This has GOT to be a joke...

The W3C -- my absolutely positively most favorite standards body ever -- has just come up with an XML namespace for emotions! I must say that I fully support this specification... who on earth would ever want to type something as confusing and ambiguous as this:


When we can do The Right Thing™ and use XML instead:

    <emotionml xmlns="">
            <category set="everydayEmotions" name="Amusement" />
            <intensity value="0.7" />

Ugh... If this were released on April 1st instead of November 20th, I would have been amused. But now I'm just plain sad. As Wearehug said, "It is becoming increasingly difficult to distinguish W3C specs from Onion articles."

(Hat Tip Aristotle)

A little too far

Since 1986, the National Institute of Standards and Technology (originally the National Bureau of Standards) has produced and periodically revised a standard for the interchange of fingerprint (and, subsequently, other biometric) data. Initial versions of the standard used somewhat cryptic field separators.

However, in 2008 they released an alternate, XML-based version of the standard. While the XML strings are longer than the legacy strings, they are more readable.

Unfortunately, it appears that this emoticon namespace is not significantly more readable than the original emoticon - in fact, it's fair to say that it's less readable, especially if your native language is not English.

Oh well, we can't stop progress, can we?

Re: A little too far

What drives me nuts is that everybody keeps trying to use XML for things it was never meant for... I prefer JSON formatting for data interchange. Its much simpler, and you can always embed it in an XML document at a later date if you ever find a use case where JSON won't cut it.

When you look at the

When you look at the contributors, it makes more sense.
Text to speech and avatar animation.

Re: When you look at the

There's a perfectly acceptable, universal, and terse solution for this already... emoticons.

Just because you can do it with XML, doesn't mean you should.

"You can drive your car with your feet if you really want to; that doesn't make it a good idea!" -- Chris Rock

Emoticons aren't universal or always acceptable

I disagree.
Firstly, this is not a replacement for emoticons (or emotions). They are not pitching it as such. They don't expect people to read the XML, any more that people read HTML. They expect computer programs to interpret the XML and render it in an appropriate format (might be visual, might be audio).
Secondly, emoticons are NOT universal. There's no standard for them, so if you were trying to instruct a computer program to interpret them, there's a good chance you'll get some wrong or miss some. Given your comment regarding your choice of captcha, it would be interesting to see how screen text-to-speech converters interpret emoticons and plain punctuation.
Thirdly, yes emoticons are terse. Often good. Sometimes bad when you need to be totally unambiguous.

Whether XML or JSON or some other format is best is open for debate. I find XML bloated for many tasks. Its best use is generating publically automatically consumable information where the recipient isn't constrained to use every component of the data.

If you don't have a use for this, fine. Some people obviously do think they will use it, and are agreeing a standard to work with. They've probably spent a lot more time thinking about it than you have. It doesn't hurt you, so why belittle someone else's efforts.

if only it WAS a joke...

Is there anything that w3c, XML and all the rigmarole of the last 5 years of so-called and self-promoted "web standards" stand for other than:
waste of space?

They're looking forwards to

They're looking forwards to web 4.0.

The semantic web will be old school by then, instead we'll be surfing the 'emotional(TM)' web.

You heard it here first folks.

Your article made me:

Your article made me:

<emotionml xmlns=""> <emotion> <category set="everydayEmotions" name="Amusement" /> <intensity value="0.7" /> </emotion> </emotionml>

Shock and awe

I almost wrote "LOL" in response to your post, but I'm afraid I didn't mark it up properly and it failed the emotional validator...

When you were laughing…

… You had forgotten to look at the context and the authors. This is not a W3C Specificiation (it is not under TR/). It is the work of an incubator group, you know like Google labs, Mozilla labs, etc. An Incubator Group is usually started by a few members to look a round a specific need they have and explore what they can do.

The real issue is that Aristotle doesn't know how to read. :)

ps: the other joke is your captcha

don't knock the captcha...

I was looking for a captcha that didn't discriminate against the blind... and this one was the most effective at blocking SPAM :-P

Also... this may just be an Incubator Group... but the fact that it made it this far -- and has had W3C support since 2006 -- is at the very least tacit approval by the W3C of the goals in this working draft.

I'm not the only one tho thinks EmotionML is nutty... I'm wondering if the W3C is trying to distance itself because of the backlash?

I belittle for multiple reasons...

If the W3C were just a research organization working on technologies that might be useful, I would cut them some slack. Likewise, I don't mock every RFC that comes out... except for RFC 2549 of course. However, the W3C has a long history of cramming obtuse XML ontologies down developers' throats, without doing one shred of analysis about the practical limitations of their technologies. Then we have to replace something that works with something that doesn't work, and then find workarounds, all for the joy of being "standards compliant."

"They've probably spent a lot more time thinking about it than you have."

They probably spent more time thinking about it than I ever will... that doesn't mean they're right.

Human emotions are not as cut-and-dried as their schema would like. Lie detectors and facial recognition tools use a plethora of arrays to create a "guess" about the state of a human. They both have verbose output formats with raw data alongside multidimensional statistical analysis, all which give hints for later analysis tools to find errors in their judgment... what does EmotionML have? One single confidence attribute which goes from 0 to 1.


This format is entirely unsuitable as a data interchange format for describing human emotions between systems that analyze human emotions. At best its good for a high-level report suitable for human consumption... in which case, there are much more suitable formats... Like a paragraph of text.

Secondly, emoticons are NOT universal. There's no standard for them, so if you were trying to instruct a computer program to interpret them, there's a good chance you'll get some wrong or miss some.

This is basic internationalization and localization. Some people use 1/31/09 whereas other use 31/1/09. Some people use :-) while others use ^_^

Its a totally solvable problem, provided sufficient context... and it requires a lot less effort than forcing everybody to encode their emotions in EmotionML.


Thanks Bex. Checking out the W3C Incubator Group is a smile-a-minute*.

I do think, however, that the EmotionML crowd was outdone by the Common Web Language committee ( Theirs is a much more ambitious, impractical and unusable mission. I'm sure it will be at least as successful as Esperanto.

* hmm, does that present a problem for EmotionML? I don't see a way to encode duration or frequency in their markup.

history repeating itself...

I remember when they called XML "Esperanto for the web" with no sense of irony... Yep, if everybody just all agreed to speak the same language, the world would be a much better place! And RDF is the pinnacle of XML because it has subjects, predicates, and objects!

yep... and if everybody agreed to all wear gray and eat gruel, we could streamline this pesky "free will" thing and make everybody happy.

Incubators are a space for experiment

Note that the W3C incubator groups are a space for experiments -- which, I'll submit, is an important thing to have around a standards body. Launching one takes three interested member companies.

So, please don't mistake an incubator report for a standard!

Re: Incubators are a space for experiment

Good point... I may have have been a tad harsh, mainly because I've suffered a lot of anguish because of 3WC standards being crammed down my throat... so my initial reaction is one of bitterness.

I'm all for experimentation. As I tweeted earlier, in order to have good ideas, you need to have many ideas. However, I'm a tad surprised that this experiment has been going on for 3 years, and it doesn't appear that the designers chatted much with their intended audience... because if they had done so, this data format would look a lot different.

Recent comments