Had a great meeting with a potential partner earlier this year to discuss a complex customer requirement.
In summary, with reference to Figure 1.
- The end-user is the “bring your own yellow sweet wrapper” on the left
- The back-office has 5 sugar cubes
- Meta data about the sugar grains within the cubesĀ is aggregated into an upturned sugar bowl
- The upturned sugar bowl feeds a subset of the meta-data to a remote scrunched-up ball of paper
- The comms between scrunched-up paper and sugar bowl are un-reliable
- The scrunched-up paper and sweet wrapper are in an un-trusted environment, so can only have the absolute minimum data needed
- The user needs to search the sugar cubes for a specific sugar grain
- When found the grain of sugar needs to be securely transported to the sweet wrapper
- The user is not entitled brown sugar, so this need to be filtered
- A some point in time a golden nugget (small sugar cube) may come on-line that matches an old search, so the sweet wrapper needs to be informed
The coffee cup, nothing to do with the scenario, but without it, would not have had sugar on the table!
Anyway, looks like we have a good solution with this partner. But if you have an alternative suggestions please let me know in the comment fields below.
The coffee looks cloudy to me and I think we can assume the invisible doughnut knows the molecular make up of your sugar …
LikeLike
Nice case of “improvising using the tools available”. Would have benefitted from some toothpicks for differently-domained networks – guess there were none to hand, though.
The way I originally learned to write label_encodings files involved Lego and Post-It notes, so I’m not knocking it :-).
LikeLike