What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 133 Stack Exchange communities.

I'm quite satisfied with the Stack Overflow sites, but there's a (sort of) minor feature I'm missing, which would be rather easy to implement: microformats.

In essence, microformats serve the purpose of adding semantic, machine-parseable information to websites. Some more common uses are

  • vcard-like information about authors
  • Tagging events with in-line iCal calendar markup
  • Establishing relations between websites, users, and links.

The general principle is that microformats are added by using the class attribute of HTML elements. This way, there's absolutely no problem with compatability, as browsers just ignore classes they don't know what to do with.

The hCard and hCalendar microformats allow embedding of vcard and iCal information into the flow of the page, for example. Also, there are definitions (along with XDMP profiles)

What I was thinking of that could benefit the sites:

  • Adding `hCard` info to profile pages as well as any place where user information is displayed.
  • Using `rel="tag"` notation to designate tagspaces.
  • Using `rel="me"` notation on profile pages to designate the user's home page. (A useful feature for identity consolidation, which slots in quite well with OpenID.)
  • Not sure about this one, but using the `hAtom` format to define an in-line Atom feed. This could also be enhanced by simultaneously adding the Atom Publishing API and autodiscovery headers for said API.

A simple example, using my profile badge:

<div class="user-info">
  <div class="user-action-time">
    answered <span title="2009-07-14 12:35:48Z UTC" class="relativetime">Jul 14 at 12:35</span>
  </div>
  <div class="user-gravatar32">
    <a href="/users/11716/towo">
      <img src="http://www.gravatar.com/avatar/fabc12ff04817d5768e0388e94a0c868?s=32&amp;d=identicon&amp;r=PG" alt="" height="32" width="32">
    </a>
  </div>
  <div class="user-details">
    <a href="/users/11716/towo">towo</a>
    [SNIP: badge info]
  </div>
</div>

Would simply look like:

<div class="user-info vcard">
  <div class="user-action-time">
    answered <span title="2009-07-14 12:35:48Z UTC" class="relativetime">Jul 14 at 12:35</span>
  </div>
  <div class="user-gravatar32">
    <a href="/users/11716/towo">
      <img class="photo" src="http://www.gravatar.com/avatar/fabc12ff04817d5768e0388e94a0c868?s=32&amp;d=identicon&amp;r=PG" alt="" height="32" width="32">
    </a>
  </div>
  <div class="user-details">
    <a class="url fn" href="/users/11716/towo">towo</a>
    [SNIP: badge info]
  </div>
</div>

Some further reading on MFs:

share|improve this question

4 Answers 4

I think I'd rather have a decent way of getting the information in JSON format via specific queries - I don't see much point in adding extra payload in the HTML being generated when 99.9% of views aren't going to care about that data.

share|improve this answer
1  
Valid point, but IMHO, it's not really much. The hCard (as presented) adds 33 bytes, which isn't really that much (cosindering that users could just have, like, long user names). Other microformats are similar, so even when going radically overboard, you wouldn't even get close to 1kb extra on average. –  towo Jul 24 '09 at 16:44
1  
33 bytes per user, per answer, per page view. How many people would actually benefit from this extra cost, vs people benefitting from tools built on top of an API which has no cost to "normal" page views? –  Jon Skeet Jul 24 '09 at 16:59
1  
Micro formats send us right into web 3.0 (semantic web). Why not adding it if there's going to be mandatory in a few years? –  backslash17 Jul 24 '09 at 19:32
    
@backslash: Well for one thing, because there are more useful things the team can be doing. It's not like there's a shortage of feature requests which would actually make a difference. –  Jon Skeet Jul 24 '09 at 19:43

I am implementing the basic hCard for the user page.

I am kind of opposed to adding hCard classes to the user <div>, as stated in the question above, since

  • it would then appear on (almost) every page of every Trilogy site
  • it doesn't seem particularly useful to have this all over the place
  • the bandwidth cost adds up pretty fast if it's a core part of the user <div>

… but on the user page, I think it's worth the minor cost. Still not convinced it's all that useful, but it's easier to implement than is to argue about it. Which is I suppose an argument in favor of the hCard microformat.

Note that we already supported rel="me" and rel="tag" for many many months, and yet we got no credit for that.. :(

share|improve this answer
3  
Hey, I love the rel="me" in profiles, in fact, I want it also in careers! meta.stackexchange.com/questions/29668 –  jmfsg Dec 8 '09 at 0:03

Maybe I'm just too unfamiliar with the idea but I don't see the overall point in having to go about implementing it. I agree with balpha in that I would much rather have an API released that allows me to jigger up whatever fun toys I want.

share|improve this answer
    
The point here being that an API is a local thing (at least no-one has defined common cross-site query APIs yet), whereas the mfs work indiscriminate of the actual site in question. –  towo Jul 24 '09 at 16:56
    
I don't exactly see what there is to gain from these... –  Ian Elliott Jul 24 '09 at 16:59
1  
The ability to automatically parse information regarding the site and being able to interpret its content is the main purpose of microformats. It allows parsers (like browser plugins) and crawlers to identify certain information and use it. –  towo Jul 24 '09 at 17:04

From the latest podcast, I get it that we're not terribly far away from a Stack Overflow API. When that's done, there's no more need for something like this.

share|improve this answer
1  
Not a valid point, I'd say, as an API requires site-specific consideration by the user, whereas microformats can and are implemented as simple browser plugins or parser programs that work without resorting to a specific API. –  towo Jul 24 '09 at 16:46
    
Hmm... I can't say I totally agree with you, but I do see your point. –  balpha Jul 24 '09 at 16:50
3  
SIX TO EIGHT WEEEEEKS –  Jeff Atwood Nov 4 '09 at 8:43

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .