>
> The thing you have to consider though is with the exception of the
> podcasting crowd, (I hear that form given often on Diggnation and TWiT) most
> people, when asked, will say something like:
>
> "I'm duck1123 at/on twitter{.com}"Verbally that is easier to say, but with text you often see people using URI (i see twitter.com/username on IRC all the time), often times because of auto-linking. But, you hear myspace.com/bandname a lot, so people can easily adapt. It there was a reasonable expectation that [EMAIL PROTECTED] could be > easily translated tohttp://twitter.com/duck1123 then I think more people > would use it. At that point you'd get an awful lot of confusion as to whether that is an email address that people could send an email to. One of the problems with using a URI as the username is, what do you do > about the protocol part? Are we expecting people to always write @ > http://twitter.com/duck1123 or just @twitter.com/duck1123? What about when > the distinction between http and https is significant? I think you drop the "http://"; from it and I can't think of an instance where the protocol would be significant. When people see "domain.tld/resource" they are going to assume HTTP 99.9% of the time. Just like when people see [EMAIL PROTECTED] they are going to assume it is an email address. ---- The more I think about it, if we went the direction of pointing comments at URIs, and it is up to the target server to determine what to do with a comment directed at it, could this be used for blog commenting? There could be an optional "TO" field which explicitly states what URI is being targeted. If standardized, RSS readers could include this capability and we could remove the nasty process of having to visit a blog & sign up for an account just to drop a comment, instead it could be done from the RSS reader itself. Another random thought... many want to shift towards using XMPP as the transport protocol for micro-blogging, but 70% of XMPP is overhead for "presence" (the first P in XMPP) and there is no presence with microblogging since I'm always "on". Is there a better method? Something within HTTP so then it can all remain under 1 protocol? Hooks? Twitter has wanted to implement hooks for awhile. I wonder if there is any technical reason why they haven't (other than innovation over there happens at a snail's pace). http://webhooks.org/ On Tue, Sep 23, 2008 at 5:42 AM, Daniel Renfer <[EMAIL PROTECTED]> wrote: > On Tue, Sep 23, 2008 at 6:25 AM, Derek Gathright <[EMAIL PROTECTED]> wrote: > >> a lot of people on identi.ca self-identify like this " >>> identi.ca/brianjesse" -- I noticed it at #bearhugcamp when a reporter >>> was asking people for their URLs. Zach Copley said "identi.ca/zach" >>> twitter people do that too. and everybody on myspace. >> >> >> I do this too, and to me that is the appeal of it. We'd be creating a new >> technology around an already established practice. To be honest, prior to >> this discussion I had never heard of XRI. While yes, it could work and >> could be integrated later, I'm not sure that I see the benefits over URI >> since so many people are already familiar with URIs. My business card is >> full of URI's, but I think it will be a long time before I'd ever put an XRI >> on there. >> > > The thing you have to consider though is with the exception of the > podcasting crowd, (I hear that form given often on Diggnation and TWiT) most > people, when asked, will say something like: > > "I'm duck1123 at/on twitter{.com}" > > It there was a reasonable expectation that [EMAIL PROTECTED] could be > easily translated to http://twitter.com/duck1123 then I think more people > would use it. > > One of the problems with using a URI as the username is, what do you do > about the protocol part? Are we expecting people to always write @ > http://twitter.com/duck1123 or just @twitter.com/duck1123? What about when > the distinction between http and https is significant? > > I'm not sure that [EMAIL PROTECTED] makes anything better really unless > you go with Jack Moffitt's suggestion of using DNS records to perform that > lookup. It was an interesting idea, but I think you will find too many > domains on cheap hosting where the don't have very reliable control over > their DNS records, and the others it would be difficult to access them. (I > have to go through like 10+ clicks to get to the right page on godaddy) I > think a better solution would be using XRDS or a link in the head of the > root webpage of the domain. > > >> >> The @ symbol has been used as the seperator between usernames and domains >>> by so many other services, that it only makes sense to keep it there as >>> opposed to using something else. >> >> >> The problem is you are never going to retrain microbloggers to use >> something other than @username to direct their messages to a user, which >> leaves us with @[EMAIL PROTECTED] >> >> [EMAIL PROTECTED] made a lot of sense when it was first used in '71 (25 >> years before HTTP was standardized). But now with HTTP being the most >> important in a mix of protocols that will hopefully shape microblogging, I >> think it is very important to include it when determining how to identify >> people. It's almost like URI should be used for identification, HTTP used >> for discovery / pull transport, and XMPP for push / pull transport. >> >> >> Thinking more about it, assuming we go with the email-style >>> [EMAIL PROTECTED] identifier, do we even need a character to identify >>> that you are mentioning a person? >> >> >> I've thought about that too, but I think the inherent hurdle with that is >> again... we'd have to retrain people to not use @username and instead use >> [EMAIL PROTECTED] That's a big hurdle to overcome since people have been >> doing @username since '05 on Twitter. Then what if you want to @reply >> someone on Twitter from within Twitter? You don't want to require the >> @twitter part for intra-service messaging cause 99% of people won't use it. >> At that point, you are just left with {username}, which won't work because >> usernames are often just common words. >> >> IMO, the @ symbol prefixing the string is a requirement because it has >> been used for years and people just aren't going to shift away from it. >> Sure Laconi.ca peeps can go their own way and change the @reply rules, but >> we really want something that jives with Twitter users too so Twitter can be >> persuaded to hop on board with this cross-service routing scheme. >> >> >> --- >> >> >> It's good to discuss all these additional alternatives, but I still >> haven't seen one that beats @URI, it just makes sense too much sense. A use >> like this is exactly what URI was designed for. >> >> http://en.wikipedia.org/wiki/Uniform_Resource_Identifier >> >>> In computing, a Uniform Resource Identifier (URI) is a compact string of >>> characters used to identify or name a resource on the Internet. The main >>> purpose of this identification is to enable interaction with representations >>> of the resource over a network, typically the World Wide Web, using specific >>> protocols. >>> >> >> In any message prefix a URI with an @ symbol and that server is ping'd >> with your message and can determine what it wants to do with it. Heck, it >> is almost like trackbacks. >> >> This is more of a pipe-dream, but I'd love to be able to type into some >> message box myspace.com/tom or facebook.com/zuck, and have it >> actually delivered to their inbox on myspace or facebook. At that point, >> could micro-blogging be integrated with the same intra-site messaging we've >> been using on message boards, forums, social networks, etc... for the last >> decade? Could this be a full replacement for email? If this is a >> standardized, the micro-blogosphere could go from 4 million to 104 million >> users overnight if MySpace or Facebook hopped on board. But, some contend >> adoption by the masses (*cough* AOL) killed usenet, could it kill >> micro-blogging? Those are questions for another thread, but it is something >> to keep in mind when developing ideas for where we want to steer this boat. >> > > I'm not suggesting that we turn off the @username option for people that > are still using it. @username would probably still be useful when sending a > reply to a user on the same system as you are. (much as many mail systems > will allow you to send an email to just username) > > Don't forget, we could always turn this into a URI if we ever decided to > register the protocol: > > microblog:[EMAIL PROTECTED] > > It would be a good match for the xmpp: and mailto: URI's that use the same > form. > > If only the 140 character limit wasn't such a concern. Finding good ways to > make it easy to identify a friend is completely orthogonal to finding a good > format of that identifier to be federated out into the wild, if only it > weren't for the fact that every additional character in the address is one > less in the body. > > I could live with @[EMAIL PROTECTED], but I would prefer that the first > @ be optional when specifying a full address to a user. > > Then again... it would be sweet to have the @ character trigger an Ajax-y > search of my address book / FOAF file for users that have microblog > addresses. > > @Ev <down><down><tab> I know exactly how you feel <ret> > > > >> >> >> On Mon, Sep 22, 2008 at 10:45 PM, Brian Hendrickson <[EMAIL PROTECTED]>wrote: >> >>> >>> My apologies to anybody who I offended when I shared & "tested" my idea >>> for cross-site addressing. I just wanted to see what it would look like >>> >>> http://identi.ca/notice/625565 >>> >>> http://identi.ca/notice/625553 >>> >>> my test generated an incredulous reply from Jack Moffitt, who may have >>> thought me an upstanding person before that moment. kshep too. sorry guys >>> >>> Jack please delete my comment from your site (if you didn't already), I >>> don't want to start a messy discussion there. Evan mentioned the URL idea >>> and credited me so I thought I should just explain it and be done with it. >>> >>> But I totally agree with what Derek said about naming. and his "slow >>> motion moment" haha. He inspired me to share my idea, i'm fine with it if it >>> goes no further. >>> >>> a lot of people on identi.ca self-identify like this " >>> identi.ca/brianjesse" -- I noticed it at #bearhugcamp when a reporter >>> was asking people for their URLs. Zach Copley said "identi.ca/zach" >>> twitter people do that too. and everybody on myspace. >>> >>> obviously e-mail format is a much more common meme, and is used for XMPP >>> JID best-practice lately. >>> >>> and then there's the proper XRI stuff which is technically sound and >>> demonstrates the potential for simplicity (@ahynes1 and that's it) but maybe >>> could be difficult for laypersons to take advantage of, and not free >>> >>> -- Brian >>> >>> >>> On Sep 22, 2008, at 11:23 AM, Derek Gathright wrote: >>> >>> Oooo... yeah, I hadn't even thought about people using a domain as their >>> microblogging ID, but it makes sense for single-user laconi.ca (or other >>> platform) instances. >>> >>> On Mon, Sep 22, 2008 at 12:08 PM, Joe Cascio, Jr. <[EMAIL PROTECTED]>wrote: >>> >>>> Absolutely agree. A URI is the only way. >>>> I think the most compelling reason, other that being a well-known >>>> standard already, is that a URI makes discovery possible. So, for instance, >>>> I could be "http://joecascio.net";. Just like my blog home page declares >>>> my OpenID server and delegate, so it could declare my microblogging server >>>> and ID. This also helps to attack the problem of ID proliferation. The >>>> individual sub-IDs I may be known by for email, IM, microblogging or >>>> whatever now can be subsumed by one master ID, or as many as I want to have >>>> to serve my various on-line activities, sort of like carrying multiple >>>> credit cards. >>>> >>>> JoeC >>>> >>>> On Mon, Sep 22, 2008 at 12:53 PM, Derek Gathright <[EMAIL PROTECTED]>wrote: >>>> >>>>> How exactly we namespace micro-blogging usernames was a topic Evan >>>>> discussed at Bearhug Camp and unfortunately I wasn't able to be in >>>>> attendance to throw in my 2 cents. But to me this is an extremely >>>>> important >>>>> issue that deserves discussion, so I'm bringing the debate here. >>>>> >>>>> Here's the problem (as I see it): If >>>>> microblogging/micromessaging/tweeting/whateveryouwanttocallit is going to >>>>> truly be cross-platform, there needs to be a way to direct messages not >>>>> only >>>>> to users within your own platform (i.e. Twitter, Identi.ca, etc...) as >>>>> well >>>>> as direct messages to users on other platforms (like how email works). >>>>> Also, when your message/tweet is sent to another platform and it has an >>>>> @reply in it, how is that @reply portrayed on that other platform? >>>>> >>>>> Example: Currently there are Identi.ca users that make use of a bridge >>>>> to relay their messages from Identi.ca to Twitter, and when those messages >>>>> contain an @reply, those also get carried over to Twitter. That's fine & >>>>> dandy until someone sends an @reply to identi.ca/bob who is different >>>>> from twitter.com/bob, and twitter.com/bob starts getting all these >>>>> tweets in his reply timeline that are not really supposed to be directed >>>>> at >>>>> him. The purist in me says that is a big issue that needs to be resolved >>>>> before more people start doing the same thing (*cough* >>>>> http://laconi.ca/trac/ticket/68) because it can have a detrimental >>>>> effect on the experience for users on other systems. >>>>> >>>>> Unfortunately I don't remember all the options Evan had written on the >>>>> whiteboard at Bearhug Camp, but here are some that I had thought of a few >>>>> weeks back when this issue arose >>>>> >>>>> @identi.ca/derek >>>>> @derek/identi.ca >>>>> @derek::identi.ca >>>>> @[EMAIL PROTECTED] >>>>> @http://identi.ca/derek >>>>> etc... >>>>> >>>>> You can see patterns develop, and really it just comes down to what >>>>> symbols you want to use. So what are the similarities/differences between >>>>> them? Well, all of them are made-up URI's aside from the ones that >>>>> actually >>>>> point to the user's true URI, @http://identi.ca/derek & @ >>>>> identi.ca/derek. >>>>> >>>>> As a client developer that has played with mixing twitter & >>>>> identi.catimelines (unlike Twhirl for example which separates them into >>>>> different >>>>> windows) I've really thought about this issue, and the only one that >>>>> really makes sense to me is the true URI. If micro-blogging >>>>> proliferates as much as we hope, multi-platform clients are going to be >>>>> fed >>>>> many @reply messages directed at users that aren't hosted on their >>>>> platform. >>>>> If I get a message that contains @derek/twitarmy in my client, I would >>>>> have >>>>> have zero idea where to actually point for that user's URI or what >>>>> platform >>>>> "twitarmy" even is unless I rely on a list of all the micro-blogging >>>>> platforms out there (bad idea). However, if my client gets a message that >>>>> contains @army.twit.tv/derek and I have never heard of "army.twit.tv", >>>>> it's no big deal because I have a great idea of where to point my user to >>>>> in >>>>> order to find more information about "derek". Platforms and/or clients >>>>> can >>>>> also of course hide the service domain if it doesn't make sense to display >>>>> that info (i.e. if the recipient is on the same domain as the sender). >>>>> >>>>> Just think about how different the internet would be if email addresses >>>>> weren't "[EMAIL PROTECTED]" but instead would be "gmail.com/drgath". >>>>> That would in fact be your true URI where people could send messages to >>>>> via >>>>> email, could visit via HTTP to see who that person is, could chat with >>>>> that >>>>> person via XMPP by adding that user to their buddy list, could be used as >>>>> an >>>>> OpenID, etc... Social networking would have evolved much differently and >>>>> there may not be the need for developer unfriendly silos like MySpace and >>>>> Facebook. Social networking could be... *gasp*... distributed! We can >>>>> finally use a "Universal Resource Identifier" to actually be a universal >>>>> way >>>>> to identify and access a person. >>>>> >>>>> Now, adding all of the additional modules to handle that functionality >>>>> may or may not ever happen, but the potential is at least there. >>>>> >>>>> Back to Bearhug Camp... I didn't catch all of the conversation >>>>> surrounding this namespacing/routing issue and where the conversation left >>>>> off. But I did see Evan erase the "@http://identi.ca/username"; option >>>>> and said he was comfortable with the other approaches. It was one of >>>>> those >>>>> slow-motion "nooooooo!" moments and I wanted to raise the issue to see >>>>> what >>>>> other developers thought. Am I crazy? >>>>> >>>>> >
_______________________________________________ Laconica-dev mailing list [email protected] http://mail.laconi.ca/mailman/listinfo/laconica-dev
