Categories & Tags Mayhem (∴ SQL)

Aside

I was just about to write on a few things I’m reading – Michel Serres, Black Metal Theory mostly – and suddenly felt an urge to clean up, merge, rename, delete some of the 1300 tags on supernaut, when I noticed something dead weird: some of the tags were simultaneously categories. Much messing around and the inability to edit or change taxonomies caused me to have a look at the database, where I found, horribly, they indeed existed as both. I guess this is either a hangover from the Movable Type port some years ago, or a plugin that didn’t do its job properly around the same time. So I manually deleting 50 or so of these and their relationships. And then did some more merging and renaming. And now I must eat.

techno…blah…ti, or the meaninglessness of social networks

A while ago, I ditched a slew of social networking vampires that had ceased to have any relevance to me. Now there is only Technorati remaining, and considering its uselessness I’m fairly sure it’s on the way out too. I like the concept of social networking, and think for a while, up until say earlier this year, the idea was working or at least was throwing up some interesting practical usage on the internet, but like everything, it got subsumed into spam, and the infernal noise of vapid negative-content that seems to overwhelm every good idea.

So, displeasure with Technorati, failure to update when it gets pinged, missing posts, missing links, not listing tagged posts under their tags, tags full of endless volumes of blog-spamorama. Do I use it to search? Only if I’m bored. Against all the MySpace bullshit (and don’t expect me to link to you if that is what constitutes your ‘web presence’), and other assorted blah-i-ness of the increasingly unusable internet, I’ve noticed, for want of a better term, flow of information has an approximate bullshit content equal say, “Great Wall visible from moon“, or “Castrated Imperial eunuch discovered Rhode Island in 1421“.

Mostly stuff comes to me from emails, mailing lists, news feeds (of which I am a preternatural whore for), following links from these, and very infrequently from dry searching, which I suspect from the outset won’t turn up much useful stuff.

So in all of this, when the relevance of Technorati to me is minimal, and its utility to anyone reading this about the same, why am I bothering? Oh, probably laziness, caused by a certain degree of redundancy in that categories already do the job within my blog, and the Technorati Tags were meant to, I guess, like … connect me to the universe … dude …

A possible solution, that would require effort from me would be to recode (:ack: :cough:) the tags so they search supernaut. This would mean I’d have to be a bit more, I dunno, thoughtful in my choice of tags, and I’d hate to dig myself a hole where I felt obliged to go back over almost a thousand entries and keyword them all proper, like. Another solution is to just delete them altogether – quicker than the time it took to expectorate this mess.

Anyway, decision probably by Sunday, and I’m thinking about opening comments back up, though considering I closed them over two years ago, hacked all the code out of Movable Type’s templates, deleted the offending cgi scripts, torched a few churches and sacrificed goats and a lamb for good measure, I dunno how easy it it to extricate myself from this particular situation.

… edit …

A sweaty evening of recoding and I think it works, but y’know, maybe I see something that works and everyone else sees blllarrrrgh! Actually it’s probably the most ill-conceived few lines of code I’ve butchered since the bad days of Flash 5, but wtf? screw you and your standards compliant escaped ampersands. (Funnily enough, I did escape it. Who’da thot?) … (oops, “alt” should be “title”)

<h6>
<MTPerlScript>
my $keywords = "<$MTEntryKeywords$>";

if ($keywords eq "") {

print '<br />No Keywords for this post.';

} else {

my @split_keywords = split(/,/, $keywords);
my $split_keyword;
print '<br /><a href="<$MTCGIPath$><$MTSearchScript$>" title="search supernaut">Search Keywords:</a> ';
foreach $split_keyword (@split_keywords) {
print ' - <a href="<$MTCGIPath$><$MTSearchScript$>?IncludeBlogs=<$MTBlogID$>&search='.$split_keyword.'">'.$split_keyword.'</a> ';
}

}
</MTPerlScript><br />
</h6>

categories and stuff

I’ve changed some of the names of the categories to make them a little more representative of what gets assigned to them, as often they weren’t really a good fit, and I don’t want to make a new category for every second post. I assign multiple categories to every post anyway to make something of a collective mess of the idea of a category, and additionally am a slut for Keywords, which I’ve linked to Technorati.

So, books has become words, to cover anything that has words in it, magazines, fiction, non-fiction, e-books, blogs, and generally all things texty.

Film has become video, to deal with the fact I don’t often see bits of celluloid spinning by at 25 frames a second, but do see lots of moving pixels. It’s pretty much a linguistic change.

What was net-art, a repository initially for browser-based interactive art, installation network-driven art, and so on is now internet. This is probably the least accurate fit, but I don’t want to have two separate categories for internet stuff and net-art stuff.

Also, I’ll be changing the category which is currently extermination into one where all my art stuff goes. I’m just trying to come up with a name…

technorati tags and blog aggregators

(Geek warning, coding follows) T-Salon had a post on Technorati and their new blog aggregators, which are based around tags and include Flickr photos and del.icio.us links. Currently it’s not a particularly representative aggregator, as it just scrapes any blog on Technorati via the category for each entry. As most China blogs – or any that are on a specific topic don’t need to explicitly state that topic in every post, it tends to collect generalist blogs who have made a specific post about the topic or tag.

It is however, capable of supporting keywords as tags, and this is where it becomes both powerful and much more specific. Keywords are just that, they list the most relevant words from an entry, and provide a much more precise description of the entry subject and content. Also they are dynamic and are drawn from the entry itseld, unlike categories which are fixed and to which an entry is assigned.

Using the keywords to build the tags can all be done dynamically, as John’s Jottings shows in Movable Type. It’s a fairly straightforward process using Ben Hammersley’s keyword code and Brad Choate’s PerlScript Plugin.

ecto the blog client already supports keywords in an entry, and it probably won’t be long until some kind of keyword tag function is incorporated into it.

[edit]

I’ve spent about an hour installing the plug-in and getting it working, along with the MTKeywords plugin, which automatically fills the keywords metatag in individual posts’ head tags with keywords parsed from the entry, it doesn’t support non-western languages though, so I might have to get dirty. All pretty easy. I made some changes to the code from John’s Jottings though, preferring to split keywords on commas instead of spaces. This allows for people’s full names to be one tag rather than splitting into first and last names, which is fairly meaningless. I also used Ben Hammersleys’ if-else condition to hide the keywords if an entry doesn’t have any, coz I don’t have time to go back through 350 or so entries and keyword them right now. Overall pretty bloody simple. Code here for anyone who wants to see.

<MTPerlScript>
my $keywords = “<$MTEntryKeywords$>”;

if ($keywords eq “”) {

print “<!– No Technorati Tags to print out //–>”;

} else {

my @split_keywords = split(/,/, $keywords);
my $split_keyword;
print ‘<br/><a href=”http://www.technorati.com/tags/” title=”technorati”>Keywords:</a> ‘;
foreach $split_keyword (@split_keywords) {
print ‘ – <a href=”http://www.technorati.com/tags/’.$split_keyword.'” rel=”tag”>’.$split_keyword.'</a> ‘;
}

}
</MTPerlScript>

[edit] While I was redesigning my site, I reworked this because the layout was dependant on there being something in the Keyword line, and the “<!– No Technorati Tags to print out //–>” left that line blank. I’m not a perl scripter, so maybe this is common knowledge, but having “”, double speech marks around this line containing real text causes the search function in Movable Type to fail

print “<!– No Technorati Tags to print out //–>”;

The solution is to use single quotes to contain any text or html:

print ‘No Technorati Tags available’;