Logo GISdevelopment.net

GISdevelopment > Proceedings > GITA > 1998


GITA 2002 | GITA 2001 | GITA 2000 | GITA 1999 | GITA 1998 | GITA 1997 
Sessions

Application

Data Distribution

Data Evolution

Field Applications

Integration of the Enterprise

Invited Presentation

People Issues

Scada and Real-Time systems

System Development

User Presentations

User Solution


GITA 1998


Data Evolution


Data Conversion - Many Souces, One Target


Sources - Svmbolo~v and Annotation
Up to this point, this could apply to converting any type of data set, Map conversion has the special consideration of graphics. What do they look like today? What should they look like when in the target system?

Documenting symbology and annotation of existing sources should be done as part of the source documentation. This part of the documentation is particularly crucial for paper sources. All symbology variations need to be defined. If conversion has many sources, chances are there are many symbol and annotation variations between them (the source documentation would verify that). Some Geographic Information Systems may allow for multiple symbol and annotation variations for same item. Hand drawn paper maps almost insure this will be the case. A conversion project is the perfect opportunity to set common standards.

Symbology and annotation for the new system needs to be documented. Many of these can carry over from the source documentation, but as conflicts arise how can one be selected over another. Adopting industry standards (ANSI, IEEE) can speed up the process of choosing the graphical representation of the data. This is a very subjective activity and everyone has an opinion. While this is started with the source documentation it cannot be completed until the data model is finished.

Sources - Key Identifiers
Each item needs at least one attribute that can uniquely identify it from all other objects. Some may need to rely on two or more attributes, but the fewer the better. It is important when selecting key attributes to select those with the most positive results and chance for fewest errors. As an example, a symbol that that is printed, viewed, reviewed and recognized by many, is preferable to a level designation, which maybe accessed by only a few people. This is even more critical if the source application allows these two to be separately entered and maintained.

When these are determined, any item not conforming to the standards specification must be scrubbed before conversion. Document the key identifiers when the sources are documented. This portion of scrub can be done once they are set and must be done before conversion begins.

TarRet Data Model
The second most important piece of the data conversion puzzle is defining the data model. This is by no means trivial, but since there are many others who can and will cover this in better detail, here it is only stressed how important it is in the data conversion process. The data model is best designed within the confines of a specific project and must be completed before conversion can begin.

Page 3 of 7
| Previous | Next |

Applications | Technology | Policy | History | News | Tenders | Events | Interviews | Career | Companies | Country Pages | Books | Publications | Education | Glossary | Tutorials | Downloads | Site Map | Subscribe | GIS@development Magazine | Updates | Guest Book