| … | … | @@ -123,14 +123,21 @@ a more flexible system: |
| 123 | 123 | * Read priority and caching are possible. |
| 124 | 124 | * Write mirroring and write sharding are possible. |
| 125 | 125 | |
| 126 | Some aspects are missing: ProxyNodes can only supplement AtomSpace |
| 127 | contents; they do not provide any ability to hide existing Atoms, |
| 128 | or to present an alternate view of an Atom, in the way that the layers |
| 129 | do. The layering system provides a "view" of the AtomSpace. It has |
| 130 | many of the same problematic issues that OverlayFS does, namely, |
| 131 | complexity and performance. |
| 132 | |
| 126 | 133 | This suggests that Frames should be replaced by ProxyNodes. Which again |
| 127 | 134 | elevates a StorageNode (this time, in the form of a Proxy) to be at the |
| 128 | 135 | same conceptual level as the AtomSpace. So we now have three distinct |
| 129 | 136 | "remote" operations: read, write and execute. |
| 130 | 137 | |
| 131 | 138 | ### Reinventing Cmputing? |
| 132 | | The rwx certainly make it feel like we're re-inventing computing. Lets |
| 133 | | take a closer look and make sure we are not missing anything. |
| 139 | The `rwx` certainly make it feel like we're re-inventing computing. |
| 140 | Lets take a closer look and make sure we are not missing anything. |
| 134 | 141 | |
| 135 | 142 | * The `rwx` perms on i370/s390 mainframes came with a storage key, |
| 136 | 143 | granting specific access to different memory regions. Do we have |
| … | … | @@ -184,6 +191,7 @@ How can we leverage the above for the next-gen design? |
| 184 | 191 | So we're effectively going for version 3, here; lets get it right, |
| 185 | 192 | instead of being half-assed about it. |
| 186 | 193 | |
| 194 | ### Redesigning layering |
| 187 | 195 | |
| 188 | 196 | ### Are SensoryNode's AtomSpaces? |
| 189 | 197 | By logical extension of the above thoughts, SensoryNodes should be |