E-mail
Attending: Joel Farrell, Chair; James Fiore, Steve Kinney, Carl Singer, Valerie Smothers, Tarang Shah
Recent MedBiquitous News
Valerie commented that not much had happened in the Virtual Patient working group with regard to developing a roadmap for future work. Nabil had offered to circulate something when his vacation finished in August.
Valerie then described a new Educational Achievement project underway. The work is intended to document achievement of competencies or milestones. An example of a milestone is something like “Able to perform a newborn exam with no supervision.” Joel emphasized the need for this work to fit in with related standards development, including our competencies work.
Attachments
Joel summarized that those who exchange professional profile and competenciesdata needs binary data like pdf to support and document credentials or competencies. We looked at 3 ways this could happen:
1. Web services (use MTOM) Seems like organizations in need of attachment support are not web services enabled. They will need a different approach.
2. Atom archive tag, with foreign coding. This is fine for small binary objects. For large ones, it is inefficient. In addition there are performance and storage implications.
3. URL references, either outside the firewall or at to some outside filesharing site. This would be easier for many members. They could make files available even through FTP, or through a mechanism that can be read in through web browser.
There is a valid use case for each. We need to do a guidelines document that explains what to use in what case.
James commented that ECFMG is using zip files right now and base 64 encoding. Size was a concern for them. They have scanned documents: images, not so much text.
Joel proposed putting the following in a guideline: if its part of web services, use MTOM. We don’t have standard web services. As a general guideline, if you are defining web services that include binary content, MTOM is the way to go.
If you are trying to include binary in existing XML standards, for most cases, URL reference to the binary object is the preferred method. If the document is small enough (we would need to define), you could also encode in text value of XML elements.
Valerie added that there is currently no place to put url references in the XML. Where would a scanned diploma reference go?
James asked if we should define a separate element that could indicate whether the data is uuencoded or referenced.
Joel commented it should be a separate element. Atom has the archive tag in case you want to uuencode. It has a link element, too, but not specifically for binary things. We could make an element, doesn’t have to be same name all the time, unless we want to go all out.
James commented that adding a specific element may not cover our needs. We need 0 to many occurrences, a unique id and description, type of attachment and field for attachment. This would be similar to the way activity report documents continuing education activities. There is a Credit id, number of credits, unique id from domain issuing, activity name, module name, status. If you add an element for a scanned diploma, they will have a whole history of medical student milestones.
Joel commented the tag will have to be in the context of other elements that describe what it is. It would be up to the particular standard to determine where attachments can go.
James summarized that the types of attachments could be uuencode, base 64, or url. The element would contain a resource locator or encoded document. We could have some additional data to describe what it is, including attachment type. Perhaps mimetype.
Joel commented that for a diploma tag, within that is a url tag, within that what kind of file; we could have a link tag and attachment tag. The Link tag or inline tag/archive tag would need an attribute indicating what kind of encoding so that recipients know how to reconstitute the binary object.
Valerie commented that it sounded similar to atom.
Joel questioned how much do we want to define up front and how much to leave to the individual standard. Valerie commented that working groups would be grateful for a consistent approach they can drop in.
Joel added there are some examples where there is a set of binary data. Where is that additional metadata kept? Making a statement would make a legitimate guidelines document.
- Valerie will draft guideline for the next call.
- The TSC will come up with examples separately. We’ll look at samples next time and run them by the AAMC and ECFMG to see if this makes sense, or if they are looking for additional guidance.