Rebase namecoin on Bitcoin 0.6 or libbitcoin

[virtual_master]

I am not an expert with the client source code but what about this modular option:
- namecoind rebase on libcoin (all libraries, inclusive Namecoin specific ones must be available already as I understood and must go fast) - they say the bitcoind based on libcoin is 100% compatible and 3 times faster
- GUI based on namecoind with additional features, based on namecoind(libcoin based or not), libbitcoin libraries and an easy language(like Python)
Advantages:
- more quickly and parallel development
- interchangeability and more variations like:

• - namecoind(actual libcoin based) + Python GUI
  - namecoind(actual bitcoind based) + Python GUI
  - namecoind(actual libcoin based) + webserver + PHP based GUI
  - namecoind(actual bitcoind based) + webserver + PHP based GUI

- Retroshare seems to come in age and could have some Namecoin extensions also, ZeroReserve is using already that, a modular structure would make easier to integrate Namecoin also in them.
Namecoin could be traded automatically with Bitcoin with a ZeroReserve integration and further Retroshare connections could be Namecoin ID based also. A private and very secretive Retroshare network may be will not disclose their connection data but for those who wish to trade each other and anyway publish it(like on ZeroReserve, a Namecoin ID based connection could be more convenient.
- Probably other projects could also profit from an available Namecoin library and modular Namecoin structure when integrating ID or other features in them.

[domob]

I am not an expert with the client source code but what about this modular option:
- namecoind rebase on libcoin (all libraries, inclusive Namecoin specific ones must be available already as I understood and must go fast) - they say the bitcoind based on libcoin is 100% compatible and 3 times faster
- GUI based on namecoind with additional features, based on namecoind(libcoin based or not), libbitcoin libraries and an easy language(like Python)
Advantages:
- more quickly and parallel development
- interchangeability and more variations like:

• - namecoind(actual libcoin based) + Python GUI
  - namecoind(actual bitcoind based) + Python GUI
  - namecoind(actual libcoin based) + webserver + PHP based GUI
  - namecoind(actual bitcoind based) + webserver + PHP based GUI

- Retroshare seems to come in age and could have some Namecoin extensions also, ZeroReserve is using already that, a modular structure would make easier to integrate Namecoin also in them.
Namecoin could be traded automatically with Bitcoin with a ZeroReserve integration and further Retroshare connections could be Namecoin ID based also. A private and very secretive Retroshare network may be will not disclose their connection data but for those who wish to trade each other and anyway publish it(like on ZeroReserve, a Namecoin ID based connection could be more convenient.
- Probably other projects could also profit from an available Namecoin library and modular Namecoin structure when integrating ID or other features in them.

This sounds like a very good plan! I've already worked with connecting to namecoind using multiple platforms and languages (as most here probably know), and it works quite well. Maybe also my own Namecoin library (see https://dot-bit.org/forum/viewtopic.php?f=9&t=1118) can be of use to a UI or other high-level application (as it will be for the Moneychanger integration with Namecoin).

[snailbrain]

i think libbitcoin and libcoin are different ?

[biolizard89]

i think libbitcoin and libcoin are different ?

Yes, they are. libbitcoin is by genjix; I don't know much about libcoin (anyone got a link?).

[virtual_master]

Libcoin is written by 'libcoin', who discovered the bug and helped us to fix it.
I think there is a fix also in libcoin.
https://github.com/libcoin/libcoin
LGPL v. 3. license

http://ceptacle.com/

[biolizard89]

Libcoin is written by 'libcoin', who discovered the bug and helped us to fix it.
I think there is a fix also in libcoin.
https://github.com/libcoin/libcoin
LGPL v. 3. license

http://ceptacle.com/

libcoin looks really nice actually... using it for the rebase would have the advantage that when Bitcoin adds functionality, we get the changes for free; no need to maintain totally separate codebase. While my opinion doesn't carry much weight since I won't be involved in the rebase, I think using libcoin for the rebase is a quite good proposal.

[domob]

Another remark: Isn't libcoin basically a full implementation of all functionality that's also in namecoind? (That's what I understood it to be ... maybe except wallet management?) In that case, a "rebase" of namecoind on libcoin would be more like implementing a thin front-end for libcoin that has the same RPC interface as namecoind used to have, wouldn't it? What about doing that but not to replace namecoind and letting the existing codebase go, but rather to have a second, independent, alternative implementation? As with Bitcoin, I think that having alternative full-node clients available is a good thing. I would definitely welcome a modernised namecoind based on the original code-base as well as one based on libcoin along side each other.

[phelix]

Another remark: Isn't libcoin basically a full implementation of all functionality that's also in namecoind? (That's what I understood it to be ... maybe except wallet management?) In that case, a "rebase" of namecoind on libcoin would be more like implementing a thin front-end for libcoin that has the same RPC interface as namecoind used to have, wouldn't it? What about doing that but not to replace namecoind and letting the existing codebase go, but rather to have a second, independent, alternative implementation? As with Bitcoin, I think that having alternative full-node clients available is a good thing. I would definitely welcome a modernised namecoind based on the original code-base as well as one based on libcoin along side each other.

I would be happy with one up to date version for now

Not sure if there is full (merged) mining code in libcoin...

[virtual_master]

Another remark: Isn't libcoin basically a full implementation of all functionality that's also in namecoind? (That's what I understood it to be ... maybe except wallet management?) In that case, a "rebase" of namecoind on libcoin would be more like implementing a thin front-end for libcoin that has the same RPC interface as namecoind used to have, wouldn't it? What about doing that but not to replace namecoind and letting the existing codebase go, but rather to have a second, independent, alternative implementation? As with Bitcoin, I think that having alternative full-node clients available is a good thing. I would definitely welcome a modernised namecoind based on the original code-base as well as one based on libcoin along side each other.

I would be happy with one up to date version for now

Not sure if there is full (merged) mining code in libcoin...

Good question to think about the mining also. Very important.
....................
from ceptacle.com
'libcoin

libcoin is a crypto currency library based on the bitcoin/bitcoin "Satoshi" client.

libcoin maintains a version of bitcoind that is 100% compatible drop-in replacement of the bitcoin/bitcoind client: You can use it on the same computer on the same files and you can call it with the same scripts. And you can easily extend it without touching the basic bitcoin source files.

The libcoin/bitcoind client downloads the entire block chain 3.5 times faster than the bitcoin/bitcoind client. This is less than 90 minutes on a modern laptop!

In libcoin, the Satoshi client code has been completely refactored, properly encapsulating classes, removing all globals, moving from threads and mutexes to a pure asynchronous approach. Functionalities have been divided into logical units and libraries, minimizing dependencies for e.g. thin clients.'
So I interpret this phrase that it is 100% compatible with bitcoind, inclusive mining.
'libcoin maintains a version of bitcoind that is 100% compatible drop-in replacement of the bitcoin/bitcoind client'
............
As far as I know bitcoind is doing only CPU mining. And who uses it any more ? Namecoin mining on CPU would be even worse.
GPU and asic mining is done with a different program.
So if libcoin/namecoind cannot make CPU mining wouldn't be important at all. At least I think so.
Te separate GPU/ASIC mining app connects to a running bitcoind with JSON-RPC API without using its CPU mining function so theoretically it could work without this function also.
And the same must be valid for Namecoin also.
Is that so or I know it wrong ?

[phelix]

CPU mining really is not important any more and you are right about RPC.

Especially the merged mining part is critical as it is Namecoin specific.