You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A proposal: In order to ensure the quality and correctness of the template, it's necessary that it's somehow runnable in itself.
So how do we run the template in order to see that it works?
There are two ways to approach the issue...
1. Apply template substitution (then run it)
If the template is full of substitution variables, it would need to be "instantiated" before running it. That also means testing the application of the template itself, which is good.
2. Write a template without substitution (so it's always runnable)
In this setting, the template has all its substitution variable documented but doesn't contain any placeholders in areas that need execution.
The template could for instance have a name, Pseudo BB.
All BBs that would wish to adopt from it, would refer to a list of places in Pseudo BB that needs to be customized individually, such as the name "Pseudo".
Which one?
I lean a bit towards number 2 because it doesn't deal with a substitution layer, i.e. some template engine and its quirky syntax. Why? Templating is about not just substitution of file contents, but also file and folder names. This quickly becomes messy to look.
I'd definitely suggest to avoid a mix of 1 and 2.. i.e. if you cannot have a complete and automatic substitution mechanism, then don't have one at all and go for documentation.
The template should perhaps also be structured so it looks a bit more like:
template/
docs/
README.md # A README for the template, not a template for a README! :D
This could also be achieved with separate git repos, but that ruins release flows and will make documentation drift.
That was all :)
The text was updated successfully, but these errors were encountered:
A proposal: In order to ensure the quality and correctness of the template, it's necessary that it's somehow runnable in itself.
So how do we run the template in order to see that it works?
There are two ways to approach the issue...
1. Apply template substitution (then run it)
If the template is full of substitution variables, it would need to be "instantiated" before running it. That also means testing the application of the template itself, which is good.
2. Write a template without substitution (so it's always runnable)
In this setting, the template has all its substitution variable documented but doesn't contain any placeholders in areas that need execution.
The template could for instance have a name, Pseudo BB.
All BBs that would wish to adopt from it, would refer to a list of places in Pseudo BB that needs to be customized individually, such as the name "Pseudo".
Which one?
I lean a bit towards number 2 because it doesn't deal with a substitution layer, i.e. some template engine and its quirky syntax. Why? Templating is about not just substitution of file contents, but also file and folder names. This quickly becomes messy to look.
I'd definitely suggest to avoid a mix of 1 and 2.. i.e. if you cannot have a complete and automatic substitution mechanism, then don't have one at all and go for documentation.
The template should perhaps also be structured so it looks a bit more like:
This could also be achieved with separate git repos, but that ruins release flows and will make documentation drift.
That was all :)
The text was updated successfully, but these errors were encountered: