Add option for having status values use shields.io

[eoghain]

Github users have been conditioned to look for shields.io tags as a quick way to understand the state of a repo. It'd be great if the Status section that is generated could replace the STATUS with a shields.io image instead of using just plain text.

Examples:

For clarity I'm thinking of a command line switch: adr --shields new URL Scheme adr --shields new -s 1 Updated URL Scheme

[npryce]

That would make it difficult to write scripts that work with ADR files. Maybe this could be done in the publishing pipeline, rather than maintained in the source

[PeterMTaylor]

To suggest from a Powershell point of view. This article shows a way that state of files being tested which updates the status badge directly to the readme.md itself. http://wragg.io/add-a-code-coverage-badge-to-your-powershell-deployment-pipeline/

[eoghain]

I don't see how it's any more difficult than plain strings. It's just a different regex. As long as the tool defines the the set of possible strings it'd be simple enough. Also I'm thinking of this as an optional choice that people can opt into using not a replacement for the plain strings.

[npryce]

I think that this is not a feature I can support in adr-tools, because:

• ADRs are intended to be stored long-term to document how an architecture evolves over time. (Some of the systems we have are 25 years old, for example. If successful, I expect our new systems to be around at least as long). Therefore the machine-readable parts of the documentation format are designed to be as stable as possible. Web services have a shorter lifetime. Links to a service will likely become out of date and so shouldn't be used for the machine-readable parts of the document.
• Other tools also parse ADRs in the format managed by adr-tools -- Documentation pipelines, architecture visualisation tools like structrizr, etc. They would have to be understand the configuration to correctly parse out the data.
• Displaying status information visually is a view concern, and is best managed in the document publishing pipeline where it can be varied independently of the long-term storage of stable information. This can be done with a Pandoc filter, for example.

[eoghain]

@npryce Thanks for taking a look. Everything you've said makes sense. Moving this into a visualizer is probably a good idea.