So, I’m definitely +1 for renaming the “Fields” to “Tags”.

And it seems my LinkedIn thread also seemed to solidify that willingness to 
change.

A good argument for renaming was, that in Historians, which store data of all 
sorts of automation products, the term is “Tag”.


Chris

From: Łukasz Dywicki <[email protected]>
Date: Monday, 7. November 2022 at 13:48
To: [email protected] <[email protected]>
Subject: Re: [DISCUSS] Rename Fields -> Tags?
Hey Chris,
I do not insist on any side. Knowing how hard it is to get a "common
understanding" on certain things I think it is easier if we stick with
project specific concept.

Other point, we do not need to re-use a PlcField and field notion
everywhere. For example output from browse api might be a descriptor
which can be used to construct a field address. After all, a browsing
functionality might provide more information than needed to fetch data -
ie. human readable name, description or other elements which are
irrelevant for driver to get data.

As a side note, I do acknowledge that best time to do naming and larger
API alignments is prior 1.x release.

Best,
Łukasz

On 6.11.2022 13:32, Christofer Dutz wrote:
> Hi Lukasz,
>
> even in protocols like ADS and EIP at Rivian everyone is referring to any 
> data point as a “Tag”.
> So far, I haven’t come across a single person saying something else on 
> LinkedIn.
> https://www.linkedin.com/feed/update/urn:li:activity:6994584721582088192
>
> And keep in mind: PLC4X is meant to be the bridge between IT and OT, and we 
> chose a lot of stuff (Like the address patterns, etc.) to match the OT 
> expectations. After all, I will most probably be an IT person asking the OT 
> person: Please give me the address for Field/Tag XYZ. So, I think naming it 
> “Tag” would be better.
>
> I would like to name it to match the most used term: I know that this isn’t 
> always a perfect match in all protocols, but I guess that’s the difference 
> between providing a “shared API” or building individual drivers for each 
> protocol.
>
> And currently we name it “Query” in Go … so you currently say: “AddQuery” 
> instead of “AddField”.
>
> Chris
>
>
> From: Łukasz Dywicki <[email protected]>
> Date: Sunday, 6. November 2022 at 11:45
> To: [email protected] <[email protected]>, Christofer Dutz 
> <[email protected]>
> Subject: Re: [DISCUSS] Rename Fields -> Tags?
> Hey Chris,
> I am not certain if "tag" is standardized or not. Earlier, knowing only
> modbus registers and bacnet objects, I been confused multiple times what
> the tag is. For regular IT tag is rather a marker placed on something to
> categorize elements. Our field currently specifies rather a unique data
> point than a tag.
> If tag meaning comes from IEC standard then I'd opt in for a change. If
> its not standardized then I'd advice staying with a field. Our use is
> mixed IT/OT (with probably still more IT?), if we stick too much to
> automation industry terminology then we will need to bake definition of
> a tag, fight situations where we miss a "common understanding" cause we
> can't beg others for unification of their meaning.
>
> I've seen tag used in context of ethernet/ip (more precisely Rockwell
> PLCs), but haven't done a research of why. Keep in mind we have also
> object oriented protocols such as BACnet (with `device.object.property`)
> and CANopen (with `device.sdo` or `device.tpdo..`) thus in their context
> tag is far less meaningful than field.
>
> Cheers,
> Łukasz
>
> On 5.11.2022 12:23, Christofer Dutz wrote:
>> Hi all,
>>
>> I’m currently working on harmonizing our different API variants a bit and 
>> hopefully finalizing our Browse API (Which sort of wen’t through multiple 
>> levels of change between Java and Go, back to Java and now back to Go.
>>
>> One thing I learned at Rivian is, that everyone seems to be talking about 
>> “Tags” on PLCs. So I asked on LinkedIn and it seems pretty obvious that 
>> “Tag” seems to be the term mostly used in the automation industry.
>>
>> So, I would like to consequently rename “Field” to “Tag”.
>>
>> What do you folks think?
>>
>> Chris
>>
>

Reply via email to