From: Roebuck, Roy [RoebuckR@ncr.disa.mil] Sent: Tuesday, September 21, 1999 3:12 PM To: Diaz, Cris; Chris Burritt (E-mail); Juan at SAIC; Randy Wingett at SAIC; Bush, Cynthia; Jacobs, Chris; Richard Bartoletti (E-mail); Heather Newell (E-mail) Cc: Larry Smith; James Fesko (E-mail); Matthew Howell (E-mail) Subject: Directory, Data, and XML Junction

Importance: Low

Hi:

The Data Junction products described in the attached HTML file and links could have immediate benefit to the DII community if included as part of either the COE or as part of the Mission System Developer's recommended "toolkit". Note that the products allow full two way interchange between hundreds of structured (XML Junction, Data Junction), semistructured (XML Junction and Cambio), and unstructured (Cambio) data stores. (See postscript.)

When I was the D3 IMO in 93 through 95, and was serving as the DISA HQ Notes Administrator and supervising a contractor team developing Notes applications for the DISA CIO, I used two main products for transforming, cleaning, harmonizing, and interchanging data - Data Junction, for a few hundred dollars at the time, and another data synchronization tool (Cashal ReplicAction at $10,000 - now unused by DISA) using Notes replication technology. I found the Data Junction tool, at that time, to have the best ad-hoc transformation capability, while the $10K tool was better for scheduled replications. It looks like Data Junction has moved itself into much of the ReplicAction territory for a fraction of the cost. I'd say that if you combined Data Junction (with LDAP and XML Junction functions) and Cambio to map interfaces and provide the transport mechanism for data interchange, and XML-Authority from Extensibility (http://www.extensibility.com) for mapping and translating XML data and document schema, you'd have a fairly complete mechanism to take down the barriers between DoD system and operational data and documents.

Also, with the LDAP junction, COE could provide a mechanism outside of metadirectories and APM/CDS to harmonize directory/hiearchical-namespace data.

Roy

P.S. My term "structured data" is SQL, ISAM, XML, LDAP, and Object DBMS, deeply tagged documents, and data stores like indeces, categorized indeces (index trees like Netscape Compass, Oracle Context, or Verity Topic) where both the containers and content are highly addressable. Semistructured data is Lotus Domino, slightly tagged documents, threaded discussions/bulletin-boards/newsgroups, pattern-indexed video and images, and voice-recognition processed audio files (meetings, speeches, etc.) where the container is highly addressable (rich metadata) but the content is largely unaddressable. Unstructured data is unstyled and untagged documents, and raw video, images, and audio where the container is addressable (little metadata), but the content is not addressable.

http://www.datajunction.com/news/reviews.html

http://www.datajunction.com/news/announcements_directoryjunction_091499_prin t.htm

http://www.datajunction.com/news/announcements_xmljunction_61599_print.htm

<<Data Junction, Inc. - data warehouses, data marts - News.htm>>