Re: [apps-discuss] SenML

[Zach Shelby <[e-mail]>]

On Mar 14, 2012, at 11:50 AM, Kaebisch, Sebastian wrote:

>>> 
>>> 3) use restriction definitions of strings (define maxLength or list all
>> possible values as enumeration). Especially using enumerations results in
>> a smaller message size and avoids string comparison.
>> 
>> Enumeration is good idea but I think strict enumeration allows no
>> addition in future without breaking backword compatibility. I recommend
>> some 'dummy entries' to make room (and show the limit to keep the same
>> bitwidth clearly) for future.
>> 
> 
> If there a need of flexibility, then we should do that. In my understanding, the unit symbol table in 10.1 will be standardized and will be fixed then. 

The unit representation part of the SenML draft is the last thing that is an open issue to fix actually. The value will need to be open-ended as even in the current form this would need to be an IANA registry and thus new values would be added over time. However, we are looking at a model where external unit references can be made instead of trying to standardize a list of units in the IETF. For now at least we should leave this as a string. Let's look at it again when the unit model has been updated in the draft.

Zach

-- 
Zach Shelby, Chief Nerd, Sensinode Ltd.
http://www.sensinode.com
http://zachshelby.org  - My blog "On the Internet of Things"
http://6lowpan.net - My book "6LoWPAN: The Wireless Embedded Internet"
Mobile: [phone]