Hello Chris,

On Tue, 2022-02-15 at 12:20 +0000, Christofer Dutz wrote:
> Hi all,
> 
> I'm currently working on Plc4C. Here I need to save the datatype
> along with the plcvalue as there is no such thing as a "instanceof"
> in C.
> While initially this was fixed to a pre-defined list of built-in
> types, I'm currently making this more dynamic.
> 
> This got me thinking: Wouldn't this generally make sense in other
> languages too?
> 
> For example, in KNX there are multiple datatypes which result in the
> same type of PLCValue. However, one is a Power-Consumption in Watt,
> another is a Temperature in °C.
> When parsing this generally isn't an issue, however when serializing
> the representation might vary (One might be that crazy KNX 16-bit
> float, another might be a real IEC Float).
> So we need to keep the datatype in parallel with the PLcValue.
> 
Single precision floating point is one thing this could help with I
guess, but it would really be beneficial when you are thinking about
the complex data-types most of the large-ish PLC's & PAC's have, like
the axis data-type for instance. This makes working with things like
Rockwells Logix family of processors easier I would think. And also
very likely Siemens family of Sinumerik S7 PLC/PAC.

> I think having it IN the plcvalue would be better.
> 
> What do you think?
> 
> 
If it makes sense then yes. Are there caveats to this?
> Chris

Stephen

Reply via email to