Since my latest post I have been busy distilling a set of principles that will serve as the foundation for media management. SAMSN stands for Strongly Attached Metadata with Semantic Namespaces. OpenSAMSN is an organization to raise awareness of the principles and build a community of practice.
SAMSN is about establishing a foundational platform for enabling the media management industry to get to the next level of efficiency, interoperability and in general adding overall value to the integrating system(s). As I wrote in the my previous posting, I believe there is an opportunity for optimizing digital supply chains through the use of metadata in a similar fashion to how the barcode revolutionized the retail and physical supply chains.
SAMSN is about establishing that common touch point, in the same way that the barcode established that common touch point across the various systems in the retail industry. A critical requirement of this foundation is that it should be open and accessible to all participants in the industry. It should also be a standard in the sense that it will not significantly change over a number of years. After all, the UPC barcode hasn’t changed in over 40 years – yes there has been improvements with QR codes, RFID, etc. In all that time, the touch point has remained the same but the systems that integrate with it have evolved – that is what the industry needs to move forward.
Strongly Attached Metadata is about ensuring that the metadata can easily travel with the file. XMP is the best candidate since it provides a common specification for how to embed and describe metadata within a large number of formats. But it doesn’t cover them all. Proprietary formats such as MS Office don’t preserve XMP metadata, rather they have their own way of storing metadata. That’s ok – as long as there can be an open, publicly documented way in which other systems can take advantage of that format. Additionally, the OS provides ways in which the metadata can be stored as alternative file streams or resource forks. This can be useful in general, but specifically for formats such as plain text files. The challenge is in preserving those files streams as the files are emailed, FTPed and pass through unfriendly legacy systems. Raising awareness of this data loss can help to plug the holes and ensure the metadata remains strongly attached.
Semantic Namespaces is about recognizing that every company can and will eventually create their own namespace and property definitions. This will naturally arise because of the business requirements to get their own media management house in order. So it will be critical to chose a metadata format that can support namespaces and respect their existence without deleting them. Since everyone will have their own internal way of describing the world, there needs to be a way to connect the meanings of the property descriptions. Efficiencies will be realized when metadata can flow between systems through the integration of the data itself. Making this connection can be as simple as defining mapping tables or as complex as defining industry ontologies. Hence the eventual need for semantic technologies to help out. Good news is that XMP is also a good choice since it is based on semantic technologies such as RDF.
I’ve established OpenSAMSN as a way to raise awareness of these principles and bringing together industry leaders that are interested in collaborating to make this happen. Enough technology already exists, what is requires is human collaboration and agreement to make this work. After all, the barcode was patented in 1952. The first real retail implementation didn’t occur until 1974, after the industry got together to agree how to use the technology. We shouldn’t have to wait 20 years.