Skip to content

Commit

Permalink
Add decision graph for diagnostics
Browse files Browse the repository at this point in the history
graphviz dot code:

```dot
digraph G {
     A [ label = "Does the message contain\ncode snippets derived\nfrom real code?"]
     A -> B [ label = "Yes"]
     A -> C [ label = "No"]
     B [ label = "Are there multiple\nsnippets that may apply?"]
     B -> "Guess" [ label = "Yes" ]
     B -> D [ label = "No" ]
     D [ label = "Are there situations\nwhere the snippet is plain wrong?"]
     D -> E [ label = "Yes" ]
     D -> "Suggestion" [ label = "No" ]
     E [ label = "Can you detect those\nsituations programmatically?"]
     E -> G [ label = "Yes" ]
     E -> "Guess" [ label = "No" ]
     G [ label = "Suggestion where correct,\nGuess or no code hint otherwise"]


     C [ label = "Is the message reasonably short?"]
     C -> "Note" [ label = "Yes" ]
     C -> "Help" [ label = "No" ]
}
```
  • Loading branch information
oli-obk committed Mar 3, 2017
1 parent 417c1ac commit aed4d38
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion text/0000-guesses.md
Original file line number Diff line number Diff line change
Expand Up @@ -201,7 +201,7 @@ This list would get extended by Guesses under the code hint category
The complexity is not in the diagnostic levels, but in choosing the correct level for a situation. This is fairly straight-forward for errors vs warnings, but in the sub-diagnostic and code-hint levels, it is less clear.
TODO: add decision graph
![chart](https://cloud.githubusercontent.com/assets/332036/23554529/be228b76-0025-11e7-99e7-627d60f5a328.png)
# Drawbacks
[drawbacks]: #drawbacks
Expand Down

0 comments on commit aed4d38

Please sign in to comment.