| … | … | @@ -63,6 +63,12 @@ It seems to be slow, presumably due to the huge number of scheme/C++ |
| 63 | 63 | transitions. To validate this, need to run a pair-count-hacked |
| 64 | 64 | experiment, and actually measure. |
| 65 | 65 | |
| 66 | The `add-storage-count` API fetches counts from storage, if not in |
| 67 | atomspace. This needs to be used instead of the `SetValue` default. |
| 68 | It implements atomic update. |
| 69 | |
| 70 | There's also a marginal counter, but its unusued... |
| 71 | |
| 66 | 72 | ### TODO/Open issues |
| 67 | 73 | Some unresolved issues. |
| 68 | 74 | * Parser wants strings as `Node`s; it would be nice to be able to work |
| … | … | @@ -80,3 +86,6 @@ Some unresolved issues. |
| 80 | 86 | that this resembles a generic fold. The alternative to this is a |
| 81 | 87 | `AtomicSetValueLink` which holds a lock to make sure the update is |
| 82 | 88 | atomic. The risk is that naive users will discover deadlocks. Hmmm. |
| 89 | * There lready is a `FetchValueOf` but it needs to be fixed to take |
| 90 | default, just like `ValueOf` |
| 91 | * Need a 'count-key etc. comparable to matrix API |