team repository featuring collaborative editing and commenting.   The tool runs on Silverlight, meaning it doesn’t work on Chrome or Edge browsers, and it doesn’t appear to support business knowledge models.

Signavio

Signavio provides what I believe to be the most full-featured DMN editor today.  Users can create DRDs and link them to decision tasks in Signavio’s BPMN editor, which shares a common repository and dictionary.  It also provides a nice decision table editor and data-aware expression editor for literal expressions.  The decision table syntax is not strictly S-FEEL, but the decision table editor includes a Verify feature that reports on gaps and overlaps in the rules, a great feature.  Signavio provides its own reasonably friendly expression language and function library, but it is not the standard expression language, FEEL.  Signavio provides a DRD extension called Multi-Instance Decision to solve the need for iterated decision logic.  Best of all, Signavio’s Simulation feature allows end-to-end testing of DRD logic: supply values for the input data elements, and Simulation reports the output of any selected decision in the DRD.  Signavio can include BKMs in a DRD but cannot model their decision logic.  Signavio is the tool I currently use in my DMN Method and Style training.

FICO

FICO recently made available a free DMN Modeler.  In its current version, it fully supports DRDs, decision tables, and literal expressions using S-FEEL.  As such, it is probably the only tool today that technically can claim DMN Conformance Level 2.  (Although Signavio is more capable today, it does not adhere to S-FEEL syntax, the basis of CL2.)  According to a source at FICO,

“The next stage for FICO will be to introduce DMN as the primary user interface for our cloud-based Decision Management Suite.  Users will be able to configure an entire decision management solution as a DRD in which BKMs can be any functional components which can be created on DMP:  Blaze Advisor (sorry, Decision Modeler) rulesets, decision tables or decision trees;  SRL functions;  analytic models imported from SAS or from FICO Analytic Modeler;  optimisation models;  etc etc.  The defined solution can then be deployed directly as an executable on DMP.  We already have a first implementation of all this, not yet available as a product.  CL3 compliance is a secondary objective and FICO may decide to use SRL rather than FEEL for the primary decision logic language in DMS.  We might provide SRL-FEEL conversion for openness but SRL has a good deal more functionality than FEEL so not all models created with DMS would be capable of representation in pure DMN.”

Camunda

Camunda provides a free DMN tool focused on simple routing decisions in BPMN models.  It does not support DRDs at all, just decision logic modeled as a single decision table.  Its decision table editor supports S-FEEL and various programming languages as well, and provides a Simulation feature to allow testing of the decision logic.  Decision logic modeled in the DMN tool can be deployed as a decision service within the Camunda BPMS.

Trisotech

Trisotech offers a DMN Editor as part of its Digital Enterprise Suite of modeling tools.  Unlike most other vendors, Trisotech is committed to supporting the full DMN standard, including FEEL and boxed expressions, something I very much welcome.  From a drawing perspective, the current offering supports DRDs, decision tables, the full range of boxed expressions, including Contexts and Relations.  Through a partnership with BRE vendor Idiom, Trisotech can simulate S-FEEL-based decision tables.  According to a Trisotech source, “Topics remaining to address for us include: FEEL/S-FEEL Expression Editor, FEEL/S-FEEL Validation, Decision Table Completeness Validation, Test Case Validation, execution of FEEL. (We currently support S-FEEL execution in decision tables only.)”

Oracle

Oracle’s DMN Modeler is still in the lab, but was demonstrated at bpmNEXT 2016 (watch for their presentation video, coming soon).  The current implementation includes full support for FEEL and boxed expressions, but – somewhat surprisingly – not yet DRDs, just a linear chain of subdecisions (effectively a Context for the top-level decision).

What This Tells Us

What this tells us is, first, that DMN tools are beginning to proliferate but are still immature.  What is more worrisome is that few vendors are feeling the pressure to fully comply with the spec, meaning implementing FEEL and boxed expressions.  Without those things, DMN cannot achieve its objectives.  Hopefully that pressure will come!

Note: The summary above reflects my best understanding.  If I have misrepresented a product, please contact me and I will endeavor to correct the post.  If you have a DMN product or are building one and would like to be included in future updates on this topic, please contact me.

By | 2016-12-29T13:40:00-08:00 May 13th, 2016|DMN|1 Comment

About the Author:

One Comment

  1. jamet123 May 16, 2016 at 10:07 am

    Bruce
    Nice summary. I would only add two things regarding DecisionsFirst Modeler:
    – Users need only be fluent in the rule editors of their target BRMS – ODM’s Business Console or RedHat’s Business Center for instance. They don’t need to have any skill or experience with the technical environment because the models can be deep linked to the correct decision table, decision tree, ruleflow, ruleset etc.
    – BKMs are fundamentally about reuse. DecisionsFirst Modeler does not support them but as the link from Decisions to Implementation Components is many:many, any Implementation Component representing reused decision logic could be linked to all the decisions that use it. Thus Implementation Components fulfill the use case for BKMs even if they are not (yet) diagrammed as such

    I also think it is worth remembering that the ability to fully define a working implementation is only one of the three use cases defined for DMN – requirements for automation and defining manual decisions are considered equally valid use cases by the standard.

    Finally it would be worth identifying the execution environments supported by each tool.

    James

Leave A Comment