OMG Internet Platform Special Interest Group

Minutes of Meeting #17

June 8, 1998
Orlando, Florida
OMG document internet/98-06-01
          OMG Internet Platform SIG homepage:



Interoperability Clearinghouse Initiative, John Weiler, OBJECTive Technology Group

John Weiler, OBJECTive Technology Group (and OMG's Government Liaison), gave a brief presentation on the Interoperability Clearing House Initiative. This joint government/industry initiative is being headed by the OBJECTive Technology Group under a DARPA contract, with oversight by the Office of the Secretary of Defense. The primary mission is to provide both architecture tools and a supporting Web-based repository that help in the selection, management, and validation of interoperable software components. An initial goal is to define a mapping showing how various component technologies fit together in an enterprise systems architecture. By enabling links between interdependent components, users will be able to determine which set of standards and associated products interoperate to form a virtual Common Operating Environment. Partners in the effort include standards organizations, user groups, component suppliers, system integrators, and independent software validators. More information on the initiative can be found at and

Web Compatible CORBA Metadata, Ron Zahavi and Robbie Tischer, Concept 5 Technologies

Notes: Additional metadata capabilities are needed to address a number of problems in CORBA-based systems. IDL provides mostly syntax, not semantics. Support is required for both large- and small-grained objects. It is also desirable to be able to look at individual differences among instances, and have dynamic discovery and introspection. MOF is useful, but the metamodel is not extensible, it supports only IDL, and no standard transfer format is defined. Beans provides no access to fine-grained content, and metadata only on interfaces and methods. Web-related technologies that can help provide the needed capabilities include XML, DOM, RDF, and XML-Data, and work on specific metadata properties such as Dublin Core. Providing CORBA metadata in XML would allow access both within the CORBA environment and by Web browsers.

The basic idea is to describe CORBA representations using XML metadata, and add tags to define more precise type information. The XML tags are defined by starting with the OMG BNF syntax as the basis of the DTD, defining XML tags for CORBA meta concepts. Additional tags can then be added that further constrain types (e.g., along the lines of COBOL-like PIC definitions).

XML allows arbitrary mixing of tagged information that represents both metadata and data. As a result, it is possible to start with a set of XML tags that represent a type, and then create instances by adding additional tags that hold the instance-specific data. The approach also provides an extensibility framework. By adding XML to an existing DTD, a new DTD can be generated which includes new tags defined by that XML.

Next steps for this technology include integration with RDF and XML-Link. If the ISIG thought such technology was promising, it might cnsider issuing one or more RFPs in the area of XML/CORBA integration.

Update on W3C HTTP-NG Effort, Mike Spreitzer, Xerox PARC

Abstract: The HTTP "Next Generation" project in the World Wide Web Consortium is an experiment in updating the HTTP protocol to a distributed object protocol. This update will present the current status of that experiment. More information on the project is available at

Notes: HTTP-NG is an experiment aimed at redesigning the Web as a distributed object application. The work is being done in three levels:

The Web application layer provides an object-oriented network API for HTTP 1.1, and is designed to support "TCWA" (The Classic Web Application). The Messaging layer essentially defines middleware support, and provides an OO RPC protocol. The Transport layer defines a set of protocol stacks (some portions being composable), and includes support for multiplexing and bi-directionality. HTTP-NG uses binary formats on the wire to reduce parsing overhead and the number of bytes transmitted. Advantages of HTTP-NG include efficiency (including minimization of round trips), extensibility (e.g., to support new applications like WebDAV--Distributed Authoring and Versioning), and simplicity.

Phase I of the project ends this summer, with the first release of drafts scheduled for early July. A basic feasibility demo will also take place. The results will be presented to IETF at their August meeting in Chicago. A more extensive briefing will be given to the ISIG at its Seattle meeting.

Component Technology for Dynamic CSCW Environments, Eric Hughes, MITRE

Notes: This talk described how component technology, such as JavaBeans, can be used to support CSCW (Computer Supported Cooperative Work). Eric described current capabilities for using CORBA in this type of application (e.g., OrbixTalk, the Event Service). The general architecture is a classic 3-tier architecture. Components such as JavaBeans can run in anyof the three tiers. Enterprise JavaBeans (EJBs) are designed for the middle tier, supporting transactional, data-intensive applications. When using them, clients are provided with stubs to access these middle-tier components. It is not always clear how to map CSCW applications onto this 3-tier architecture, although in some cases the mapping is straightforward. Eric also described FACELIFT, an FY98 C4I demonstration, based on this approach. It uses JavaBeans at the client tier, EJB + CORBA + an HTTP server in the middle tier, with databases and legacy systems at the bottom tier.

XML|IT Toolkit, Juggy Jagannathan, CareFlow|Net

Notes: This brief talk described CareFlow|Net's XML|IT toolkit, initially targeted at healthcare applications. The toolkit automatically generates XML-tagged CORBA service calls, and provides utilities to convert Java structures to and from XML. CORBA requests are translated into XML based on the IDL interfaces, having tags like:
<servicename> (having an

IOR as value)






which can be sent to servers using a Java DII call (referencing the structure getaddress.xml, in this case). The basic idea is that much healthcare data will probably be represented in XML anyway, so it will be useful to have everything represented in XML so that common manipulation techniques can be applied.

Formation of Working Groups

Following the presentations, the ISIG voted to form three Working Groups: The first two of the WGs were scheduled to have concurrent breakout sessions in the afternoon, followed by a cross-briefing, after which there would be a plenary session for the third WG (this general pattern is planned for future meetings, with the WGs rotating having the plenary session).

Web/OMA Integration Architectures Working Group Breakout, Frank Manola, Object Services and Consulting

The WOIA Working Group defined as its general goals: These are becoming very important, given the increasing number of systems being built which incorporate both CORBA-based distributed object technology and Web technology, and the increasing frequency with which Web technologies such as XML are being referenced in OMG Task Force activities.

The group also identified as important activities:

Manola indicated that he would produce a draft initial charter, and distribute it via email to get discussion started prior to the Helsinki OMG meeting in July. He also solicited input on Web/OMG applications. Overall, there was a great deal of interest and enthusiasm shown for this effort.

NB: The OBJS Web Object Model report (containing material presented to the ISIG at the East Brunswick meeting in December) provides an introduction to many of the relevant technologies

CSCW Working Group Breakout, Henry Rothkopf, MITRE

The CSCW Working Group is looking at developing open standards to support component based, "tailorable", collaborative, place-based or instantiation-based environments. Through these, a user may meet with others, either synchronously or asynchronously, in a "room", leave materials in the room for review, etc., share applications, voice, video, and so on. The "room" might be an instantiation, perhaps a node in a self-building work flow diagram. Some of these capabilities are just now coming to market, but not in an interoperable, web-based way. There is much work to do. So far, MITRE, Hitachi and Lotus are involved.

The WG's Mission is: To examine and report on opportunities for fullfilling emerging requirements for Computer-Supported Cooperative Work (CSCW) within distributed, component-based enterprise.

A password accessed discussion has been created for the Workgroup, located at g/cscw. Send e-mail to Henry Rothkopf if you want to join.

Working Group Cross Briefings

Following the breakouts, the first two WG's exchanged cross-briefings. Henry Rothkopf and Frank Manola described the general goals of their respective Working Groups, as reported above. Henry additionally noted that the CSCW WG intends to work on a Green Paper, and cooperate in depth with the other WGs. Frank Manola noted that WOIA, in addition to standing for "Web/OMA Integration Architectures", might also stand for the general question it attempts to answer: "Why Objects and the Internet Anyway?".

Work Session on OTAM, Mike Bigrigg, CMU

Mike Bigrigg led a working session on OTAM (effectively, the OTAM WG meeting). This is aimed at defining an OMG facility, based loosely on the ISO FTAM specifications, for a form of virtual file system that allows access to, and maintenance of, virtually any type of data. Mike circulated both a draft OTAM Green Paper and a draft RFI. There were a number of questions raised as to the need for the proposed facility, how it related to existing OMG specifications, scenarios in which it would be used, etc. More information on these subjects is needed before an RFI can reasonably be issued. Frank Manola had written up a list of questions/comments on the Green Paper, which he agreed to send to the entire group via email following the meeting.