Let’s go back to that metadata snowball I hinted at a few posts ago.
In my previous experiences, I have been involved in efforts to centralize the capture of metadata in one place to be a resource for everyone to use. The diverse aspects of the entire metadata record aren’t entered at once. Instead, it all starts at the beginning of the processing workflow as a small core set of metadata about the asset…ideally this has been captured at the creation source or shortly after that or by parties connected to the origination as to be able to supply the most relevant information. This core record should never need to be re-entered by other groups anywhere else along the workflow into the same data repository and any other database. Instead, this becomes the core of the metadata snowball that rolls down thru workflow gathering additional metadata such as rights, search cataloguing, production and distribution info, etc.
This snowball effect of starting with a small amount of information and building a large, rich (and accurate) data record is an efficient way to reduce inaccuracies and inconsistencies. The days of different departments keeping different versions of the same data in separate places has to end. There should be one pool…one resource…always.
With the snowball, I don’t need to know everything about something to start as long as know the basic information then we can add to the record and expand on the information to completeness later. This adding of information compartmentalizes the process and reduces the organizational burden to the groups involved. There responsibility in this data assembly line is to be the masters of their area of expertise and apply the additional information to the metadata moving thru their part of the assembly line not to manage redundant synchronization of information as it passes thru their world.
But what about user generated content? How could this model be used with the same standardization and controls in the metadata? It seems to me that the upload process must include a required data set from the user that is including the same controls for internal processing. For example, sports leagues have tightly defined structures and metadata sets..players, teams, leagues, various stats, etc. These pre-defined variables can be preprogrammed into the controlled lists and be part of the required data for the UGC upload.
The challenge here is to tap into the power of information that can be gained from this efficiency and from the knowledge base from the UG stream….we need to create tools to make this easy. Don’t build a snow fort and dodge the snowballs, set them loose on a hillside and see what you get.
Comments