[
Lists Home |
Date Index |
Thread Index
]
- From: Andrew Layman <andrewl@microsoft.com>
- To: "xml-dev Mailing List (E-mail)" <xml-dev@ic.ac.uk>
- Date: Wed, 16 Dec 1998 14:47:08 -0800
Thomas Bergstraesser asked that I forward this to you:
> XMI is a proposed OMG standard for the exchange of meta data based on MOF
> (Meta Object Facility). MOF is an OMG standard for repositories, which
> consists of a set of basic types and relationships used to implement
> information models. MOF is a CORBA-specific object model. To apply XMI to
> other component models, such as COM, requires wrapping or bridging with
> their inherent integration and performance penalties.
>
> XMI defines the mapping of MOF objects and relationship into XML. However,
> to be useful to store and interchange meta data, XMI must be used with
> information models that capture the semantics of types, such as table,
> column, transformation, or COM component. Such a type specification is
> required for meta data to be accessed or interchanged in a meaningful way.
> In absence of such information models, MOF and XMI are only enabling
> technologies.
>
> One of the standard information models needed for meta data exchange is
> UML. UML is a notation and a meta model that describes the semantics of
> the notation. MOF and UML are two different models. However, the UML meta
> model can be interpreted using MOF and therefore UML models can be
> exchanged in XMI. The exchange of UML models between MOF/CORBA repository
> prototypes of IBM /UNISYS / Oracle was demonstrated at an OMG meeting on
> November 15, 1998. In a related activity, OMG issued an RFP for a Common
> Data Warehousing Model (CDWM) with a submission deadline of Q4 1999.
>
> Microsoft currently does not see much customer demand for a
> MOF/CORBA-based repository implementation, which in any case would be an
> alien implementation in the COM world. Microsoft does, however, see the
> value of the technology independent UML standard and has supported and
> implemented it. UML is the core of the Open Information Model and
> Microsoft's Visual Modeler and is the base for the UML model interchange
> initiative currently supported by 60+ vendors. The initiative was
> kicked-off with an interchange demonstration between Microsoft / Rational
> / Platinum / Popkin / LogicWorks and others in January, 1997.
>
> Microsoft uses XML to interchange meta data described by the Open
> Information Model (OIM), which covers analysis and design (UML), component
> development and deployment (CDE), and database design and data warehousing
> (DBM). The OIM is described in UML, and the XML Interchange Format (XIF)
> for OIM is generated directly from its UML representation. It is therefore
> technology independent and vendor neutral and can be implemented by any
> repository or application. Microsoft and the leading repository vendors
> have implemented the XML-based interchange for the Open Information Model
> to enable customers to integrate meta data from many heterogeneous sources
> in the application development and data warehousing domain.
>
> The use of pure XML to interchange meta data described by OIM seems
> prudent if one considers current standard efforts in the W3C around XSL,
> XQL, and XMLdata. These technologies will constitute the standard
> framework for the exchange of meta data on the Web. The OIM as
> specification of the semantic of meta data elements in the application
> development and data warehousing domain will make use of these W3C
> technologies whenever they become available and therefore protect your
> investment.
>
> To summarize the different approaches for the interchange of meta data:
> Open Information Model / XML:
> * Open Information Model including UML shipping in SQL Server 7.0
> * XML Interchange Format generated from UML model shipping in SQL
> Server 7.0
> * Implementation on multiple repositories available
> (Informix, Microsoft, Softlab, Unisys, Sybase, Viasoft, NCR, Siemens,
> Platinum)
>
> XMI/MOF:
> * No common information model (UML prototype implementation, OMG RFP
> for CDWM in Q4 '99)
> * XML Meta Data Interchange (XMI) based on MOF/Corba model
> * MOF/Corba repository implementation required (Unisys; announced by
> IBM, Oracle, Select)
>
> Sorry for this lengthy memo stuffed with technlogy mumbel-bumble. From
> Microsoft's perspective
> the interchange of meta data in the application development (AD) and datra
> warehousing (DW) area is (and should be) very simple:
>
> - OIM is the semantic description of meta data in the AD (includes UML)
> and DW domain
> - UML is the specification language for OIM to communicate the structure
> and semantics
> of its meta data typs in a formal way
> - XML (and XQL, XSL, and XML schema description) is the interchange format
> for meta data
> described by OIM
>
> - thomas
>
> -----Original Message-----
> ---forwards removed---
>
> -----Original Message-----
> From: Dave Carlson [mailto:dcarlson@ontogenics.com]
> Sent: Wednesday, December 16, 1998 10:31 AM
> To: xml-dev@ic.ac.uk
> Subject: RE: DTD-to-DB SCHEMA
>
>
> After a brief scan of Microsoft's site on OIM and the associated XML
> Interchange Format (XIF), I'm concerned and confused about the
> conflicts
> with the OMG's XML Metadata Interchange (XMI) specification. XMI is
> in it's
> second draft as a standard, and most of the large UML tools vendors
> have
> voiced support (IBM, Platinum, Rational and Oracle are co-authors).
>
> It appears that OIM and XIF are more focused on database and data
> warehouse
> integration, but they also claim full UML metamodel support.
> Currently, I'm
> focused on OMG's XMI as the likely standard for UML model
> interchange.
> However, I'm very interested on other opinions.
>
> Dave Carlson
> Ontogenics Corp.
> Boulder, Colorado
>
> At 12:04 PM 12/16/98 -0600, lauzon@us.ibm.com wrote:
> >Thanks for all the response I got from this note. Two of these
> >particularly look interesting. One is the OIM from Microsoft that
> is
> >described by Thomas Bergstraesser, and the other is the Metadata
> >Interchange Specification (MDIS) from the Meta Data Coalition
> (MDC).
> >
> >The MDIS seems to be almost exactly what I'm looking for to
> describe
> >relational databases. Unfortunately it doesn't really seem to be
> in XML,
> >although they do mention using XML to distribute these MDIS
> documents. OIM
> >has been submitted to the MDC to become a standard, but it seems
> overly
> >complex for what we're trying to do, although that might be just
> because I
> >haven't had enough exposure to it. What I'm not sure about is how
> >widespread support is for MDIS in the XML community, or if there is
> any
> >thought about it whatsoever. Or is the trend for a standard based
> upon
> >OIM?
> >
> >Shawn Lauzon
> >Department MMB - San Francisco Database Persistence
> >email: lauzon@us.ibm.com
> >phone: (507) 253-6966 T/L 553-6966
> >
> >
>
>
> xml-dev: A list for W3C XML Developers. To post,
> mailto:xml-dev@ic.ac.uk
> Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
> To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
> (un)subscribe xml-dev
> To subscribe to the digests, mailto:majordomo@ic.ac.uk the following
> message;
> subscribe xml-dev-digest
> List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)
xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)
|