MultiValuedDatabases

The Pick Multi-Value database has been around since the late 60's. The History of Multi-Value is interesting, diverse and also full of Colorful Characters. Pick and other Multi-Value databases have numerous monikers; Multi-Value, MV and Multi-dimensional are just a couple of them.

Wikipedia has a good overview article for those new to MV: WikiPedia

The following characteristics are common to most MV systems:

• A HashingFileSystem (as a minimum - newer systems have many different types of files)
• Dictionaries that describe the record layout, but do not constrain the data
• Records with fixed primary keys, delimited by AttributeMarks
• Records can be extended on the fly
• MultiValued fields within the records, traditionally seperated by ValueMarks
• A Basic programming language that understands files, dictionaries and records
• A query language, that deals with multivalued data structures

MV DBMS products traditionally run as a Virtual Machine over an operating system. In addition to other connectivity methods, many users can connect to these environments using Terminal Emulators like Accuterm, Winnix or Wintegrate. These terminal emulators can host graphical interfaces and provide other capabilities that integrate server and client.

For those who are interested in some of the more technical details of Pick and the Pick Data Structure should review some of the links found at the bottom of this page. There is a general Philosophy of Pick for those who just want some more general info.

A concerted effort is now taking place to create a formal definition of MV, led primarily by members of U2UG and the Oliver mailing lists.

A partial list of commercially available MV systems:

• Caché
• D3
• jBASE
• mvBase
• OpenInsight
• OpenQM / QM
• Reality
• UniData
• UniVerse
• UniVision

A partial list of Open Source and/or freebie MV systems

• Bart
• MaVerick
• OpenQM
• Winter

A very partial list of extinct (or nearly so) MV systems:

• ADDS Mentor
• Advanced Pick
• Advanced Revelation
• Altos
• C-Itoh
• CDI ??
• Climax
• Datamedia
• Fujitsu
• ICL Clan
• IN2
• mvEnterprise (nearly so)
• Siemens-Nixdorf
• Pertec
• Power95
• Prime Information
• R83
• Sanyo/ICON
• TAU ??
• Ultimate
• UltPlus
• Wicat

Other Notes

For a comparison of MV Databases, there is a fine paper by MIS Modular Information Systems @ http://www.utwo.co.za/MultiValueReviewWhitePaper.pdf
former link to CDBMA no longer valid.

A very good basic manual of PICK/BASIC can be found at: Jonathan E. Sisk's "Pick/BASIC: A Programmer's Guide": January, 2000 WWW Edition http://www.jes.com/pb/index.html

For an introduction to using the Pick Query Language, see "Getting Started in OpenQM" from http://www.rushflat.co.nz/downloads.shtml

IBM's white paper on comparing MV Architecture vs: "Normalized" DB structures can be found @: ftp://ftp.software.ibm.com/software/data/u2/pubs/whitepapers/nested_rdbms.pdf
former link to CDBMA no longer valid.

Search on Key words "Nested Relational Databases" pulls up a significant number of sites with good information for those interested in the subject.

A basic intro to the MV DBMS Model for someone more familiar with relational: http://www.nebula-rnd.com/products/mvdbms.htm

Notes / Discussion

• [TG] Would be nice if someone helped owner of this site to correct his MV-related links.
• [TG] Can we host the linked PDF files here so that they don't disappear anymore?