Create requirements-summary.csv

[mmccool]

Proposal for a way to organize the requirements as an incremental step towards completing the coverage table as discussed in Issue #189.

• Use title of use case section to reference a particular use case (section numbers are too subject to change right now). See coverage.csv for a complete list.
• For use cases under a more general category, use parenthesis, for example: "Smart City (Geolocation, Dashboard)"
• Please put all string values in quotes to avoid issues with embedded commas
• Give each requirement a new number, adding to the bottom of the table. If we have to delete some later there may be gaps, which we can live with (in final publication we can do a renumbering if necessary, but for now the important part is to assign short unique IDs for requirements)
• Group detailed sub-requirements using "a", "b" etc. but provide a summary requirement when possible. So the summary requirement might be R{N}, where {N} is a number, and related subrequirements would be R{N}a, R{N}b, etc.

[mmccool]

Propose to cite use cases, use the short names (section titles) identified in the coverage table.

[mmccool]

ok, I did a bunch, basically working through the existing filled-in MD templates under REQUIREMENTS for geolocation, discovery, and agriculture-greenhouse. Need to do another pass to complete cross-references to actual use cases, however, and I had to "re-interpret" some of the requirements in some cases, so this probably needs some review.

But at least this will give people an idea...

I do think the "Discovery", "Security", and "Privacy" requirements are a bit out of place, because these are "horizontals". In particular, motivating "Discovery" requirements by the "Discovery" "use case" (alone), for example, is not really good, such features should be motivated by "vertical" use cases, not a "horizontal" technology. That is why I have left the Discovery ones as TBD, I want to go through the "vertical" use cases and motivate each of these specific requirements by those, not by the horizontal one.

[mmccool]

As expected even using the "short names" for use case references makes the rows quite long. If this was in an HTML table though we could put each use case reference on a different line which would be much more readable. Unfortunately the github CSV renderer is not smart enough to wrap lines.