META Group

[Kelly Dilks]

 Hello,
 
 Is there anyone in the digital library community who has been in
 contact with anyone in the META Group?  A brief article about their
 goals is below this message in ascii format.  I am interested
 in their efforts since they are dealing with formats for metadata and
 I work with the FGDC Clearinghouse Working Group so am interested in
 anything that deals with formating of data about data.
 
 Any help will be greatly appreciated.

 Please send any replies to me directly at dilks at vinny.cecer.army.mil
 and I will forward info onto anyone who requests it.
 
 Thank You
 
 Kelly Dilks
 U.S. Army Corps of Engineers



> ----------
> X-Sun-Data-Type: default
> X-Sun-Data-Description: default
> X-Sun-Data-Name: METAgroup
> X-Sun-Content-Lines: 86
> 
> Managing Metadata 
> 
> META Group and vendors to develop minimum standard 
> 
> By Bradley F. Shimmin 
> 
> Through a flexible standard for metadata, or information about information, META Group Inc. and six data-management and
> development-tool vendors hope to do away with the complexities surrounding the management of heterogeneous datastores,
> client/server applications, and data warehouses. 
> 
> Because data-management and applications-development tools often reference and represent information about information, they tend
> to create functional yet inaccurate situations similar to this sentence: '"Is not preceded by a quotation" is not preceded by a quotation.' 
> 
> To combat this issue, the Metadata Council last month proposed a standard for the exchange and storage of metadata called the
> Metadata Interchange Initiative. 
> 
> Expected to be published as a working document by year-end, the initiative originated through an effort by META Group to give tool
> vendors a single standard through which they could exchange metadata, according to Karen Rubenstrunk, director of applications-
> development strategies at META Group in Stamford, Conn. 
> 
> The Metadata Council consists of six tool vendors--Arbor Software Corp., Business Objects Inc., Cognos Corp., Evolutionary
> Technologies International, Platinum Technology Inc., and Texas Instruments Software, a division of Texas Instruments Inc. Together
> these companies plan to produce a two-phase standard for metadata interchange. 
> 
> The first phase will create a standardized ASCII file format, which vendors can use to import and export metadata. The second phase,
> scheduled to appear one year later, will consist of a standardized API through which vendors can directly and automat- ically exchange
> metadata. 
> 
> However, the standard does not stem from a simple need to exchange information, according to Rubenstrunk. Rather it arises from the
> fact that each tool represents data and metadata differently. "Within any tool, whether it's a computer-aided software-engineering
> tool, a model developer, a data dictionary, a database itself, or a decision-support software tool, the way they work is to create
> metadata," she explained. "But that data is always proprietary, so these tools can't talk to each other." 
> 
> For example, an Oracle Corp. RDBMS (relational DBMS) of 300 tables might maintain a data dictionary that houses thousands of
> table relationships. For the database administrator, that information is easily accessible, but for a decision-support software (DSS) tool
> user, that same information is hidden from view. To gain access to that information, the DSS user must input and manage the thousands
> of pieces of metadata separately from the original metadata container. 
> 
> Yet the additional management costs in coordinating this information is not the entire problem, said Neal Hill, vice president of
> marketing for Business Intelligence Products at Cognos, a vendor of online analytical-processing tools in Burlington, Mass. Because
> the metadata in different tools is stored in different ways, a change to the representation of data in one tool may cause unanticipated
> consequences within another tool or even the data itself, he explained. 
> 
> "Each time you increase or decrease capacity in manufacturing, you may be causing a ripple effect in other related data sets," Hill said.
> "Over time, you may be sowing a series of future time bombs for yourself completely without knowing it." 
> 
> These time bombs occur because administrators and developers change one tool's description of a data set without changing the
> description of that data set within other tools, said John Mann, director of client/ server computing services at the Yankee Group, a
> consulting firm in Boston. "Before, managers saw a number on a report and believed it," Mann said. "Now they've gone into their PCs
> and have seen the raw data and have seen how fuzzy that data is, how it is not so absolute as they thought." 
> 
> For Robert L. Scheier, senior analyst with Hurwitz Consulting Group Inc. in Newton, Mass., the Interchange Initiative's goal of
> providing a single standard for metadata interchange, though worthwhile, will be difficult to logistically and ideologically implement.
> "The underlying problem, which is true for all such groups, is that every vendor is simultaneously attempting to be as open as they have
> to be to meet market requirements, while at the same time being as proprietary as they can be to add value to their own products,"
> Scheier explained. "This is a difficult enough balancing act within one vendor much less trying to get five or six together trying to
> agree on a common agenda and a common rollout for actual deliverable products." 
> 
> The Metadata Council hopes to solve these problems by creating a minimal, obligatory set of metadata-component descriptions. These
> elements include only information about a component's identification, description, owner, and type. Outside of this information, each
> vendor is free to maintain other information specific to that vendor's tool, said Katherine Hammer, president and CEO of Evolutionary
> Technologies Intl., a vendor of data-transformation and -migration tools in Austin, Texas. 
> 
> Like the LISP (linked list processing) programming language, which houses objects containing attributes that cannot be seen by other
> objects, the Interchange Standard lets each tool's metadata components carry similar information that is unavailable to other tools,
> explained Hammer. "Rather than cataloging every piece of information," she said, "it lets tools come up with a standard way to
> communicate these things and a way to carry along the private things." 
> 
> Because the Interchange Initiative is made up of only a fraction of the data-tool vendors, some analysts believe its loose structuring of
> metadata may prove successful by not locking users into a particular vendor or set of vendors. It is analogous to the 1992 IBM-led SQL
> standard, which defines a core set of features while leaving room open for each vendor to add value, said the Yankee Group's Mann.
> "Just like the SQL standard, it might allow the customers to feel more comfortable," he said. "It is nice to have standards so you feel
> OK about buying this metadata product." 
> 
> However, Mann warns that while the standard promises to lessen the amount of metadata management within an organization, it will
> exact its own level of management overhead. "The mere act of setting up a repository and putting metadata in it still begs the question
> of how well the data is understood and how clear the description of the data is," he said. "Therefore, there need to be human beings
> involved in the management of metadata."  
>