Question 5
What is different about a database?
GIS has only partially come to terms with database management.
Lots of database management buzzwords get thrown around in marketing,
so it is important to tell when they are being used properly.
You should be able to use the follow terms correctly:
- Schema
- Foreign key
- Join
- distributed database (alternatives: federated, centralized,
etc.)
You should have a sense of what distinctions people are making
when they distinguish:
- a file system from a database
- "normalized" relations (tables, tuples, or whatever)
In response to this question, pick at least one
of the options below:
- How do you merge information from two different sources if
they have adopted two different schemas (hint: you might want
to read Tim Nyerges' paper on schema integration ask me for a
copy)
- What makes transactions more complex in a spatial database
(compared to say a ticket counter)? What kinds of activity would
you have to consider? How would you manage the situation where
two users wanted to edit the same map area?
- How far can you go with a distributed database? Discuss the
tradeoff between performance (speed) and integrity (one known
master copy)?
- Tomlin's map algebra, Arc/INFO procedures (and their predecessor
in ODYSSEY) relied on transformations that made copies of the
whole database. How could this be replaced in a database situation?
Can you allow cumulative transformations? What about the other
users?
Your answers shouldn't be much longer than one page.
Index from here: Back to 465
Main Page | Questions / Assignments
| Glossary from Exploring
GIS | Schedule | Bibliography
| Course Policies | Resources
Version of 7 February 2001