jasinlee
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 02:40:51 AM Last edit: April 29, 2013, 03:01:35 AM by jasinlee |
|
There's a few things going on - hence trying to clear up things.
1) Nodes What: Long standing dvc nodes - ideally placed on fast connections & configured with a large number of connections.
Why: To allow new & existing clients to bootstrap quickly. To allow a complete fall-back when other systems aren't functioning.
How: Coded dns names in the code - that allow cname & a lookups.
2) Locators - client discovery What: Review the current infrastructure to ensure it's suitable for scalable growth. IRC, DDNS (DNS Seeding), Client cache
Why: It's good to review where you are and where you're going. DNS Seeding should be reviewed and checked to see if the current setup is optimal. DNS Seed names (DNS entries) should be checked & updated - are we using the most appropriate record types, what's the change control process. As you mention - does it fallback properly should those DNS Seeds be taken out? Can geoDNS be leveraged to help users download the blockchain from their closest neighbours. Tech has moved on - DNS has improved a lot - with ipv6 more prevalent now - srv & locator records can be used and you have more granular control over the how & what. geoDNS is now pretty affordable and no longer limited to the big-boys - it can make a huge difference to the blockchain startup/catchup time for end-users. The existing locators aren't setup to gather & supply metrics - each DNS query, irc join is a statistic that's useful in gauging adoption, growth & measuring scale.
How: Test, metrics, target goals for how this works.
3) DNS What: Using existing DNS names to make config & setup easier & simpler for end users.
Why: Long term success in pretty much all things comes down to making something accessible to end users.
How: Update the way a handful of (1) are recorded in DNS, update the documentation for users telling them they can addnode=dvc.devcoin.me and you'll be up & running in no time. Using something like dvc.devcoin.me is way easier than 15 addnode=abc.def.ijk.lmn entries when a user is complaining that its taken 20h for their blockchain to be downloaded.
I'm overly cautious when it comes to DDNS because there's risk - it's easy to pollute and as it's a centralised public resource it's an address book of hosts.
I would like to suggest a bounty for bringing the devcoin client current to be capable of using the seed nodes. Of all the alt coins posted to an exchange it looks like dvc is the furthers behind on the clients. To keep the currency relevant we need to plan ahead, this looks to be a necessary step to continue the development of the currency. Over the past 10 years css and xhtml have become more prominent, this is a similar case where the standard method is changing and I would not want the clients to get so far behind that they would cause new users of dvc to give up in frustration. Good design will facilitate the new & update speed of users with regard to the blockchain, help tx propagation across the network, scale with as the number of users grows and provide a resilient mechanism for clients locating other clients. I would also like to see the receiver list integrated in some manner. One idea would be: Lookup this months distribution address in the blockchain (via the dvc client) - pull up the first tx - grab the addresses the payment was sent to - distribute the y shares between these addresses. Manual payments that are done at the moment (non-generating shares) are the remainder of y. Send that amount to the this months originating address. This would all be automatically handled by the dvc client at the mining pool. Pools wouldn't have to change anything. Not sure if this has had a bounty requested yet, but this would address more than half of the new devcoin users that have trouble getting the client working. They are used to bitcoin and other clients, you install and no files needing downloaded and installed in specific directories. Exactly, and awesome, you got the python so thats an additional language.
|
|
|
|
pyra-proxy
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 03:15:51 AM |
|
Would it be possible to have a bounty for someone to update electrum to work with DVC as well? Think it would be a good project to support those who want to run a DVC client but don't want the weight of yet another block chain to download.
Bounties for server and client (solo coin and merged coin variants) might be needed/appreciated as well.
|
|
|
|
jasinlee
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 03:29:57 AM |
|
Would it be possible to have a bounty for someone to update electrum to work with DVC as well? Think it would be a good project to support those who want to run a DVC client but don't want the weight of yet another block chain to download.
Bounties for server and client (solo coin and merged coin variants) might be needed/appreciated as well.
For the second portion I think I remember a bounty for it already with the multi wallet whether it be a client or web wallet I dont think mattered. Unthinkingbit, what is the waiting bounty on the multiple currency wallet? And do we have one for the electrum wallet port to dvc or multi coins?
|
|
|
|
emfox
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 04:27:22 AM |
|
Would it be possible to have a bounty for someone to update electrum to work with DVC as well? Think it would be a good project to support those who want to run a DVC client but don't want the weight of yet another block chain to download.
Bounties for server and client (solo coin and merged coin variants) might be needed/appreciated as well.
For the second portion I think I remember a bounty for it already with the multi wallet whether it be a client or web wallet I dont think mattered. Unthinkingbit, what is the waiting bounty on the multiple currency wallet? And do we have one for the electrum wallet port to dvc or multi coins? I have tried to make electrum work with dvc, but faild. It is not that we should update electrum to work with dvc, but should update dvc to work with electrum. because electrum need several api calls and features from those *coind. Once meet the requirement, electrum could easily be ported to any altcoin.
|
Earn Devcoins by WritingBTC: 1Emfox1WswYcd2YucUskRzqfRWKkcm1Jut DVC: 1Emfox1WswYcd2YucUskRzqfRWKkcm1Jut IXC: xnRKo3qSDdcPJ4pgTLER3orkquUVQXeLwf
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 04:51:57 AM |
|
Adding our stable nodes to the client is, I am pretty sure, trivial, it is mostly waiting for an actual list of actually stable nodes. It is silly to need a bounty for that, it just needs me to input some IP addresses into the code in the same place where long ago I removed the IP addresses of bitcoin nodes. A minute or two's work, with the puublishing of the new version being motivated mostly by the desire to get another checkpoint out there to further secure the cold-wallet devcoins the DigiDeVCoin tokens on the Digitalis Open Transactions Server represent.
Doing things using hostnames instead of IP addresses I am less eager about because as I mentioned before i0coin dies on boost's failure to look up an IP address from a hostname. It is possible that the reason I0coin is the only coin I see doing that could be because it is trying to look up a network-time server by hostname, it might not, or might not *always*, be trying to look up a seed node by name or do some lookup related to clients it is connected to or trying to connect to. But it indicates that using boost to look up hostnames might not be a great idea.
Also, we are supposed to be plugging in STABLE nodes. If they are STABLE we should reasonably be able to expect at least some of them to still be at the same IP address for at least a few years. That is part of being STABLE.
The bASIC guy hijacked our entire devtome even though that was based on domain name not on IP address, so really if you cannot rely on someone to keep a server at a STABLE IP address I do not really think there is a whole lot more chance, really, that they will maintain a STABLE hostname either. A fuckup is a fuckup, I tried to warn the people ordering bASICs that the guy could not even administer the complicated task of renewing a domain when it expires (or was deliberately stealing our domain and content) so expecting him to actually produce an ASIC was massively stupid. They shrugged it off as if oh well who cares, so he cannot do the simplest frikking thing, so what, he will do this amazingly complex thing just fine...
So basically if people cannot even just frikking keep paying for the same damn IP address for a few years expecting them to maintain a DNS is like are you frakking crazy that actually takes maybe some knowledge not just throwing money at a hoster saying "I demand a fixed IP address".
If it IS safe to do DNS lookups in the code though, I don't mind putting in some hostnames on a domain whose DNS I control, and plugging our list (where is that list, by the way, I do not see it ha ha please lets come up with one) of STABLE nodes' IP address into DNS as hostnames. I can even spread the DNS providers / domain name providers out, as I have domains with Network Solutions and with Godaddy and with NameCheap, so even if two out of three of those providers goes down there will still be one provider left doing DNS for a host record that is (supposedly) the IP address of a STABLE node.
But, again, if the IP addresses are good its almost easier to put them into the code than into all three DNS/domain providers' DNS systems. And that allows us not to use boost's seemingly not all that great to use DNS-lookup code.
Maybe we should consider putting a "time to die" trigger in the code too, so people HAVE TO update their development-coin code regularly, it is after all all about development...
-MarkM-
|
|
|
|
metazilla
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 75
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 06:11:32 AM |
|
Maybe we should consider putting a "time to die" trigger in the code too, so people HAVE TO update their development-coin code regularly, it is after all all about development...
-MarkM-
This seems pretty reasonable to me. How difficult would it be to have the (official) client automatically push updates?
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 06:38:27 AM |
|
Pushing updates should probably be left to the distros, e.g. apt-get or yum or Windows Update or whatever a particular platform uses to automatically update its users' machine to latest versions of software.
-MarkM-
|
|
|
|
Unthinkingbit (OP)
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:08:52 AM |
|
.. Unthinkingbit, what is the waiting bounty on the multiple currency wallet?
I don't remember making a bounty for a multi currency wallet. I don't plan on making one until at least after the problems with the devcoin client are fixed. And do we have one for the electrum wallet port to dvc or multi coins?
No, because a port to electrum is waiting on devcoin being ported to 0.8.1+: I have tried to make electrum work with dvc, but faild. It is not that we should update electrum to work with dvc, but should update dvc to work with electrum. because electrum need several api calls and features from those *coind. Once meet the requirement, electrum could easily be ported to any altcoin.
Once devcoin is ported to 0.8.1+, then there will be a one share bounty for an electrum wallet.
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:11:38 AM |
|
Once bitcoin settles down into a stable version it will be time to start a repo that has just bitcoin with merged mining patches applied.
Then from that each merged mined coin can start making its own mods to turn that into each specific merged mined coin. Devcoin is only one of several coins waiting for that.
-MarkM-
|
|
|
|
psybits
Legendary
Offline
Activity: 1386
Merit: 1000
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:12:13 AM |
|
Guys - can someone please make a Devcoin client for windows that is professional and easy to use and claim the bounty for that ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) I think the windows exe files already exist so it shouldn't be too hard for someone with the skills. This will help in more people using Devcoin!
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:18:55 AM Last edit: April 29, 2013, 08:29:38 AM by markm |
|
Can we please try to make a list of all STABLE nodes so I can plug them into the code so when someone builds a windows client it will be able to find connections?
I have a node intended to be stable at 198.154.60.183
Are there any others out there?
Once upon a time we were paying to have five stable nodes, I guess maybe as soon as we stopped paying they all shut down?
(Maybe part of the reason for choosing five was so each could get 1/5 of a generation share and it would add up to a whole number of shares?)
Note that coin daemons apparently take more memory as they get more connections, so we want these to be big nodes, not little virtual machines with half a gig or a gig of RAM. I am running mine on my 16 gigs of RAM dedicated server but my second dedicated server has 24 gigs of RAM so if the 16 gig ever isn't enough I could switch to using the 24 gig box. (Though that has a different IP address (on the same class C subnet) so I guess maybe putting a name in a DNS could be useful if DNS lookups work well, so that I could switch to the other box without needing to change an IP address hardcoded into the code.)
(I guess we should think up some standard names we will use for at least five stable servers, maybe the same names on different domains, like dvcstable1.knotwork.com, dvcstable1.decoin.org, dvcstable1.devcoin.com, dvcstable1.devtome,org and so on, maybe all of a given number pointing to the same stable node, and put into the code five or so different domains on which those hostnames can be found?)
-MarkM-
|
|
|
|
jasinlee
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:41:41 AM |
|
108.61.57.117 and 5.9.233.104 are the seed nodes you were after before, did you ever test it ?
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:54:29 AM |
|
What the heck, I might as well run devcoind on my second dedicated machine too, so there will be no later worries about changing the IP address to move from one machine to the other. Both are on the same class C subnet so any one user's one client will not connect to both, as there is code that makes them try to pick from different class C subnets if possible.
Thus, my second "stable" server is the 24 gigs of RAM machine at 198.154.60.61
The list put into the code though should include at least five distinct class C nets, even if that means more than five actual machines.
-MarkM-
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:57:49 AM |
|
108.61.57.117 and 5.9.233.104 are the seed nodes you were after before, did you ever test it ?
I guess not since my devstart.sh script has -addnode=50.19.210.139 -addnode=107.20.209.11 \ -addnode=107.20.228.112 -addnode=209.217.250.10 \ -addnode=206.71.179.116 -addnode=99.254.112.129 \ -addnode=187.75.230.175 -addnode=201.6.123.98 \ -addnode=80.74.157.31
I expect most of those are the ancient ones that have long gone away. So, I have commented that whole block out and now instead have -addnode=198.154.60.183 -addnode=198.154.60.61 \ -addnode=108.61.57.117 -addnode=5.9.233.104
-MarkM-
|
|
|
|
jasinlee
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 08:59:35 AM |
|
Alright well those 2 ips should work they are the ones we put up the other day. Let me know if they work out.
|
|
|
|
metazilla
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 75
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 09:13:17 AM |
|
A few updates to my DVC Ticker app (should I post a separate thread for this?). You can now specify an amount of coins to convert, and the API requests are cached to reduce load on Vircurex (more exchanges are next on the list). Here's an example of how this could be used: There are 180000000 DVC generated every 4000 blocks (one "round"). Based on http://www.devtome.com/doku.php?id=daily_script_results, there are 107 shares (for devtome writing) and based on https://github.com/Unthinkingbit/charity/master/bounty_23.csv, there are 71.6 other shares, for a total of 178.6 shares. 180000000 / 178.6 ~= 1007839 DVC. A round's worth of DVC = ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2F180000000dvc.png&t=665&c=7xyT9haOxkjpig) A single share = ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2F1007839dvc.png&t=665&c=BamMJ9ppXBmUEQ) Note that a bitcoin is worth ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2Fbtc.png&t=665&c=qX9om-cAQzM2yA) Also FinShaggy currently has 61 shares from word count alone! Worth ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2F61478179dvc.png&t=665&c=2ogDg70RSAO6jw)
|
|
|
|
markm
Legendary
Offline
Activity: 2954
Merit: 1090
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 09:24:29 AM |
|
Also FinShaggy currently has 61 shares from word count alone! Worth ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2F61478179dvc.png&t=665&c=2ogDg70RSAO6jw) He wrote 61,000 words of original material? Or adequately re-worked in his own words three times that much of other people's material? Even Stephen King doesn't write that much that fast does he? -MarkM-
|
|
|
|
twobits
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 09:25:31 AM Last edit: April 29, 2013, 10:06:27 AM by twobits |
|
I would like to suggest a bounty for bringing the devcoin client current to be capable of using the seed nodes. Of all the alt coins posted to an exchange it looks like dvc is the furthers behind on the clients. To keep the currency relevant we need to plan ahead, this looks to be a necessary step to continue the development of the currency.
Brilliant man, just brilliant. You suggest a bounty to make the client capable of doing what is already is capable of doing. You really like low effort bounties don't you. Since this was told in discussions in the very thread that you quoted you either did not do your research or just felt like posting nonsense? Over the past 10 years css and xhtml have become more prominent, this is a similar case where the standard method is changing and I would not want the clients to get so far behind that they would cause new users of dvc to give up in frustration. Good design will facilitate the new & update speed of users with regard to the blockchain, help tx propagation across the network, scale with as the number of users grows and provide a resilient mechanism for clients locating other clients.
You really do take that 'Blah Blah Blah' under your avatar seriously don't you? Not sure if this has had a bounty requested yet, but this would address more than half of the new devcoin users that have trouble getting the client working. They are used to bitcoin and other clients, you install and no files needing downloaded and installed in specific directories.
This is being addressed already, do you read the posts? I propose that we limit the number of proposals one person may make in some time span without paying a fee to propose more things. This may make it so people actually think about, and research their proposal before doing something as silly as proposing a bounty for something that already is in the code.
|
█████ █████ ███████ █████ ███ █████████████ █████ ██ █████████████████ █████ █ ██████ ██████ █████ ████ ████ █████████████ █████ ████ █████████████ █████ ████ █████████████ █████ ████ █████ █████ █████ █ ██████ ███████ █████ ██ ███████████ █████ █████ ███ █████████ ████ █████ █████ ███████ ██ | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | | | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | ►WhitePaper ►One-Pager | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | █████ █████ ███████ █████ ███ █████████████ █████ ██ █████████████████ █████ █ ██████ ██████ █████ ████ ████ █████████████ █████ ████ █████████████ █████ ████ █████████████ █████ ████ █████ █████ █████ █ ██████ ███████ █████ ██ ███████████ █████ █████ ███ █████████ ████ █████ █████ ███████ ██ |
|
|
|
|
metazilla
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 75
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 09:30:16 AM |
|
Also FinShaggy currently has 61 shares from word count alone! Worth ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fdvcticker.appspot.com%2Fvircurex%2F61478179dvc.png&t=665&c=2ogDg70RSAO6jw) He wrote 61,000 words of original material? Or adequately re-worked in his own words three times that much of other people's material? Even Stephen King doesn't write that much that fast does he? -MarkM- I was skeptical too, but looking through his userpage( http://devtome.com/doku.php?id=wiki:user:FinShaggy), every article I've checked looks original. ![Shocked](https://bitcointalk.org/Smileys/default/shocked.gif)
|
|
|
|
twobits
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
April 29, 2013, 09:30:30 AM |
|
Guys - can someone please make a Devcoin client for windows that is professional and easy to use and claim the bounty for that ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) I think the windows exe files already exist so it shouldn't be too hard for someone with the skills. This will help in more people using Devcoin! Devcoin is pretty much the same client as bitcoin, at the time it was created. What is the issue with the client you want to see addressed?
|
█████ █████ ███████ █████ ███ █████████████ █████ ██ █████████████████ █████ █ ██████ ██████ █████ ████ ████ █████████████ █████ ████ █████████████ █████ ████ █████████████ █████ ████ █████ █████ █████ █ ██████ ███████ █████ ██ ███████████ █████ █████ ███ █████████ ████ █████ █████ ███████ ██ | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | | | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | ►WhitePaper ►One-Pager | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | | ███ ███ ███ ███ ███ ███ ███ ███ ███ | █████ █████ ███████ █████ ███ █████████████ █████ ██ █████████████████ █████ █ ██████ ██████ █████ ████ ████ █████████████ █████ ████ █████████████ █████ ████ █████████████ █████ ████ █████ █████ █████ █ ██████ ███████ █████ ██ ███████████ █████ █████ ███ █████████ ████ █████ █████ ███████ ██ |
|
|
|
|
|