Extracting Metadata - Retrieving Metadata from External Sources
(Page 5 of 22 )
The sources you’ll retrieve metadata from may be quite varied and include sources such as mainframes and Excel. Think broadly about metadata. What are you working with? What metadata information do you need, and how can you get it? Are you working with laboratory information, process control, or home automation with specific types of sensors and controls you can describe with metadata? Do you need to mimic the structure of mainframe data, either for synchronization or for an initial import of your data? Can you grab the structure and build CREATE stored procedures as well as import and synchronization processes? How can you capture this information?
TIP Keep your mind open to incorporating other metadata sources as you run across them. You can use any predictably formatted information as a metadata source.
Remember that metadata doesn’t include information that might change at runtime because you won’t regenerate to incorporate changes at runtime. (See Footnote 4.) In general, metadata will not include specific network layout information, client configuration information, and so on. If these are part of your application, your metadata might include the structure that holds such details, but not the details themselves. This distinction may be easier to understand if you consider a simple home automation example. Information about individual types of sensors and controls are valid metadata. Each type of sensor or control has a logical set of data, such as a thermometer having a property that can contain its location and another property that can contain the current temperature. The temperature isn’t metadata, and in most cases the specific location isn’t metadata. The fact that a thermometer is a type of sensor with a current temperature and location is metadata.
(Footnote 4. Strictly speaking, you can, but building and compiling source code at runtime is rarely justified.)
Once you understand what metadata you’ve got and where it currently resides, you can design a freeform XML block. Based on that schema, you can write an XSLT to organize the metadata if it’s already in an XML format or write a program to extract it. How hard that program might be depends on how messy the source’s interface is.
 This is from Code Generation in Microsoft .NET, by Kathleen Dollard (Apress, ISBN 1590591372). Check it out at your favorite bookstore today. Buy this book now. |
Next: Extracting Metadata from Design Tools Such As UML >>
More Database Articles
More By Apress Publishing
/ 18 
