The CyberStacks(sm) approach is more holistic in presenting a resource profile. It seeks not to delineate all relevant elements in describing a resource, but to characterize the resource sufficiently so that users can judge its potential usefulness and to simultaneously integrate the resource within an appropriate framework that offers a context in which the relationship to other resources are indicated. We believe that the meaning and value of a given resource may be conveyed as well, or better, by structure and organization. We believe that the CyberStacks(sm) scheme provides meaning through context; a context that is appropriate and relevant to users seeking a resource which may satisfy an information need.
We are fully aware of efforts to create and use a standardized record format for cataloging selected World Wide Web and other Internet resources. Of particular note is the OCLC/NCSA Metadata Workshop Report, a report of the March 1995 Metadata Workshop, sponsored by the Online Computer Library Center (OCLC) and the National Center for Supercomputing Applications (NCSA).
Although we have decided not to standardize the format of data within a record, many of the categories recommended for describing a resource within the CyberStacks(sm) scheme are identical or similar in function to the metadata elements of the Dublin Metadata Core Element Set (Dublin Core) described in the workshop report.
While the Dublin Core seeks to identify and isolate the data elements of an Internet resource, the summary information provided for a selected resource within CyberStacks(sm) is intended to describe the resource only to the extent that users are able to judge the potential value of that resource.
While we prefer to describe resources within CyberStacks(sm) more holistically than analytically in a default presentation, we understand the need and desire to identify and define core metadata elements. Although we do not believe that an exhaustive, analyzed set is necessary for inclusion within the resource profile within our scheme, we believe that a set such as the Dublin Core can provide an additional level of descriptive and subject cataloging for World Wide Web and other Internet resources that will facilitate their incorporation within emergent as well as conventional information storage and retrieval systems and services.
We believe that in the Age of HyperText, catalogers, and other librarians, need to consider more fully the 'presentation form' of a resource as well as formats needed for high-quality, consistent bibliographical control. We must look beyond the reconfiguration of conventional data elements as the sole means of representing an object to an alternative that includes additional data fields that seek to present the object as a whole, and not as a collection of encoded fields that users must decode, decipher or translate.
We do not advocate the abandonment of the MARC format for providing the necessary bibliographical control for Internet resources, but desire that it be enhanced and extended in a manner that could further facilitate control, access and presentation of Net resources. Data that is primarily intended for manipulation by the local or remote systems, or which primarily serves the needs or desires of catalogers, should be placed within a linked 'background' record, retrievable on demand, in a manner similar to the 'Long View' of many OPACs or the MARC alternative display option within the InterCat database. Information that facilitates successful identification and use of Net resources by users should be presented in an appropriate form as the new default option in a redesigned 'presentation' format. Within the hyper-environment of the Web, efforts to catalog Net resources must consider the Foreground as well as the Background.
One way in which these multi-faceted goals might be achieved would be through the conversion of conventional MARC tags into a parallel collection of HTML-MARC tags that would have the base functionality of conventional MARC tags and the hyper-functionality of HTML tags. As an illustration, we have marked the MARC record above (Fig. P-2) using a set of HTML-MARC tags (Fig P-4.):

<000>    cmm  Ia</000>
<001>    32671041</001>
<003>    OCoLC<</003>
<005>    19000000003138.0,</005> 
<008>    950619m19939999enkn     d        eng d</008>
<040>    UOK   $c UOK</040>
<050> 4  QA76.15 $b .H6 1994</050>
<082> 04 004/.03 $2 20</082>
<100> 1  Howe, Dennis.</100>
<245> 14 The free on-line dictionary of computing $h [computer file] /  
       $c Dennis Howe.</245>
<246> 3  FOLDOC</246>
<256>    Computer data.</256>
<260>    [England? : $b s.n.,]  $c c1993-</260>
<538>    System requirements: PC with modem.</538>
<538>    Mode of access: Internet World Wide Web. Address:,/:538>
<500>    Title from title screen.</500>
<500>    "This dictionary started in 1985 and now contains 8611
         definitions totalling 3.0 megabytes"--Modified ...
         May 18 ... 1995.</500>
<500>    Includes searchable online index.</500>
<520>    "FOLDOC is a searchable dictionary of acronyms, jargon,
         programming languages, tools, architecture, operating
         systems, networking, theory, conventions, standards,
         mathematics, telecomms, electronics, institutions,
         companies, projects, products, history, in fact anything
         to do with computing ... entries are cross-referenced to
         each other and to related resources elsewhere on the net
<650> 0  Computers $x Dictionaries.</650>
<650> 0  Microcomputers $x Dictionaries.</650>
<856> 7  $2 http  $u</856> 
Fig. P-4

With fuller data element delineation and a well-structured and defined Document Type Definition (DTD) within SGML, this HTML-enhanced MARC format may be the means by which the CyberStacks(sm) approach could be scaled to include all types of Reference works or all types of media. A more feasible option may be to adapt the USMARC DTD proposed by McDonough and the DTD for library, museum, and archival finding aids proposed by Pitti (Pitti, 1995), both of the University of California, Berkeley.
Pitti, Daniel V. "Standard Generalized Markup Language and the Transformation of Cataloging." Serials Librarian 25, nos. 3-4 (1995): 243-253.
Any or all of these approaches have the potential of offering not only enhanced presentation of Net resources within conventional Web browsers, but, with the extension of MARC and related formats, provide the basis for an integrated database and collection of Internet and non-Internet resources accessible from within the next-generation OPAC.