Don’t quite understand it yet — can you give some examples of how this might apply in a Twitter conversation?

Hmm. I see more people using RT: (retweeting) or saying “Via: @username” than using a canonical source, which seems incidental to the content — not part of its meaning.

“Seems like an em dash or hyphen is more traditional.” -Chris Messina

The reason why hashtags have succeeded is because they solved a problem of ad hoc association or affiliation because Twitter never supported groups. When I introduced the concept of the hashtag, that was the primary use case, and people were constantly griping about how Twitter — a technology — was getting in their way. I don’t see the same demand for specifying source, but your proposal is still interesting — just unlikely to spread until people feel that the solution solves an immediate need for them.

I agree, Chris. It will be picked up only if there is a need for it. Aside of simplicity of use and making a distinction between people and written sources, it will take some some time to show if there are additional benefits of using asterisks as source reference.

I could imagine that sites using references listed by source would add value for users, in addition to grouping content by name or hashtag.

I think you need to show some examples of usage

1) What are you going to reference and in what exact format: URIs only?

2) What is the delimiter of reference start and stop (blank)… propose a RegEx

3) I think * is already with blod/emphasize… looking a wiki syntax ^ could be an alternative

Please show more example of intended usage. For me * is occupied with bold / emphasize.

1) What do you want to reference in what syntax: URI only?

2) What is the start and the stop delimiter so it can be efficiently extracted. References in body text too? Best is to supply a RegEx to be discussed

3) Should the reference be canonical? I.e. using a referencing service like http://www.doi.org/ (or much simpler)?

Looking at wiki syntax ^ could be a preferred alternative

At first glance and when thinking about this in the context of twitter, I thought you were proposing the asterisk as a replacement for “via” when attributing something to a person or source. Upon further reading, maybe that’s an oversimplification of what you’re suggesting, but maybe that would be a good place to start. Sure, it only saves a couple of characters, but sometimes you really need those two characters.

Simplicity is ONE of the considerations for adoption. The other is that it “feels right”

To me, using the asterisk for “source” doesn’t feel right. There’s too much prior usage and too much other noise to decide that it just works for that purpose.

I’d argue that @ works specifically because it was never used much outside of email, comment threads, and forum posts. It’s mostly been used for attribution since forever ago.

I guess I’m saying, I like the idea, I dislike the implementation. Ben’s got the right idea with RT fast becoming the standard for nanocontent. Mainly that’s due to TweetDeck’s adoption as the big gun for hardcore twitter users.

Interesting article, it seems logical to follow traditional sourcing. Maybe it will catch on as the medium become established.

Great idea although I don’t think the asterisk is the correct symbol it implies a wildcard to most technical types already therefore I think the idea is a good one but the choice of symbol is not.

@Prydie

@jstuker, my intentions are simple. I want to expand the set of @ as a reference to a name, and # as a reference to a tag, with * as a reference to a source. Yes, the source is canonical in most cases.

As a person we make statements, and those statements are sometimes cited. Institutions and companies make statements, and they are cited too. But not every source is a statement of someone. Sometimes it’s a canonical source, sometimes it’s out of an article where you don’t know the author, sometimes it’s out of a publication with multiple authors.

“I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do.” — A famous line out of the movie “2001: A Space Odyssey”.

Whom did I cite? The author of the movie script? The director of the movie? The movie itself (a product) or the protagonist (a computer)?

Of course I quoted the protagonist, but the source is the movie:

“I am putting myself to the fullest possible use, which is all I think that any #conscious #entity can ever hope to do.” @hal9000 *2001aspaceodyssey

This sort of application will not work everywhere. But I believe in its potential for microblogging and subsidizing services, such as WeFollow.

hi guys,

great discussion, very sensible suggestion. I’d like to add to the debate, because I believe there is a great need for semantic standards in the sphere of user generated content.

@ and # turned up because they solved the most important relationship languages solve in general - identity of someone/something. There is nothing in the world around us that is more important than identity.

the * standard you are suggesting is solving a different problem - description of a specific relationship, which means it is competing with many other relationships between entities - why not have a symbol for ‘i like’ and ‘i hate’ also? I’d love to have it, but there are so many similarly important relationships, that basic symbols aren’t enough to code them anymore.

this shows in your own example even. you suggested writing ‘@core from *wikipedia’, where you actually duplicated the relationship information using both, ‘from’ and asterisk. if what you proposed really made sense, it should be written and readable as:
@core *wikipedia
… which I think is not sufficient information, specially when we start combining with #…

i think our brain can assume prefixes express something essential for particular entity, and in generalized usage this can only apply to identity by default. everything else needs explanation.

so i can imagine using * as a default relationship denominator within well defined topical domains - for example meaning a source in a community that is structured around sharing and pointing back to content.

to solve this in general we invented basic semantic web technologies and there is a beautiful universe of opportunities evolving from them. however, we should not expect that end-users would express themselves in triplets, the tools they are using should translate their thoughts transparently.

best, b

Inventing[1] a syntax for a theoretical need[2] nearly always fails.

The “@” convention on Twitter succeeded because:
1. People demonstrated a need to refer to other people - evidence: numerous uses of references to people and usernames in tweets, even early on in Twitter’s history.
2. People converged (perhaps re-used from online forums?) the “@” symbol as a way to explicitly reference usernames - well before Twitter took advantage of the emergent behavior by hyperlinking such references and thus explicitly adopting the syntax that the community had organically converged on.

Existing Twitter users have already solved the semantic citation problem with inline shortened URLs (e.g. see most of http://twitter.com/timoreilly ).

For end-citations/expansions, there is established practice in the web design/development communities (most W3C mailing lists[3], see examples[4]) of using bracketed numbers (as above) in plain text media to provide footnotes which can be used for either explanations, citations, or both.

I don’t know the origins of the bracket end/foot-note reference, but presumably it’s due to limitations of the plain text medium (no superscripts) compared to print, and a desire to not overload traditional means of superscripting in plain text, e.g. the “^” for exponents.

In a constrained medium such as Twitter where every byte counts, it may make sense to reconsider using “^” as @jstuker suggests. I.e. just use a standalone ^ for one reference in a tweet, and number them for more than one, i.e. ^1 ^2, and then reverse the order for the end citations, 1^http://… 2^http://… . E.g. these real world examples could be redone with end-citations:

1. One citation. http://twitter.com/mkapor/status/1449285878 redone:

Obama to bank CEO’s: I’m the only thing between you and the pitchforks.^ This has to be the line of the week. ^http://bit.ly/18OlTW

2. Two citations. http://twitter.com/timoreilly/status/1449710728 redone:

Obama to bank CEO’s: I’m the only thing between you and the pitchforks^1 This has to be the line of the week.^2. 1^http://bit.ly/18OlTW 2^http://twitter.com/mkapor/status/1449285878

(full twitter permalink URL used for demonstration purposes, in practice all URLs would be shortened to better fit Twitter’s constraints).

This comment became much longer than I intended, I’ve archived it on my wiki where I may iterate upon it.[5]

Tantek

[1] I say “inventing” because no real world evidence, e.g. links to actual examples (as opposed to an artificial demonstrative example as in the above post ) on the public web, has been provided.

[2] I say “theoretical” because no real world need has been shown for providing end-citations (as opposed to inline-citations) in a medium like Twitter.

[3] http://lists.w3.org/

[4] http://lists.w3.org/Archives/Public/www-style/2002Sep/0067.html

[5] http://tantek.pbwiki.com/CaretNotes

Interesting points, Bostjan. I agree that “from” becomes redundant — in fact that’s the point of using an asterisk, so that the source (from) declares itself.

Referring to a source is a meta element, unlike identity like you describe. My intentions were to think about adding a tool to the toolset we have, without making life more complicated.

So we will see how this works and if it will be picked up at all. Especially on the Web, people only start using something more commonly if it works for them — independently of its logics or boundaries set by other meaning or usage of the same element.

Who cares if * and / are used to style text as bold or or italic in wiki markup? It doesn’t matter, if this application within the context of microblogging works well enough to carry on.

Jeff Croft sent me a tweet today:

“Good idea, but is it solutioneering?” And I replied with a citation:

“Twitter is the messaging service we didn’t know we needed until we had it.” @biz *colbertnation

I agree with @Tantek that the immediate need to cite something on Twitter is already met by the use of (shortened) URLs. This addition seems a bit too subtle and a bit awkward. Like @bostjan said, we shouldn’t expect end users to be ‘expressing themselves in triplets’.

The @ convention is immediately useful because it has a distinct function within the Twitter system, and it addresses a very immediate need for the user - replying to a particular person. We’re all so used to that with email.

To me, the #hashtag is already in danger of being a little bit contrived - it appeals to me as a web geek because I know that it enables people to do interesting things on http://hashtags.org. But I would guess that the majority of Twitter users don’t bother with them.

Asking them to consider subtle things like what kind of source relationship we have is going to be even less appealing. I mean, people don’t even bother with using apostrophes and hyphens properly these days (which also have a clear semantic function), so you can see how little energy people are willing to devote to syntactical correctness.

English isn’t my native language, so “semantic reference” may be the wrong expression, which may have been misleading.

The blurry definition of the idea was intentional though. I could have worked out a full fledged proposal, prepped with real-life examples and spiked with historical references. Instead I just put it out there, for the bright minds of the Web to play with it.

Not surprisingly, this discussion has formed three subject layers: logics, technology and social.

On the logical layer, I agree with Tantek (@t): the convention of using a symbol as reference isn’t logically solved. However, I do not believe you need to mark both, citation and the referred URL to make a source-symbol work.

During the Web 2.0 Expo 2009 opening, @timoreilly said: “I think there is a really important concept, and that is, the meaning doesn’t have to be formalized. It can be sophistically extracted.”. This is an important realization that touches both, the logical and the technological layer of this discussion. “Reference to information is no longer ruled by people typing it in”. (@timoreilly *web-2.0-expo-2009).

To make the links work with these citations, I entered the URL-references manually. What if a service would do that for me? What if the service is intelligent enough to sort out the URL handling? No longer would we be required to copy and paste or route our tweets through URL-shortening services.

This is what happened with @-signs and hashtags. Twitter recognized the potential and incorporated it in its service and API.

What if * replaced http://www in microblogging? It could work somewhat like wiki markup text, just simpler. It would use the meme of the source reference, without duplicating its exact functionality, like @t suggested.

But how can we guess the DNS? How do we know the precise URL? We don’t have to. For general references to *wikipedia, *imdb or *apple, we wouldn’t have to enter full domains. The system “knows” which one we mean. And if there are multiple options for DNS resolution, like .net and .com addresses with the same domain name, the system should be intelligent enough to guess the most relevant one.

Admittedly, it leaves room for misinterpretation and failure. But if you haven’t noticed, so does the DNS system in general. Just because it works 99.9% of the time doesn’t mean it’s perfect. It’s just the best thing we have so far.

What * or ^ does is sort of expanding the options to play with the data entered. Suddenly you could sum up all references made to *apple, in combination with the tag #iphone. The wonderful thing is, you’d be able to distinguish the root of information from the subject.

With this, I already touched the technology layer. The benefit would apply to the social layer, which @bostjan touched in his response too.

It will take off once people find it accessible and easy to use, and if it really works in all directions (say, if Twitter picked it up and made it part of its system). Of course for that to happen, there would have to be more than one user actually using * or ^ for source reference.

The idea of this draft was, take the burden off the from the user’s shoulders. Don’t make users copy and paste long URL’s, trying to figure out how to squeeze everything in 140 characters. Don’t make them go to use external URL-shortening services. Let technology handle accessibility and let users focus on accessing, expressing and referring to information.

Great discussion. I have a point and a thought.

I thought is was interesting that you picked an asterisk—it made immediate sense to me because the Kindle for the iPhone (and I am guessing for the Kindle as well) does use an asterisk for footnotes. When you click on the asterisk, this takes you to the reference note, and provides you a link back to the text. Obviously this wouldn’t work for Twitter, since you would have the URL count against your character count and you would have to provide somewhere to house the note.

My thought is that the other two shorthands provided easy automated functionality, meaning, any tool can quickly and easily understand what to do with them. For the @ you can link to the user’s Twitter page and for the # you can link to a Twitter search for that word. This is great, because it provides a text entry short hand, and a tool can help the reader make use/sense of it.

This wouldn’t work as eloquently with references due to the need to include a link that won’t conform to the same standards. What would be cool though is if Twitter would let you enter something like: *http://bit.ly/LWEgV and Twitter would:
A). not count the URL chars against you
B). turn it into something else like perhaps just a * with a rollover title tag that told you the page title of the reference Web page

Just my thoughts.

Many comments here have been distracted by the Twitter aspect of the proposal.

Semantic representation of footnotes is critically important and needs a common form of implementation.

I am in favor of having multiple footnotes such as “^1 notestuff” and “^2 nextnotestuff”.

Also, footnotes serve as a “note” relating to specific portion of content. There should be a way to convey the meaning of the “note” in a manner that fits within and facilitates the semantic web AND a way to ties that to the specific piece of content. While many footnotes are “notes” about a reference source (for example “page 123 of X book” or “personal interview by author with Y person on Z date”) other footnotes are indeed “notes” that have semantic value and should be conveyed (for example “X publicly repudiated this view in Y speech delivered on Z date before W group” or “compare with X”).

@all why can’t we have #asterisk simply meaning *link? Who needs all this http:// when whe can generate new & www shared semantic standards?