View source for Talk:Bitcoin/address validation

==Task definition==
Is there a link to the actual definition of what is a valid BTC address? The wikipedia link (at time of writing) doesn't say anything on the topic other than that it is “human readable” and “around 33 characters long” and “always starts with a 1 or 3”; that's far too vague for a Rosetta Code task! (Yes, I could try decoding some of the existing examples, but I greatly prefer to be able to independently code tasks if I can. Independent implementations help ensure that the task itself is properly possible.) A link for this sort of thing is entirely adequate for now. –[[User:Dkf|Donal Fellows]] 00:16, 28 November 2012 (UTC)
: Not sure if [https://en.bitcoin.it/wiki/Technical_background_of_Bitcoin_addresses This] is the definitive thing.  On the other hand though, is it really necessary to have a task with such a narrow range of application? --[[User:Ledrug|Ledrug]] 03:39, 28 November 2012 (UTC)
:: The bitcoin wiki is a more accurate source than wikipedia indeed.  As far as whether it is "necessary" to have this task in RC:   no, it is not.  I don't think it would hurt, though.  On the over hand, I know there are some political issues with the concept so I won't insist much if whoever is in charge here doesn't accept it, even as a draft.--[[User:Grondilu|Grondilu]] 04:07, 28 November 2012 (UTC)
::: I'm not too worried about the narrowness of the task spec, just that it's possible to work out how to implement it without reading someone else's solution and that it is possible (even if not necessarily ''easy'') to verify the correctness without reference to someone else's code. If these are satisfied and we've got evidence of independent implementations, it's a good draft and can advance to being a full task.
::::I've added more details in the description of the task.  I'll add an example of a correct bitcoin address later.  (a wrong one is easy to forge).--[[User:Grondilu|Grondilu]] 10:46, 28 November 2012 (UTC)
::: Of more interest though is whether there will be other bitcoin-related tasks in the future (specifically, the next 3 months; there's no point trying to plan further ahead than that). If not, the name of this one should change. That's independent of whether this is draft or not. –[[User:Dkf|Donal Fellows]] 09:09, 28 November 2012 (UTC)
::::I was thinking of making a "create a bitcoin address and its private key" task, but on second thought I'm not sure it can be done in a short program.  It would require [[wp:elliptic curve]] arithmetic so I may have to create a task about that first.--[[User:Grondilu|Grondilu]] 10:50, 28 November 2012 (UTC)
:::::I advise starting with “convert key system to bitcoin address”, which will include the ripemd160 step but not the key generation itself, as well as needing a base58 encoder. (This task has the decoder instead.) –[[User:Dkf|Donal Fellows]] 22:55, 28 November 2012 (UTC)
::::::Good idea.  I'll do that.  I'll call it "bitcoin/Public point to address".--[[User:Grondilu|Grondilu]] 02:44, 29 November 2012 (UTC)

===Vague task===
Looking at this further (and peeking at the proposed solutions) the task appears to be to verify that the data is base58-encoded and that the computed check-digits match the supplied ones (which is basically following the bottom of [https://en.bitcoin.it/w/images/en/9/9b/PubKeyToAddr.png this diagram], i.e., after the ripemd160 step). If this is the case, it is important to '''spell this out''' in the task definition rather than leaving it to people to guess. –[[User:Dkf|Donal Fellows]] 09:40, 28 November 2012 (UTC)

: I would second that opinion of Dkf. The only way I could work out what to do was to run another solution and find out what the intermediate values were the nconvert that to Python 3! --[[User:Paddy3118|Paddy3118]] 04:48, 29 November 2012 (UTC)