I have made a few changes to the Sidebar Tabs plugin. Version 0.4 was uploaded tonight.

The most important change was a switch from Prototype to JQuery. This should make it less of a burden on load times and prevent the plugin from being compromised by plugins that bypass WordPress’s inbuilt script scheduling.

I have also added the facility to deal with sidebars where there are more tabs than the width of the sidebar can accommodate. It show now offer the option to ‘page’ the tabs where there is more than one row instead of simply wrapping around.

Finally, I have added a fade effect to the tab switchover.

You can read more about the plugin at the main plugin page here: http://www.wp-fun.co.uk/fun-with-sidebar-tabs/

Earlier today Miriam at WordPress Garage posted a quick challenge: how to display a list of authors who had posted in a particular category. I had an idea at the time but I wasn’t in a position to test the idea, so now I am home I have written a quick function to do just that.

The function

These two functions are best placed in your functions.php file or can be added loose to the category template file, either would be fine.

function get_authors_by_cat($cat_id, $sort = true){
 
	$author_id_array = array();
	$author_details_array = array();
 
	$cat_posts = get_posts('category='.(int) $cat_id);
 
	foreach ($cat_posts as $cat_post){
		if (!in_array($cat_post->post_author , $author_id_array)){
			$author_id_array[] = $cat_post->post_author;
			$author_details_array[] = get_userdata($cat_post->post_author);
		}
	}
 
	if ($sort === true ){
		usort($author_details_array , 'author_by_cat_sort');
	}
 
	return $author_details_array;
}
 
function author_by_cat_sort($a, $b){
        $al = strtolower($a->display_name);
        $bl = strtolower($b->display_name);
        if ($al == $bl) {
            return 0;
        }
        return ($al > $bl) ? +1 : -1;
}

There are two functions here, the first retrieves all the posts from the category and loops through them to make sure. The second function is a callback for the usort function. This allows the list to be sorted by the display name of the authors.

This could be done more efficiently by querying the database directly but I wanted to avoid that option as a lot of people are not comfortable doing that themselves.

Using the Function

To use the function in your theme you need to include the following in your where you want the list to appear (replacing the number 18 with the ID of the category that you want to list authors for.

<?php
 
$authordata_array = get_authors_by_cat(18, $sort = true);
foreach ($authordata_array as $authordata){
?>
 
	<?php the_author(); ?>
	<?php the_author_posts_link (); ?>
 
<?php
}
 
?>

Notice that in the foreach loop you should be able to use the normal author based template tags. You can find a list at http://codex.wordpress.org/Template_Tags

I have released the initial version of my plugin: Fun with Microformat Pingbacks. A plugin that parses the posts that ping WordPress looking for an hReview Microformat and uses the summary as the comment text.

This page has more about Fun with Microformat Pingbacks.

If you want to display a list of archive links in WordPress the way you do it is with the template tag: wp_get_archives. But what if you want a little more control?

Unlike bookmarks and categories there doesn’t seem to be a quick and easy way to retrieve an array of archive links as an array for later use. So what do you do if you want a list of archives exactly the same as that generated from the code below, but you want to insert something else into that list?

<?php wp_get_archives('type=monthly&limit=12'); ?>

WordPress has three functions that make it much easier to generate lists of archives manually:

• get_day_link( $day , $month , $year );
• get_month_link( $month , $year );
• get_year_link( $year );

These functions each return the appropriate permalink for the date provided to them. All you need to do is to calculate the dates you want to include.

The code snippet below shows the links for this and the previous 11 months and adds a class (selected) to the list item if the page it is displayed on is an archive from the month and year selected.

for ($i = 0; $i < 12; $i++){
 
	$the_date = mktime(0, 0, 0, date("m")-$i, date("d"),   date("Y"));
	$year = date("Y", $the_date );
	$month = date("m", $the_date);
	$selected_class = '';
 
	if ( $wp_query->get('monthnum') == $month && $wp_query->get('year') == $year ){
		$selected_class = ' class="selected" ';
	}
 
	echo '<li'.$selected_class.'>';
	echo '<a href="'.get_month_link($year, $month).'">'.date('F Y',$the_date).'</a>';
	echo '</li>';
 
}

This can be amended in a variety of ways to add extra functionality to a theme, including paging, listing posts inline, and even combining with Javascript to produce a dynamic menu of years, months and days.

If you’ve been reading this blog, or Reader Appreciation Project for the last few months you may have noticed an ongoing discussion in posts and comments about the usefulness and potential replacements for trackbacks / pingbacks. Put simply I believe that trackbacks and pingbacks are almost useless right now, and I want to change that.

The key problem is that the text that ends up on the page is, honestly, useless. The text is taken out from around the link based on characters with no context or semantics. Most of the time the result looks something like this:

[...]Person X has written a post about thing y, here is a link.[...]

This tells us nothing except that a link exists, and often interrupts the conversation.

To find a solution I decided that it would be necessary to identify the meaning of the link. A, ‘Go look at this’ link has no business being tracked, it is meaningless to every visitor to the page being linked to. There is no reason for the visitor to move in the opposite direction to the link. These then can be ignored.

This leaves posts that not only link, but also comment on the content. These are useful, add to the conversation, and should be tracked.

It seems clear to me that what we are talking about is semantics. We are trying to identify the meaning behind the content, and this is what Microformats were made for.

I think that the hReview Microformat is perfect for marking up offsite comments (posts with pingbacks). What is a comment if not a review? Better yet the format already features a summary class that can be used to quick and simply summarise the points for readers of the original post.

Is this a good idea? I want to test it out and that is what this post is all about.

I have written a rough plugin to parse hReview review pingbacks, based on the hkit Microformat parser, It hasn’t had much testing, and so I am hoping that you will help me by writing a post using the html below, an hReview.

If you are using WordPress it will ping me automatically as there is a link to this post, and the plugin will parse your post looking for an hReview, hopefully including your full summary as the comment text. I hope the html is fairly self explanatory.

<div class="hreview">
<p class="item">Andrew at wp-fun has asked bloggers to help him test out his <a class="fn url" href="http://www.wp-fun.co.uk/2008/03/18/microformat-pingbacks/">Plugin that uses Microformats as part of the pingback process</a>.</p>
<p class="summary">this is where you summarise the content, this will appear in the comments of wp-fun</p>
<div class="description">
<p>This is where you say what you think in full</p>
</div>
</div>

To help things along a little I have removed the normal comment option from this post so it is pings only. I appreciate any help you can provide.

Update: I have found a problem with the implementation that I don’t really understand. It seems that there is a problem with XPath when it uses normalize-space in that it sometimes treats spaces between words as whitepace nodes and deletes them. I need to research this a bit more as this is all contained within HKIT and isn’t something I have ever used, but for the time being it means the odd space may be missing from pingbacked comments.

Update: The problem with the spaces is actually nothing do to with XPath so, thankfully, I don’t need to get into that. The problem was actually caused by a combination of Tidy wrapping content, and removing spaces in the process, and hkit then removing the line breaks.

Things will be a little quiet here this week. I am working to get my theme finished off but also have some other, non-WordPress-related stuff to do so I probably won’t get time to post. To bide over the time here are a few questions I have been mulling over for a few days.

Time limited content

Over at Reader Appreciation Project a little while back I posted about weekend blogging. About the fact that I don’t but am disappointed when others do.

That sparked a discussion about how to educate readers to expect content on the weekends and time limited content is something that has since occurred to me as a way to do that.

Here’s the idea:

A post is published. Unlike regular posts the feed contains only a brief introduction and a statement that the post will be available to read for 24 hours.

If you click through and read it within those 24 hours you get to read the post, and you get a cookie which means that you can keep viewing it. So any conversations in the comments can keep going.

If you click through outside of 24 hours the post content is not available to you. You can register to get it when it is released, which could be in a week, (presumably the e-mail notification will also arrive on a Saturday), but you can’t view until then.

My question to readers here is: would this just piss you off? or would you get it and run with it? would it make you unsubscribe?

Plugin Architecture

The second question relates to my Fun with WordPress Plugins tool.

I have in mind a major overhaul for the next version, lots of UI changes, much more integration with filters and actions; however, there is one thing I have in mind for the actual back end code, but I am not sure if there is a demand for it.

I have it in mind to develop some objects for widgets, admin pages, and possible even actions and filters and keep these outside the main plugin class. The idea would be that these would function as a limited form of MVC, with each object including pages in specific foldersl; for example:

Your plugin creates a new admin page object, sets the parent, and registers it. It looks for a specified file within the html folder to get the content of the page, and within the code folder to get the processing instructions.

This invariably makes the plugin more complex, possibly even more complicated, but I like the idea of keeping the plugin itself as empty as possible.

My question on this is, what do you think? good? bad? do you have a better solution? Do you think there is a best practice for WordPress plugin architecture irrespective of its function?

I appreciate any comments you have on any points.

P.s. I am also on Twitter now.