Namecoin Forum

domob wrote:Since I'm getting requests sometimes, I think it would be a good idea to implement social networking handles in NameID (and to specify them for id/ in general). I see two possibilities:

1) Just use values like "twitter" or "github" on the top level. This seems simplest and is, AFAIK, what OneName does.

2) Since these are not "generic terms" but instead names of particular companies, another idea is to introduce a new "social" value which holds an object with unspecified keys that can link to arbitrary social network handles. Like this:

Code: Select all

{
  "social":
    {
      "twitter": "my-name",
      "github": "domob1812",
      "sourceforge": "domob"
    }
}

The actually allowed names would not absolutely need to be formally specified and can be "in flux".

What do you think? My opinion is that 2) is a "nice" solution, although it may be overspecifying things. Also, one could argue that fields like "bitcoin"/"namecoin", "otr"/"gpg" or "xmpp"/"email" would have to be put under similar schemes for crypto-addresses, cryptographic key fingerprints and communication protocols, respectively. However, I believe it is justified to draw a distinction here between (semi-)official standards/protocols and company names.

{
  "social":
    {
      "https://www.twitter.com": "my-name",
      "https://www.github.com": "domob1812",
      "https://www.sourceforge.net": "domob"
    }
}

tosh0 wrote:is there any tool to create/format your nameid?
for what i see onename have this tool, but don't alow you to login to any site,
and nameid allows you to login to openid but lacks a way for newbies to create an id...

biolizard89 wrote:I can see (2) being useful in some cases, but I think it can be improved by specifying a URL for the service names. This way, there is a unique, deterministic way to identify the website that corresponds to a service username. (Your spec would encounter issues if 2 social networks have the same name.) I also would suggest not using the field "social", but something that also is inclusive of forums... maybe "web-username"?

Something like this:

Code: Select all

{
  "social":
    {
      "https://www.twitter.com": "my-name",
      "https://www.github.com": "domob1812",
      "https://www.sourceforge.net": "domob"
    }
}

domob wrote:

biolizard89 wrote:I can see (2) being useful in some cases, but I think it can be improved by specifying a URL for the service names. This way, there is a unique, deterministic way to identify the website that corresponds to a service username. (Your spec would encounter issues if 2 social networks have the same name.) I also would suggest not using the field "social", but something that also is inclusive of forums... maybe "web-username"?

Something like this:

Code: Select all

{
  "social":
    {
      "https://www.twitter.com": "my-name",
      "https://www.github.com": "domob1812",
      "https://www.sourceforge.net": "domob"
    }
}

That's an interesting idea - somewhat like XML namespaces or Java package names. However, in the end, it also boils down to finding a "unique" identifier for the network/service. What if I use "sf.net", someone "http://sf.net/" and the third one "https://www.sourceforge.net/" - with and without training slash, HTTP vs HTTPS and so on? I think this approach may lead to lots of trouble ... but maybe we can find a good rule to apply. (E. g., only the second-level domain name of the site, like "twitter.com" - but again, the site may have .at, .de, .us, .net, .com ... and it may not always be perfectly clear which one is the "main" TLD.) That's why I suggested to use simple names. If two sites/services/networks have the same "name", one would have to modify it to make it unambiguous - but I don't see this as a big risk. Do you have an example?

Regarding the name: "social" was only a suggestion. Other ideas could be "network", "service" or "webservice". Nothing really strikes me as "perfect" for now ... also "web-username" is quite descriptive, but sounds very cumbersome.