Hi,

If you can work out the logic for your PHP code, then you could add your custom code to your theme’s _preprocess_page function. I did this recently on a site to pull the keywords from both the node title and its taxonomy terms, and add the HTML output to a new page variable – $vars['meta_tags'] – which I then output in the theme’s page.tpl.php. hope it helps.

–starts here–

// This snippet goes in the theme's _preprocess_page function.

/* get site mission on all pages */
$vars['slogan'] = variable_get(‘site_slogan’, ‘Shariful is an expert on internet marketing, comfortable helping small and medium business’);

/* Build header meta tags */
$meta_tags = array();

// Use slogan for description
$meta_tags['description'] = array(‘type’ => ‘name’);
$meta_tags['description']['content'] = $vars['site_name'] .’ (Shariful) – ‘. $vars['slogan'] .’ is an expert on internet marketing, comfortable helping small and medium business.’;

// Build keywords from node title
$meta_tags['keywords'] = array(‘type’ => ‘name’);
$meta_tags['keywords']['words'] = array(‘this’,'that’,'the-other’,'miscellaneous’,'keywords’,'always’,'applicable’);
$node_title_words = explode(” “, $vars['title']); // get possible words from node title
$keywords_ignore = array(‘and’,'in’,'using’,'is’,'if’,'the’,'by’); // words to be ignored
foreach($node_title_words as $title_word) {
$title_word = strtolower($title_word);
// Check that word isn’t in ignore list, or already in list of default keywords
if(!in_array($title_word, $keywords_ignore) && !in_array($title_word, $meta_tags['keywords']['words'])) {
$meta_tags['keywords']['words'][] = $title_word;
}
}
// concatenate word list with commas
$meta_tags['keywords']['content'] = join(‘, ‘, $meta_tags['keywords']['words']);

// Hide horrible MS Image toolbar thing
$meta_tags['imagetoolbar'] = array(‘type’ => ‘http-equiv’);
$meta_tags['imagetoolbar']['content'] = ‘no’;

// Build HTML output
$meta_tags_output = ”;
foreach($meta_tags as $tag_name => $tag_data) {
$meta_tags_output .= ”.”\n” ;
}
$vars['meta_tags'] = $meta_tags_output;

–ends here–

The ‘description’ just uses the site slogan, and the ‘keywords’ uses some default terms plus words found in the page title (which is usually inherited from a node).

I’ve tried to remove some of the specifics related to the site I use this for, but still keep things comprehensible. Basically, I build an array of meta tags which are then compiled into a page variable which can be printed in the page.tpl.php:

–starts here–

[...]
<meta http-equiv="content-language" content="<?php print $language->language ?>" />
<?php print $meta; ?>
<?php print $meta_tags; ?>
<?php print $head; ?>
<?php print $styles; ?>
[...]

–ends here–

There should probably be some more checking in there for duplicate words, but hopefully it should give you some ideas…