Transcript A brief history of computing databases

+
A brief history of data and
databases
Spanning thousands and thousands of years
Unattributed pictures from University of Rochester History Site
+
Record Keeping – How long?
•
The earliest “databases” (organized collections of
information) may date as early as 3300 BCE (BC) with the
beginning of writing (Cuneiform).
•
Used to keep accounts and other record keeping.
•
There is evidence that other forms of record keeping (clay
tokens, notched bones) were also used prior to this.
Source: http://www.ancientscripts.com/cuneiform.html
+
So what is a database?
+
Why?

We use records to measure “stuff”.

And most of these records are not digital.
+
Today we use search engines,
but…
•
Paper records had to be cataloged in ways that made data
update and retrieval possible.
•
Census worked in precincts, within precinct by address. If
you wanted to find one person, you had to know where they
lived. When was the first US census?
–
But of course that was not the intent of the census
•
Marriage and burial records were in books by year spans.
Pages were devoted to last names in ranges.
•
Paper indexes might be used to cross reference other ways
of searching.
+
Card Catalogs
•
Interesting way to record the library collection.
•
Each document (book, journal, artifact, …) was cataloged on
a number of dimensions. (subject, author, title, record
number (dewey decimal system)
•
Cards were hand typed and hand filed.
•
Each catalog provided an index (or a pointer) to where the
book was located.
•
It didn’t indicate whether the book was available, just where
it should be found.(example catalog, example cards)
+
Other storage devices

The punched card

The first storage device of early computing

Roots in textiles in the early 1800’s

Jacquard Loom created cards with holes. The hole let a rod
through. A set of cards defined the pattern. Binary! Either a rod
went through or it didn’t. Loom

Basis for the first computing data storage.
+
Problem

As the US Census Bureau approached the 1890 census it had
a problem.

1880 census was concluded in 1887 and the population just
kept growing.

Needed to find a way to increase the efficiency of the census
processing.

Enter Herman Hollerith.
+
Hollerith’s device
pantograph
Hollerith card
Integrating
machine
+
First computers
The program, the data, the JCL – all done with punchcards
+
Electronic files
•
Originally, files were associated with particular applications
on particular computers.
•
Different kinds of data were in these different files.
•
Integration of data was very difficult.
•
Different uses of the data required program change.
•
File storage was expensive and limited.
•
File storage was primarily sequential (slow access)
+
Enter the database
•
Objects in a database can be related to one another.
•
Hierarchical – One record leads the the related record. (Like
a tree)
•
Network – Allowed for multiple relationships (like a network)
•
The databases used pointers to relate one record to another.
•
1960’s through today
+
1970, Edgar Codd
Relational DBMS
• Mathematician at IBM
• Based on Relational Calculus
U of M
• MicroDBMS
IBM
• System R
• First implementation of SQL
Led to
• Oracle
• IBM DB2
•Informix
• Sybase
• MS SQL Server (based on Sybase)
+
Relational Ideas

See prior day’s lecture
+
The Future?
•
Object Oriented Databases
–
–
–
–
–
•
Combine data and operations on those data
Allows for inheritance
Cache’ (Intersystems Corp)
Oracle (Object-Relational Database)
Postgre(open source)
XML and XML DBMS
–
–
XML designed to transport and store data (w3schools.com)
Database manages that data