Newsletter 4 - November 1997

               _____   __________________  ____  ____ 
              /  _/ | / /_  __/ ____/ __ \/ __ \/ __ \
              / //  |/ / / / / __/ / /_/ / / / / /_/ /
            _/ // /|  / / / / /___/ _, _/ /_/ / ____/ 
           /___/_/ |_/ /_/ /_____/_/ |_|\____/_/      
                                                       
               _   __                  __     __  __           
              / | / /__ _      _______/ /__  / /_/ /____  _____
             /  |/ / _ \ | /| / / ___/ / _ \/ __/ __/ _ \/ ___/
            / /|  /  __/ |/ |/ (__  ) /  __/ /_/ /_/  __/ /    
           /_/ |_/\___/|__/|__/____/_/\___/\__/\__/\___/_/     
**********************************************************************
Number 4     ***         INTEROP NEWSLETTER      ***     November 1997
**********************************************************************

                 DIF Syntax Specification for DIF V6
                 -----------------------------------
Review of the Directory Interchange Format Formal Syntax Specification
is needed.  The Specification is in its second draft and includes all
fields which have been accepted by the voting committee.  The
Specification is available online at:
http://gcmddev.stx.com/wwwdev/dif_syntax_spec_d2.0.html 

Please send any comments or questions to mddif@gcmd.gsfc.nasa.gov.
Questions or comments should also be posted to the INTEROP mailing
list (interop@gcmd.gsfc.nasa.gov) where appropriate so as to inform
the entire DIF user community.

(Note: those fields which are still under consideration by the voting
committee have not been added to the Specification.  A message will be
sent to INTEROP upon their acceptance or rejection and the
Specification will be modified accordingly, if necessary.)



                      CEOS IDN Task Team Minutes
                      --------------------------

    CEOS IDN Task Team Minutes - Stresa, Italy, 22 September 1997

On 22 September 1997, the IDN Task Team met in Stresa, Italy in the
Sala VIP at 19:00.  The meeting began with a review of the minutes
from the IDN meeting in Toulouse.  The attendees included:

Lola Olsen              NASA
Terry Fisher            CCRS
Gunter Schreier         DLR
Hiroshi Ishiguro        NASDA/RESTEC
George Saxton           NOAA/NESDIS
Ken McDonald            NASA/GSFC
Yonsook Enloe           NASA/GSFC
Wyn Cudlip              BNSC/DERA
Richard Gobel           DLR/CEO
Dirk Van Gulik          CEO
Brian Thomas            BNSC/EOS
Christian Hoffman       CEO
Holger Hoff             IGBP-BAHC
Brian McLeod            CLR
Claude Huc              CNES
Steve Foley             DERA/BNSC
Mark Nestler            Hughes STX
Lynn Halpern            Hughes STX
Osamu Ochiai            NASDA
Eiichi Sakata           NASDA/RESTEC
M. Cristine Falvella    ASI
Einer Groves            NSC
Ben Burford             RESTEC
Richard Jones           BOM
 

        The minutes from the Toulouse meeting were distilled.

At previous IDN meetings, participants expressd the need for more
formal version control and better documentation of the DIF for both
system administrators and DIF authors.  All documents (& SW) are now
under strict version control. Additional documentation to meet
extended interoperability requirements is also now available.

                          III. Documentation

A. Documentation for DIF Authors:

*Write-A-DIF: quick-look, 1-page (2-sided) listing and description of
DIF fields

*DIFguide: detailed description of fields and valids for controlled
keywords; available on the WWW (http://gcmd.nasa.gov/difguide)

                            *DIF Templates

B. Documentation for creating DIF-compatible software or database:

*Formal Syntax Specification details the type, length, number of
occurrences, etc. of every DIF field. This document is available on
the WWW (http://oxygen.stx.com/wwdev/dif_syntax_spec_d2.0.html).

C. Other:

*DEDSL - Description in the Data Entity Description Specification
Language Wyn Cudlip stated that an attempt by DERA with the CNES OASIS
software failed to machine-read the IDN DEDSLed file. They are
investigating the cause of failure. This document is available on the
WWW (http://gcmd.nasa.gov/software_docs/DIF_DEDSL_V1.0.html)

*DIF Doctype - for z39.50 searching with Isite. Available on the WWW
along with the DIF content files. See http://gcmd.nasa.gov/ceosidn and
select "Download the IDN Content".


                        IV. CEOS  IDN Outreach

T. Fisher provided 3000+ bookmarks and encouraged all to take and
distribute.  Several packets were sent for CEOS Plenary in Toulouse.


                    V. IDN Content and MD5 Status

A graph was provided of the current number of DIFs on a month/year
vs. total number of DIFs plot from Aug. '93 through Aug. '97. Current
number is nearly 5000. New population strategies continue to be
addressed.

Gunter Schreier asked if the statistics were available for DIF
updates. It was noted that these statistics are tracked and will be
plotted on a graph.

The suggestion was made to redirect users that come to GCMD to an IDN
node closer to home and also possibly from nodes with outdated content
(more than one month) to those with current content.

An inquiry was made as to whether there are commercial DIFs in the
IDN. There are commercial DIFs written, but they must be written and
quality-controlled by the data provider to minimize any effort by
public-funded staff.


Several ideas were expressed that were thught to be helpful in
increasing content and maintaining node currency. The suggestion that
nodes that do not contribute content in six months be taken off line
was unpopular. Reasons included the possibility that nodes might
withhold available data set descriptions to maintain the requisite
flow of DIFs or that nodes might indeed have exhausted descriptions of
its holdings. The suggestion will be withdrawn.

Although participants agreed in principle with the idea that nodes
should not offer outdated content, several attendees suggested that
other methods be devised to disregard outdated content through
automatic rerouting. It was also suggested that other steps be taken
to facilitate (ease) the update (synchronization) procedure by the
nodes, but minimally by announcing the availability of updated content
every month.
 
                                 VI.

The revised Aggregation Proposal was described. George Saxton
emphasized the importance of implementation of the Parent_DIF field.
This was discussed further at IDN Splinter 2. This proposal received a
generally positive response.

New location keywords and search scheme were described. Christian
Hoffman commented that they had purposely avoided using location
keywords based on political boundaries because of the inherent
instability of such terms.

The Related_URL proposal was deferred for further discussion until IDN
Splinter2. Forum participants did not endorse this proposal in its
initial form. Early comments after its revision continue to question
the implementation of this field. Therefore, its implementation has
been deferred until it becomes clear that it is needed and its
explicit purpose is presented with utmost clarity for DIF writers.


                    IX. A.  Valid Synchronization

Work on valids continues and has been extended to many new
fields. Suggested keywords are being offered for the identified fields
through the DIFguide. An attempt to define the level of effort for
identifying keywords was originally planned to be determined based on
the most often-queried fields (from the metrics), although it was
unclear that the sequence of fields as presented to the user was not
instrumental in the frequency of queries. However, Christian Hoffmann
confirmed that users seemed to prefer broad categorization-breakdowns
(based on the CEO experience even after they moved a comparable search
term to the bottom) and that he believed IDN users would continue to
seek these fields for their queries, independent of position as
presented in the interface. It was noted that as the number of
database entries increases, the searches will have to be done on
increasingly narrowing terms. This led to a discussion of the
importance of valids for bucketing as can be demonstrated in the
prototype dynamic query interface.
 
The prototype dynamic query and refinement was illustrated through
several slides.  The response seemed favorable.  The integration of
this interface for future versions will be discussed further.

                   IX. B.  Content Synchronization
 
With eleven nodes now actively mirroring the IDN, work must progress
on simplifying synchronization and moving toward a distributed system.

A second splinter session was scheduled for the following evening to
hold a more in depth discussion of the Aggregation and URL proposals.