Topic: Why only BitcoinJ? (Read 3045 times)

No

You could also consider just revealing selected private keys instead of "sending".

For things like retail shopping, where people who use cash routinely trust the sales clerk to give them any change resulting from the denominations handed over adding up to more than the amount owed, that could work fine, though its convenience could be improved by doing some shuffling around behind some curtain somewhere to get nice memorable denominations to give out.

-MarkM-

You'd have to at least be able to construct a Transaction (using private key and destination address). You'd also have to broadcast it to the bitcoin network somehow. Only thing you don't really need for sending is the blockchain storage, if I'm correct.

sounds like a great way to *receive* payments, i wonder what similar lengths one could go to to *send* payments without running a bitcoin client.

Interestingly I'm currently thinking about a different solution to this problem (accepting payment on website), that doesn't involve a bitcoin client at all:

• generate a load of addresses in advance using vanitygen
• put the public part (address) into db on webserver to give to users as payment addresses
• check payment reception using blockexplorer.com/q/getreceivedbyaddress
• keep private keys somewhere else, import as needed into wallet using sipa:showwallet patch (importprivkey)

EDIT: forgot to ask for comments on this idea? Is it workable? Anyone see a problem with this approach?

It really doesn't matter if the hosting plan is shared or dedicated. The economies of scale work the same in both cases. Its just that with the shared plan the economy of scale is more obvious. In case of the dedicated hosting it becomes more of an issue of potential bandwidth/protocol abuse after incorrect modification.

Take for example the PDO layer. It was originally written in PHP with a dependency on PCRE. PHP web developers couldn't restrain themselves and were making modifications that completely destroyed one of the original reasons of developing PDO: prepared SQL statements. Finally the PHP engine developers wrapped the PDO in a black box and prevented the average PHP website programmers from sticking their fingers deep into the gearbox.

The same problem could occur with Bitcoin protocol. By putting up a higher barrier for tinkerers the libbitcoin engine developers are hoping to maintain overall consistency and safety of the Bitcoin network that depends on the correct operation of majority of the peers.

Look, PHP is not a language for tinkerers. It is a bumper car designed to have quick, innocent and safe fun. If you are into tinkering then you aren't a target PHP audience. Switch to Perl,Python,Ruby,Java,C# or whatever else. PHP is the leaning tree that any goat can climb on.

I think you are communicating very clearly. Maybe its a problem of my dialectic. I use "you" not to mean "you personally" but "you and other primarily-PHP developers that I could envision as working on such a project".

And don't think that I'm trying to slander PHP and its users. Almost everything what I said above will be true if you substitute ASP for PHP. Except that in the case of ASP Microsoft had enough forethought to develop DAO and ADO abstraction layers from the very beginning.

perhaps you're not really talking to me, but i was saying i could customize everything NOT on a shared hosting plan.

g, i'm not having very good luck communicating in this thread am i?

Nobody's going to allow you to modify the bitcoin code on a shared hosting plan. You'll get a nice high-performance interface (FCGI/shared memory), a private wallet.dat equivalent and a shared block-chain storage and networking peer. If the libbitcoin guys won't screw up the wallet.dat will have an option of being receive only with no storage of the private keys. It will allow only receiving payments as a security measure for the shared hosting.

This is how all the PHP engines evolved: you'll get a thin wrapper and an immense choice of under-the-hood modules that are written in C/C++ and not really modifiable in any way from the level of PHP. This is both the strength and the weakness of PHP. Just scroll all the way down on your phpinfo(); read it again and you'll understand what I'm saying.

no personally i just want one because i can look at php and see what's going on, and would be able to customize everything. c++ on the other hand and i'm lost.

Plus i think that if it's possible, it's probably inevitable (just to see that it works, out of curiousity).

Because of the shared hosting plans. Hosting with access only to PHP on the back end are by far the cheapest and most popular. This wouldn't require a total implementation in PHP. The alternative is that libbitcoin will become a reality and will start getting configured into the PHP engines by the various hosting providers.

Wait, why would you want a full Bitcoin node running in PHP?

That's what I meant you meant. Oh, what a mess!

i just meant some parts had been *re-written* in php... see here for a great example: http://pastebin.com/vmRQC7ha (code by theymos)

1.) C++ and php are close? what?
2.) closeness to C++ is irrelevant. Bitcoin client can be implemented in any turing-complete language.
3.) afaik, no part of the official bitcoin client are written in php

no, not the 'default' client. i just mean some parts of it have been translated to php.

one area though where i'm not sure how it would work: connecting to other nodes and listening for incoming traffic.

manipulation of keys/addressses/transactions would probably be the easy part.

I think he means that some functions of the bitcoin client are implemented with php, not that the bitcoin client itself is implemented in php.

As C++ and PHP are close, yes it's possible
But I don't think some parts of the bitcoin client are in PHP

okay so if we establish that it's crazy and we establish that python is 'best', can i still assume the answer to my question is that it's possible?

i seem to remember seeing *some* parts of the bitcoin code done in php, like address validation, etc.

That is crazy... PHP sucks for security and is very hard to get properly stable for non-trivial projects.

I would think that a full python implementation of bitcoin would be best for web-hosting.

maybe it's crazy, but would it be possible to make the bitcoin software entirely with PHP?

i don't mean RPC to bitcoind, i mean actually do everything within PHP code?