RSS aggregators do a great job at parsing RSS feeds but there is additional metadata in the HTML posts in the form of comments and trackbacks and we'd like to parse this out.
Currently all the templating systems use the same general format for div and class names and its possible to write a parser for them.
Roughly the structure is:
<div class="comments"> <div class="comment-body"> Hey guys... this is my comment. </div> </div>
We'd like to push forward an "Open Comments" system mechanism to standardize on XHTML structure (similar to XOXO) so that RSS/Atom aggregators can also parse comments and trackbacks from HTML.
- Trackback parsing
- Comment parsing
- Q: Can't users just break their templates?
- A: Yes but they could also remove the autodiscovery tags and so forth. If they break them the parser should ignore the broken comments.
- Q: Why not just have comment feeds?
- A: Comment RSS/Atom feeds are one solution. There are some additional advantages with using HTML.
- Developers don't have to change much.
- If you already have the HTML you can index the comments with no additional IO. Imagine an email style pane for threading which was directly integrated within your browser.
- Why does it make sense to have a separate XML file when the data is right there in the XHTML file.
KevinBurton - Rojo Networks Inc.
Is "Open Comments" a good name?
Where to organize?
Where should we organize this effort? I don't want to do it on a vendor's site and would rather keep this as open as possible. Any suggestions?
onTrackback() @param author The author of the comment @param weblog The URL to weblog of the author. @param permalink The permalink to the comment post. @param content The content of the comment. onComment( String author, String weblog, String permalink, String content, Date date )
Existing Template Structure
<meta name="generator" content="http://www.movabletype.org/" />
<div class="comment"> <h3 id="c20648"> 1- <a href="http://thinkcorps.com" rel="nofollow">elvix</a> @ <a href="http://joi.ito.com/archives/2005/01/25/starting_the_car.html#c020648" title="Permalink to comment">January 25, 2005 02:25 PM</a> </h3> <div class="commentbody"> <p> i wonder if some wiki-like audio format might emerge to track /explore these collaborations? i.e., who added what and when? as collaborations like this become more commonplace (toss in mashups for good measure), could exploring the genesis of the music be as much fun as listening (or creating it for that matter)? kinda throws open the entire creative process.</p> </div> </div>
<meta name="generator" content="Blogger" />
<dl id="comments-block"> <dt class="comment-poster" id="c110594976009218540"><a name="c110594976009218540"></a> Anonymous said... </dt> <dd class="comment-body"> <p>I don't get how anyone could nail gun themselves and not know it for six days. Was his mouth open when the nail backfired or what? That's just plain weird.<br /><br />Also what's up with him not having medical insurance? I've had friends who have worked in construction and they got paid a lot of money per hour.<br /><br />Considering that the nail actually went into his brain, he is so lucky he didn't have bigger problems.</p> <p class="comment-timestamp"> <a href="#c110594976009218540" title="comment permalink">12:16:00 AM</a> </p> <span class="item-control admin-566838053 pid-2003574965"> <a style="border:none;" href="http://www.blogger.com/delete-comment.do?blogID=3187374&postID=110594976009218540" title="Delete Comment" > <span class="delete-comment-icon"> </span> </a> </span> </dd> </dl>
<a id="comment-26073"></a> <div class="comment"> <div class="w1"> <div class="w2"> <div class="w3"> <div class="w4"> <h3 class="title"><a href="node/15934#comment-26073" class="active">Funny, I was just talking</a></h3> <d iv class="author"><a href="user/4426" title="View user profile.">Boris Mann</a> - January 21, 2005 - 08:05</div> <div class="content"><p>Funny, I was just talking about this yesterday with <a href="http://www.willpate.org">Will Pate</a> of <a href="http://www.goodbasic.com">GoodBasic</a> -- he's trying to get together educators to look at Drupal as an alternative to current online learning systems.</p> <p>We had briefly talked through the creation of a "quizzes" module. The <a href="project/survey">survey</a> module is OK for short surveys, but it's meant for anonymous use and has some other shortcomings, and really the needs of quizzes is different enough to warrant a separate module. But, the <a href="project/forms">forms</a> module could likely be re-used as a supporting module.</p> <p>An additional permission would be "quiz marker" (imagine the TA of a professor marking). We had imagined that along with the quiz itself, the "answer template" would be created. There would be some questions (e.g. multiple choice) that would have just one answer and could be marked automatically, but text-answer questions would have, on the answer side, a "model" answer as well as a list of points to guide the quiz marker. Scores could be entered, as well as comments on rightness/wrongness.</p> <p>The quiz taker could view their own results. The answer template could be made public after the fact.</p> <p>For a set of quizzes, stats/reports could be generated on marking/grades. Export to Excel, SPSS.</p> <p>Some of your thoughts seem to be around an almost "question bank" type system. In which case, you would actually have a "quiz" node, as well as "question" nodes. Again, see survey/forms for similar interaction, except that "questions" would exist on their own.</p> <p>Like the profile module (or book pages?), you might want to have labelled quiz sections.</p> <p>OK, that's all I can remember for now. Will will chime in, and perhaps point some other education types here.</p> </div> <div class="links"><a href="user/login">login</a> or <a href="user/register">register</a> to post comments</div> </div></div></div></div> </div>
This format doesn't have a generator tag:
<div class="comments-body"> <p>Would like to get one invitation if possible. I already submitted my mail few days ago.<br /> Thanks in advance.<br /> </p> <span class="comments-post">Posted by: <a href="http://zecina.blogspot.com">Dalibor</a> on October 13, 2004 10:13 AM</span> </div>
Seems like they support comment feeds directly
<meta name="generator" content="blojsom v2.24"/>
<div id="comments"> <h3>Comments on this entry:</h3> <div class="comment"> <p class="blue"> Left on Wed, 2 Feb 2005 10:19 by Grumpy (<a href="http://hxr.us/grumpops" rel="nofollow">http://hxr.us/grumpops</a>) </p> <p> Very cool. And they'll be playing at a Borders only 4 miles from my house in late Feb. </p> </div>