Re: [core] Proposed new SenML Units

[Carsten Bormann <[e-mail]>]

Hi Jaime,

> On 2020-11-09, at 15:41, Jaime Jiménez <[e-mail]> wrote:
> 
> 
> Dear CoRE and SeNML experts,
> 
> the IPSO WG has some new objects being registered that contain units
> that are not currently present in the IANA SenML registry (
> https://www.iana.org/assignments/senml/senml.xhtml ) nor derived
> from it. We would like to discuss the addition of the units below:
> 
> +===========+================================+=======+
> | Symbol    | Description                    | Type  |
> +===========+================================+=======+
> | NTU       | Nephelometric Turbidity Unit   | Float |
> +-----------+--------------------------------+-------+
> | mg/l      | milligrams per liter           | Float |
> +-----------+--------------------------------+-------+


> NTU is the most common way to measure turbidity.

This one is interesting.  There are several Turbidity units, which are not in a simple linear relationship, so I would like to get some input from domain experts which ones we should have (and which are more of a historical interest).

> mg/l is used to measure
> concentration, possibly g/l should also be added as shown in the table
> below. 

We already have    
kg/m3    kilogram per cubic meter (mass density, mass concentration) float [RFC8798]

One mg/l is 1/1000 kg/m3, so that would go into table 2.


> On the secondary units table we would have the units below. 
> +=========+======================+======+==========+======+
> |Secondary| Description          |SenML |   Scale  |Offset| | Unit    |                      |Unit  |          |      |
> +=========+======================+======+==========+======+
> | ppb     | parts per billion    | /    |   1e-9   |    0 |
> +---------+----------------------+------+----------+------+
> | ppt     | parts per trillion   | /    |   1e-12  |    0 |
> +---------+----------------------+------+----------+------+
> | KB      | Kilobytes            | B    |    1000  |    0 |
> +---------+----------------------+------+----------+------+
> | VAh     | Volt Ampere x Hour   | kVAh |   1/1000 |    0 |
> +---------+----------------------+------+----------+------+
> | g/l     | grams per liter      | mg/l |     1000 |    0 |
> +---------+----------------------+------+----------+------+

VAh should be on the SI base unit, i.e.:

   VAh           volt-ampere-hour      VAs        3600   0

g/l is a synonym for kg/m3 (i.e. scale=1); that is not strictly needed but could be added if the g/l people can’t adjust to kg/m3.

> As to the KB addition, the reason is that many are still confused as to
> whether KB is 1000 or 1024 bytes.

KB = Kelvin-Byte, pretty much nonsensical.
kB is 1000 B, which is not very widely used.
KiB is 1024 B, which we already have.

> If we are taking KB as 1000 bytes,

KB is like kph, it’s not properly defined and should not be added.

> then having an entry in the table would be a good clarification to avoid
> ambiguity.
> 
> ppt could be confused with parts per thousand by some, so I would like
> to clarify in this email that "parts per thousand" would be the "/1000"
> or "permille" entry in the secondary units registry. Please confirm this
> as well. 

Yes.
So, after adding ppb and ppt, we’ll have
/  /100  /1000  ppm   ppb   ppt
1  0.01  1e-3   1e-6  1e-9  1e-12

> I believe the scale of ppb and ppt is correct, please double check just
> in case. 

Yes.

Grüße, Carsten