Improve Flow Generation field to make clear it's fairly easy to set

[j616]

Context
Some of our discussions about the generation field highlight a misunderstanding - from the description:

An indication of how many lossy encodings the Flow content has been through. This parameter provides a hint to clients as to which is the "highest qualty" Flow available to them. A Flow with a higher generation may contain less of the original information than a Flow with a lower generation. Where a Flow is captured straight from the orginating device (e.g. camera/microphone) in its highest quality, and there is no possibility of the content becoming available in a higher quality (e.g. via capture from ST2110 or SDI), it SHOULD have a generation of 0. Where the originating device outputs multiple qualities of the Source, generation should represent the encoding processes each has been through as accurately as possible.

Unfortunately it seems the first sentence is given substantial weight, and the bar for "as accurately as possible" is calibrated fairly high, causing the field not to get used very well.

In reality unless you know the input is absolutely the highest possible quality (e.g. direct raw capture of a camera/SDI/ST-2110 feed) setting an arbitrary nonzero generation at Flow creation is probably sensible, and then incrementing it every time a transcode takes place. It should be viewed as a relative measure, and comparison across unrelated Flows (e.g. Flows of differing Sources) is generally not meaningful.

Actions

• Tweak the description to be clearer that it's a relative measure, and aside from the special value 0, no other absolute value has any particular meaning
• Make a specific recommendation that newly ingested content has it set - e.g. to 1 (or maybe 10 to give room to insert higher quality content later?) and that transcoders should increment.

Outcome

• PR on TAMS with the updated parameter descriptions