Enter Semantics: Please Let Me Tag!
Tags describe an item and allows it to be found again by browsing or searching. Tagging is one of the most obvious things we can do:
- We insert metadata in HTML pages as hints for search engines.
- A lot of blogging and microblogging platforms allow tagging – think of hashtags an Twitter.
Besides tags are other useful metadata:
- Our cameras insert geo locations, date, time and even technical information like color depth, the image resolution in pictures so we can govern them.
- So same is true for audio recording: We can insert metadata into MP3 files like the creator of the record, comments and much more.
Metadata Assures Mobility
A lot of document management systems can use tags and other metadata for searching and retrieving. But there is another important reason: metadata assures mobility: When publishing media data on platforms like Flickr or Facebook often I want to publish metadata, too. When I transfer the data on another platform or to an a media player I need those metadata, too.
Fortunately the are standards like XMP (http://www.adobe.com/products/xmp/) that allow to insert metadata even into binary documents. In fact those metadata is RDF in XML syntax like Dublin Core http://dublincore.org/documents/dc-xml/ for example. The Dublin Core set provides a small and fundamental group of text elements through which media data can be described and catalogued using 15 base text fields.
Document Processes in Enterprises
Those metadata are very useful for document processes in enterprises. A lot of companies have complicated document processes starting from SAP systems. Those processes utilize many external systems like
- output management systems,
- form systems,
- print routes,
- workflow systems for controls (four-eyes principle),
- document management systems,
- batch archiving systems,
- post processing of documents,
- email servers.
Those document processes are much sophisticated and differ from printing process using Adobe Document Services: The SAP system doesn’t give out PDF but creates XML documents using standards like XSF or XFP (for Adobe Print Forms). To control those processes metadata are necessary – these are for example
- processing information for external systems like output management or documents management systems or workflow systems and
- additional IDs to for end to end control.
But there much more use cases of metadata that described in the following use cases: http://www.adobe.com/products/xmp/partners.html.
To implement complex document processes two things are necessary:
- standards like XSF or XFP have to support metadata
- APIs like form technologies like Smart Forms or Adobe Print Forms that create XML output should be able to insert metadata to the output stream.
XML without RFD is like a Pulled Handbrake
Unfortunately both conditions are not fulfilled by above mentioned standards so we have to encode those data within XSF and XFP syntax. This is unsatisfactory for many reasons:
- We have to include those metadata within the business data.
- As a consequence we can’t use standardized vocabularies defined by SAP or vendors of external systems.
For me this is hard to understand: SAP already uses XML based formats for document processes. Enhancing output technologies by embedding metadata as an XML element in a special namespace is very easy. W3C already defined with RDF an flexible and extensible metadata format. Now we only have to put everything together to achieve state-of-the interface technologies.
In heterogeneous and data driven enterprise architectures metadata is absolutely necessary for interoperability. This is well known (this of B2B EDI standards defining header elements containing meta information) and RDF allows to defines semantics units in a very generic way. In my opinion those techniques should be adopted by standards architects and introduced into SAP Business Suite.