W3C home > Mailing lists > Public > public-webpayments@w3.org > May 2015

Re: modeling wallets

From: Manu Sporny <msporny@digitalbazaar.com>
Date: Sat, 16 May 2015 21:53:27 -0400
Message-ID: <5557F497.8030601@digitalbazaar.com>
To: public-webpayments@w3.org
On 05/16/2015 09:22 PM, Melvin Carvalho wrote:
> <Alice> <com:account> <Alice:#account1> <Alice:#account1>
> <com:currency> "USD". <Alice:#account1> <rdfs:label> "Party Money". 
> <Alice:#account1> <com:ledger> <Alice:#ledger1>
> 
> I'd been using DCT : description for the "Party Money" part, I can 
> switch to RDFS label, that looks better.

IIRC, we chose rdfs:label because it was more "low level" than
dct:description in that it was a part of the core RDF stack. Either
should work just fine, but it would be nice to know why dct:description
essentially duplicates the RDFS term.

> I've been thinking quite a bit about ledgers based on accounts and 
> ledgers based on users.  My current conclusion is that I'd like to
> try both.

We used com:account primarily because we needed to group other things w/
the abstraction (name, currency, etc.). Not all accounts have ledgers,
so that also influenced our decision to use 'account'.

> The issue with user based ledgers is that they need a default 
> currency, so I've selected bitcoin for that.

I don't think you need a default, do you? You can just assign a currency
to the inbox?

> My idea for a wallet so far is that it points to an API which allows
> the user to be input as a query string.  Then it has an inbox where
> there user can send transactions based on their user name.  One inbox
> per user.

The PaySwarm API abstraction level is a bit more low level - one API per
ledger/account. So, restructuring the previous example:

<Alice> <com:account> <Alice:/account1>
<Alice:/account1> <com:currency> "USD".
<Alice:/account1> <rdfs:label> "Party Money".
<Alice:/account1> <com:ledger> <Alice:/account1/ledger>
<Alice:/account1/ledger> <com:api> <https://w3id.org/apis/ledger/v1> .

> One thing that seems to be slightly different is that I would
> consider a wallet to be a container of many accounts/users, but can
> be used with one account/user.
> 
> <Alice> <cc:wallet> <Alice:#wallet> <Alice:#wallet> a cc:Wallet 
> <Alice:#wallet> <rdfs:label> "Main wallet at ACME inc". 
> <Alice:#wallet> <cc:api> <http://acme.com/api/v1>. <Alice:#wallet>
> <cc:inbox> <https://acme.com/etc/wallet/>.

Yeah, that's another way to model it. I think we chose to say an account
can:

1) have multiple entities that have access to it
2) have N ledgers, but most likely it'll be 1 for most transactional
   mechanisms

We avoided the whole "wallet" modeling approach because it seemed more
like a "UI" thing than a fundamental modeling thing. For example, a UI
could use the concept of a wallet to group multiple accounts together,
but that would have no functional changes on how one would access
accounts and ledgers.

> The fields above are still very experimental, and I'd be happy to
> change them.  Just trying to get a prototype working and reusing some
> common patterns.

+1 - let us know how it goes. Interested to hear how the "user based
ledger" concept pans out.

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny, G+: +Manu Sporny)
Founder/CEO - Digital Bazaar, Inc.
blog: The Marathonic Dawn of Web Payments
http://manu.sporny.org/2014/dawn-of-web-payments/
Received on Sunday, 17 May 2015 01:53:52 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:07:40 UTC