Use DATA_FILE instead of RUN_TEMPLATE ?

[larsevj]

According to the ERT docs:

https://fmu-docs.equinor.com/docs/ert/reference/configuration/keywords.html#id21

For templating the Eclipse or Flow DATA file, use the [DATA_FILE](https://fmu-docs.equinor.com/docs/ert/reference/configuration/keywords.html#id4) keyword. In addition to templating it will also ensure that Ert knows if the DATA file is to be executed in parallel.

[rnyb]

For Drogon isolated changing back to DATA_FILE (we used to have it...) will not be an obvious improvement since we run Flow that does not have PARALLEL keyword. With Flow we use NUM_CPU to set number of cpu's.
However, we could change it if we regard this as a general recommended practice, but this is not entirely clear to me.

Would it make sense to investigate possibility to only have one way of doing this? E.g. make RUN_TEMPLATE handle all cases? This may be less confusing for end users. DATA_FILE also have the additional DATA_KW way of defining variables for substitution. Is this redundant or does it have some benefits over DEFINE?

Benefits of DATA_FILE:

• More robust for Eclipse users

Benefits of RUN_TEMPLATE:

• More generic
• Name of copied file is set explicitly with RUN_TEMPLATE
• We also use RUN_TEMPLATE to populate tmpl files.