We didn't make a meeting minute for meeting #19 and #20 (I forgot how we forgot to do that).
#21 is about the March schema.
We concluded the following. Starting from trivial stuff: renaming
renaming
-
<token-card>
is renamed to<token>
. Previously I thought users might confuse it with the<token>
as a root element or the<token>
inside<input>
and<output>
. Seems no one got confused, so we call it<token>
to avoid overly complicate things. -
<view-iconified>
is renamed to<item-view>
. Previously it was called 'iconified' to show that it's not interactable (like you can't have a button or input box inside an icon) and selectable (like you can Ctrl+click multiple icons). No one interpreted it that way. We gave up and give it a name that might be associated with non-interactable and selectable. Besides, we are considering @tomek's suggestion of implementing icon for each token so "iconified" got in the way. More on that in github.
changing the way view is defined.
Previously a view is a container for HTML blocks. Now it's a container of the following elements:
<xhtml:style>
<xhtml:script>
<xhtml:body>
It might change again as we make them object references so that the developers are not forced to ship them together with TokenScript but rather instead getting them from IPFS etc. I'm sorry.
The use of <xhtml:body>
is discouraged because it is not dynamic. It is there mostly for backward compatibility of preëxisting shtml files. @hboon suggested the use of JSX as its replacement, which we will consider or use in the next version of the schema.
As long as javascript is embedded, it needs to be in CDADA. When it becomes object reference all will be easier.
adding ASNX modules
This is the biggest change. You can now define an ASNX module under a contract.
adding event
as an origin
Now an attribute can be sourced from an event. An event is filtered by a filter expression.
In our first use-case (ENS), the events filtered is supposed to be reliablly once only, so we didn't have to deal with typical query parameters like sort
and unique
. We leave that to be added later, since such change doesn't force a schema namespace update, therefore can be done separately.
the design agreed in the meeting but not included in the march schema
- adding
disctinct
in attribute-type
A distinct attribute-type is one that defines individual tokens.
Both @JamesB and @hboon opted the use of distinct
attribute in place of <origin>
tag for token. Such a decision also imply the use of RDN as token identifier instead of fixed TokenID
as byte32
(traditional ERC721 tokens are therefore identified as TokenID=0x830234802abcd8034...
where it was previously identified as 0x830234802abcd8034..
)
However, as this change is non-trivial, we opted to not to implement it until a later stage.