Received-SPF: pass (sog-mx-3.v43.ch3.sourceforge.com: domain of fastmail.co.uk
	designates 66.111.4.26 as permitted sender)
	client-ip=66.111.4.26; envelope-from=jim618@fastmail.co.uk;
	helo=out2-smtp.messagingengine.com; 
Message-Id: <1372360716.14869.140661249272837.1376DACB@webmail.messagingengine.com>
From: Jim <jim618@fastmail.co.uk>
To: bitcoin-development@lists.sourceforge.net
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain
Date: Thu, 27 Jun 2013 20:18:36 +0100
In-Reply-To: <CAAS2fgRE+gj0NrDMk-WamSU+mADJM1EqgR-6Sa2MNCbwq-0Gdw@mail.gmail.com>
References: <1372353053.10405.140661249237317.77984E1F@webmail.messagingengine.com>
	<CAJHLa0Ncac9Xt-AQBnpghqqpfR-j6Xtd9qVQoUe2dPp0kJvz1A@mail.gmail.com>
	<201306271804.51009.luke@dashjr.org>
	<CAAS2fgRE+gj0NrDMk-WamSU+mADJM1EqgR-6Sa2MNCbwq-0Gdw@mail.gmail.com>
Subject: Re: [Bitcoin-development] Proposal: MultiBit as default desktop
 client on bitcoin.org
Precedence: list

A few replies, in order of point raised:

Jeff:
Arguments against multibit default:
* Less testing, field experience on desktop

Yes this is true - downloads of multibit have typically been around
1/7th to 1/5th of bitcoin-QT downloads. It helps of course that
the bitcoinj networking/ object model is also used by Andreas 
as you note.


Greg:
I think Mike has squashed the deadlocking problems with reentrant 
locks (primarily in the Wallet). I haven't seen one in at least a month.

We discussed proxy support on the bitcoinj mailing list a while ago 
and at the time the stumbling block was the Java library used for 
the networking (Netty) did not support it. Mike or Miron would 
know better than I if this is still the case.

Change address behaviour will improve significantly when HD
wallet support goes into multibit/ bitcoinj (I am hoping to get my
bit done over the summer). Matija Mazi has been working on a 
Java impl of HD wallets so it is coming down the pipe but
there is a lot to do yet.

Connections out from MultiBit are:
+ 4 bitcoind nodes on port 8333
+ multibit.org (188.138.113.201) for help, current version info
   (and probably more in future)
+ the currency ticker will make HTTP gets to the source of
   whichever exchange(s) you have set up e.g MtGox, CampBX.
   This calls should disappear if you switch the currency conversion
   and ticker off.

I think that is all the connections out I make.

Mainly due to the exchanges abruptly changing their APIs and
breaking things we are planning to put in intermediate 
"Exchange Data Provider" servers. Tim Molter is working on this
in his XChange project. That will enable us to patch the server
when things change and the multibits in the field won't be
affected. There will probably be a couple of these initially
for redundancy.

Alex: Yes I think most users migrate to blockchain.info or,
more recently coinbase.com. They are both good wallets
but I'd like to keep Bitcoin as P2P as possible.

Luke-Jr
I think you are right here on the number of full nodes versus
SPV nodes.
I don't think we even know yet what are the working ratios of
full nodes to SPV nodes. I haven't seen anybody do any 
analysis on this.

I doubt multibit will ever participate in the Bitcoin network 
other than as an SPV client. All the optimisation is to reduce
data traffic - it is effectively a mobile wallet that happens to
live on a desktop. It is not really intended to be more than
"a wallet for regular people to store and spend their bitcoin".

In English the nomenclature for direction of the transactions
is: "Sent to" and "Received with". To be honest I 
haven't transliterated the localisation files to check other
language packs but the localisers are pretty good in my
experience.


On Thu, Jun 27, 2013, at 07:41 PM, Gregory Maxwell wrote:
> On Thu, Jun 27, 2013 at 11:04 AM, Luke-Jr <luke@dashjr.org> wrote:
> > On Thursday, June 27, 2013 5:30:21 PM Jeff Garzik wrote:
> >> * Very real possibility of an overall net reduction of full nodes on P2P
> >> network
> > Even a reduction of *nodes at all*, as I've never seen a listening bitcoinj or
> > MultiBit node. :/
> > Jim, will MultiBit be adding p2p listening support?
> 
> Without validation listening isn't currently very useful. :( Maybe it
> could be somewhat more with some protocol additions.
> 
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
> 
> Build for Windows Store.
> 
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Bitcoin-development mailing list
> Bitcoin-development@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/bitcoin-development


-- 
https://multibit.org    Money, reinvented