Move away from faucet; Make all accounts generic

[phklive]

The separation between regular accounts and faucet causes significant code additions for specific checks. Moving away from faucet accounts and making all accounts generic will enable Miden to have a broader account model enabling more flexibility for the users.

(WIP)

[phklive]

@bobbinth

Could you tell us more about the rationale behind having two separate types of accounts:

• Regular
• Faucets

As you know it adds a lot of checks, specific tests, additional complexity, etc. that we could potentially prevent if we find a way to build a holistic account. I have started looking into it (but not deep yet) hence I am sure I have more things to understand regarding the reasons and specificities.

[bobbinth]

There are a few motivations, but we can examine how important/relevant they are:

1. Most accounts don't need to issue assets - so, having this functionality be available only when needed seemed like good design.
   a. There is also small overhead associated with faucets - e.g., one of the storage slots needs to be reserved. For regular accounts, this would be just wasted 32 bytes (not a big deal for private accounts, but probably not ideal for public accounts).
2. Since we have 2 different types of assets, we'd need to pick whether a regular account is a fungible or non-fungible faucet. Maybe defaulting to be always fungible is fine - but I haven't thought it through.
3. The way we have it currently, faucets are immutable (i.e., we can't update the faucet code after it has been created). The motivation here is that once the logic of an asset has been defined, it should not be possible to change it as it may lead to strange side effects. One contrived example: a faucet is created with some max_supply and issues a bunch of assets, then changes max_supply to a lower value. Now we have more assets floating than the specified max_supply.