Backward Compatibility - What and who drives the compatibility?

A tweet of Ministry of Testing read as this -- How do you decide how backwards compatible your product should be?

I learn, there is no general rule and as well no specific rule that applies when it comes to backwards compatibility, forward compatibility and the compatibility.  It is highly specific to what product we are dealing with and how the system around it are developed, or being developed or will be developed.  Also how it is deployed.

While compatibility is one of the main feature of product feature, it also signifies -- the design, the deployment style & patterns, communication and the exchange of data.

Further to be precise, compatibility is context specific task.  I have not come across where it is generic task where it is applied and applicable to any where.

If asked what software can come into compatibility spectrum, then:

>> Database System

>> Server

>> Client libraries like SDKs

>> Server Configurations

>> Operating System

>> Hardware Configurations

>> Client application versions being used by user and number of users

>> Server configuration versions being used in different client applications

>> Browser versions, it was a major compatibility concern a decade back

>> Application system version and its dependencies compatibility version

Few analytics that influences the decision on compatibility

>> Analytics showing the user on specific versions of libraries

>> Analytics showing the user on specific versions of OS

>> Analytics showing the user on specific version of hardware configurations

>> Analytics showing the user using specific feature which involves server configurations

>> Analytics showing the conversion funnel on specific setup of system

Few emotion that influences the decision on compatibility

>> Experience of user using a product on specific setup and configuration

>> Customer Support analytics on problems, queries and user satisfaction on specific setup and configuration

>> History of previous compatibility and migration

>> Competitor's presence

>> Team's opinion

Business aspects

>> Negotiation extent

>> Bearable Cost and Non-bearable cost

>> Business Value

>> Social Value

>> Strategic Value

>> All above converting to business profits

>> Survey carried out

>> Timeline

>> Budget and Resources

>> Talents -- people

Environment

>> That simulates the backward version of the system

>> That simulates the provisioned version of the system

>> Outcome of the assessment in such environments

If seen, where do tester fit here to decide the backward compatibility of the product?  In my experience, the decision of backward compatibility is most times strategic from aspect of the technicality.  In the enterprise ecosystem, the same strategic will weigh more on the business decisions and then the technical decisions.

The executing tester most time will not be involved here in such decisions.  Could be a person leading the testing team or the engineering efforts can be involved in such meetings.  In start-up environment I have worked so far, the decision meeting most times did not have the testers involved.  However, on getting the heads up for backward compatibility need, I use to assess the risks in system and its sub-systems.  Then, I shared and had a meeting scheduled if there was a need of different stakeholders and teams to clarify and resolve any risks mentioned.

In simple, the business don't want the service to be broken for a user.  The user should get the service that business provides on using the software application.  Further, the conversion what the business expects has to be what the business expects.  If not, then there is a problem to user and as well to business.