Open csjx opened 6 years ago
OK, so I only came accross this now, because we are needing to add platform to our metadata in EML. A platform is not really an instrument, but rather a place where instruments are attached to or launched from (a ship, a buoy, etc). I would really like there to be a field platform in methods, in addition to the currently available instrument. Platforms are much used in marine biology.
If you are trying to tie your methods to particular instances of Platforms or Instruments, another approach to creating a new set of fields would be to use the dataset or entity annotation feature in EML 2.2 to link to specific instrument and platform vocabularies. Would there be an advantage of using a text field over structured annotations?
Unfortunately, the destination of these data (EurOBIS, GBIF, ...) are still tied to EML 2.1.1 (and unless they recieve a push from publishers, I am told, they will not update to 2.2). We are working on updating ourselves to 2.2, so it would be great if you could give me an example of what you mean for EML 2.2. But it does not answer my problem for 2.1 which we still need to keep providing. Using structures annotaions is, indeed, better than using a text field (as long as it is clear to the human and the machine that we are talking about a platform). (FYI I realise that 2.1 cannot be updated any more, so for that part of my work I am just trying to find the best place to add platform, as well as a few other "missing" fields)
ISO 19115 has a concept of a
platform
that is defined as:and a
sensor
defined as:In EML we have a similar concept in
eml-methods
ofmethods/methodStep/instrumentation
:We've received feedback about storing platform and sensor information in a way that is translatable to ISO 19115. This would mean accommodating a hierarchy of
platform --> instrumentation
information. Looking at the ISO 19115 spec, there isn't aplatform
field per se, but aDS_Platform
is a subclass of the abstractDS_Aggregate
type, and as an aggregation, theDS_Platform
may have0..n
DS_Sensor
descriptions. What's odd to me is that the subclass design ofDS_Aggregate
can technically make for some much less intuitive structures like aDS_Sensor
with0..n
DS_Platform
sections. Such is the nature of UML - looks good until you try it for real. Here's the model from the spec: So, we'd need to discuss how this would work in a backward-compatible way for EML, if at all possible. A platform in ISO 19115 is baked into the aggregation structure, whereas EML uses a free form text field to describe this information.