Explain layering of the Store level of abstraction on the Marshal level #3567
Conversation
erights
force-pushed
the
doc-store-marshal-layering
branch
2 times, most recently
from
9373bb9
to
cb43110
Compare
| The **data styles** tag data-like containers. If a data-like container contains only data, then it is only data, passing the `isOnlyData` type test. | ||
| * A *copyArray* is a simple JavaScript array containing only passables. | ||
| * A *copyRecord* is a simple JavaScript record-like object, with only string-named own data properties containing only passables. | ||
| * A *copyTagged* is the extensibility point used by the Store level to implement *copySets* and *copyMaps*. |
There was a problem hiding this comment.
copyTagged needs more explanation. You provide various bits of precise definitional language but never say what it actually is in a more informal sense that people can actually understand. That renders the narrative much harder to follow, when I expect (but am not sure) that it's actually fairly simple and straightforward.
There was a problem hiding this comment.
Agreed. Being very precise can actually make it harder for readers to understand, when the precision requires new terms and concepts. Being a little more imprecise and using analogies to orient the reader might be helpful.
erights
force-pushed
the
doc-store-marshal-layering
branch
from
f4bdc21
to
d788f0c
Compare
| ## The Store level organized by `keyStyleOf` | ||
|
|
||
| **`keyStyleOf(p: Passable) => KeyStyle`**<br> | ||
| where `KeyStyle` is an enumeration quite similar to `PassStyle` |
There was a problem hiding this comment.
Say briefly why you need a new enumeration for keys.
| **`keyStyleOf(p: Passable) => KeyStyle`**<br> | ||
| where `KeyStyle` is an enumeration quite similar to `PassStyle` | ||
|
|
||
| `KeyStyle`: |
There was a problem hiding this comment.
Instead of re-creating a similar hierarchy, consider explaining just the differences:
- The following are identical: primitive styles (
"undefined","null","boolean","bigint","number","string","symbol"),"copyArray","copyRecord","remotable". - The following are disallowed and will throw if passed to
keyStyleOf(): errors, promises, and metaTagged. - Data with a
passStyleof"copyTagged"must have a valid representation, and will have akeyStyleof"copySet"or"copyMap"depending on that representation.
| * Store's **"remotable"** is identical to Marshal's. Every remotable is a key. On two remotables `p` and `q`, `sameKey(p,q)` compares using the remotable's unique unforgeable identity. Thus, for two remotables, `sameKey(p,q)` iff `sameStructure(p,q)` iff `p === q` iff `Object.is(p,q)` iff `sameValueZero(p,q)`. | ||
| * The remaining passables: errors, promises, and metaTaggeds, are not keys. `keyStyleOf` applied to any of them throws. | ||
|
|
||
| Primitives and remotables are keys. copyArrays, copyRecords, copySets, and copyMaps are key-like containers. If a key-like container contains only keys, then it is a key. |
There was a problem hiding this comment.
Maybe have this paragraph before explaining keyStyle.
|
|
||
| Primitives and remotables are keys. copyArrays, copyRecords, copySets, and copyMaps are key-like containers. If a key-like container contains only keys, then it is a key. | ||
| * A *copySet* can only contain keys, and so each copySet is necessarily a key. | ||
| * A *copyMap* entry can use only use a key as a key, but can use any passable as a value including unrecognized copyTaggeds. But a copyMap is a key only of all of its values are keys. |
There was a problem hiding this comment.
You said above that everything (including copyMap) must be hardened to qualify as a passable, and therefore as a key. But that doesn't make the Map contents immutable. Does the copyMap's identity (as a key) remain stable in the face of additions/removals/changes to the copyMap contents?
There was a problem hiding this comment.
Also, does sameKey on a copyMap require a full examination of the contents every time? I imagine you have some clever trick up your sleeve, but using any container as a key (in a language that lacks hashable object identities) sounds pretty expensive.
erights
force-pushed
the
doc-store-marshal-layering
branch
from
d788f0c
to
790108c
Compare
|
Converted to Draft until it is more coherent. |
erights
force-pushed
the
doc-store-marshal-layering
branch
from
790108c
to
d47e77a
Compare
erights
force-pushed
the
markm-new-passable-styles
branch
5 times, most recently
from
cb6d43f
to
a87adc6
Compare
erights
force-pushed
the
doc-store-marshal-layering
branch
from
8412a8d
to
aea0f8d
Compare
erights
force-pushed
the
markm-new-passable-styles
branch
from
72cc56a
to
dd8e9ad
Compare
|
Closing in favor of #3653 |
No description provided.