What are categories?

A dictionary definition of category is, “A specifically defined division in a system of classification.” If you break that down, and consider what is meant by “a system of classification”, you’ll come to the conclusion that this refers to a highly structured, possibly hierarchical system. For example, to borrow from the classification system for living things, you might have a tree like “Animals -> Vertebrates -> Mammals -> Horses”. Any of those terms could be considered to be a category, and they fall into a strict structure.

What are tags?

In their most common usage, tags are keywords that you attach to a piece of content. Tags are “free-form”, which is to say that there is no formal restriction on what tags you attach. So, the basic difference between categories and tags is that categories are structured and tags are unstructured. For examples of popular web services that make good use of tags, see Flickr, Technorati, and Delicious, among others.

On the other hand, the help pages at Technorati define tags thusly: “Think of a tag as a simple category name. People can categorize their posts, photos and videos with any tag that makes sense.” They’re saying that tags are categories — so, are categories and tags different, or not?

How are categories and tags different?

If a tag can be thought of as a category, then how are they different? It really just comes down to how you use them. Typically, as we mentioned above, category systems are structured. In a given system, the categories are usually chosen with as little overlap as possible. Look at your typical mainstream news site — you’ll probably see categories like Politics, Business, Entertainment, Sports, Technology, and Lifestyle. The breadth of these categories covers just about any imaginable news story.

The problem arises when a story does overlap multiple thematic categories. What if you have a story about iRobot’s newest model of Roomba robotic vacuum cleaner — do you file it under Business, Tech, or Lifestyle? In many systems, you can assign multiple categories. But even in that case, there are often reasons to specify a primary category. Slashdot is a good example — if you look at their main page, each story has a single icon which represents the primary category. This is a quick visual indicator which helps you identify stories that might interest you. Then, if you click into a specific article, you’ll see the icons of other categories that the story fits into.

And in a strict category system, visitors using your search feature might use terms that don’t appear in the content or category names. What if Alice was trying to find the Roomba story, but couldn’t remember the name of the device or its manufacturer? She might try a search for “gadget”, or “carpet”, but maybe those terms were never in the story. This is where tags can help refine searches by giving additional keywords. If our story had tags like “business, technology, tech, lifestyle, gadget, floor, carpet, vacuum, cleaning, household, home, robot, robotics” and other relevant terms, then there are lots of extra hooks for searches to latch onto.

Which should I use?

There are some people who have discarded the hard structure of categories in favor of the free-form nature of tags. There are a few folks who prefer the clean thematic separation of a category system, and avoid the messiness of inconsistent tags. But most folks should feel free to get the best of both worlds. A while back, I started using my categories as tags. I create new categories on the fly, as seems fit. But I’ve gradually started to regret that decision. What I really wanted was to have both a category to specify the overall main theme for each post, and a set of tags to aid in searching.

In WordPress prior to version 2.3, the “category” system could be used as a “tag” system simply by the nature of how the user chose to express them. The difference was simply a matter of semantics. The same holds true with the new tagging system in WP 2.3, really — it’s all a matter of how you choose to use the underlying system. But the availability of a flexible taxonomy system now provides a way to clearly delineate a separation between “categories” and “tags”, and even user-created tertiary categorization systems. Previously, this type of flexibility was only possible through third-party plugins.

Whether you choose to stick with distinct categories, tags, or a combination of both, is really a matter of your personal preferences. But whichever path you choose, WordPress will be able to give you the tools you need to organize your content the way that you like.

ShareThis

Sandbox has a really good structure, with lots of semantic classes (and microformats) added at key points. This allows the clever theme designer to turn their site into a chameleon. You can theme based on the type of page being viewed (blog posts, single post, pages, archives, categories, search page, etc.). And you can theme certain posts based on their categories, their authors, or even by the season, or time of day when the post was published.

I’m hoping that we’ll see some really stunning designs that use these capabilities in interesting ways. There’s a strong chance that Sandbox will be included as a standard theme in a future WordPress release, which would expose your work to an extremely large audience. How’s that for some extra incentive? Do you think you have good CSS chops? Design your entry and get it submitted. Just make sure you get your entry in before the July 29 deadline. And note that the deadline is 12:00am UTC, which is July 28, 8:00pm EDT, or 5:00pm PDT.

FWIW, this site is running Sandbox (albeit, an older, and slightly customized version). But I certainly wouldn’t try to hold this site up as a paragon of good design. Make something better, okay?

ShareThis

The rel="nofollow" idea had good intentions: to give content producers a way to link to another site without implying that they approve of it. The way it works is that if Google, or Yahoo!, or any other service that uses this standard sees a rel="nofollow" attribute on a link, they will ignore it. They don’t follow the link, and they don’t count the link in the destination site’s ranking calculations. One of the main use cases (as in WordPress’ case), was to reduce the effectiveness of comment spam, because the spammers would not get any “Google Juice” out of the links. Hurray for our side! We just stuck our finger in the spammer’s eye!

Unfortunately, we also tweaked the eyes of our regular readers, most of whom could probably use a little of that Juice. “Oh well,” we said. “That’s just the price we have to pay for a little peace of mind.” Well, most of us said that. Some were adamantly opposed to the nofollow idea. Many of us knew that it was just a bandaid, and that it wouldn’t really deter spammers from trying, it would just reduce their ability to get high rankings in search engines.

At the time, comment spam was a pretty major problem on many blogs, and there weren’t many effective remedies. It seemed like all of us spent the beginning of each day going through our comment queue, manually deleting the garbage that made it through the gauntlet of whatever defenses we did have in place. So, nofollow was the last-ditch attempt to deny satisfaction to the spammers when our other measures failed. But it did not discriminate. It had no way to know whether it was de-juicing a good guy or a bad guy. (Eric Meyer had some good thoughts on this subject, BTW.)

These days, many sites have better anti-spam measure in place. Akismet has been very effective, and many WordPress users swear by Spam Karma 2. With measures like these in place, hardly any spams ever make it through to be displayed on your blog. And if they do, hopefully you delete them pretty quickly after they appear. So, that’s even better than just telling search engines not to index their links. They can’t index something that they never see in the first place, right?

With that in mind, I’ve installed Kimmo Suominen’s dofollow plugin here, and configured it to remove the rel=”nofollow” attribute from comment links after two days. The two day limitation is to account for the occasional hiccup where spam might make it through over a weekend, and I don’t get to delete it immediately. The important thing is that I’ll be giving back the Juice to the comments that get to stay here. If you’ve got a WordPress blog, and you feel like comment spam is under control on your site, I encourage you to do the same.

ShareThis

In other news, Microsoft is working on interoperability between OpenID and their CardSpace™ framework. This is going to bring a lot more attention to OpenID. Microsoft has tried entering this field before, with Hailstorm. But they tied it exclusively to their proprietary Passport service. This new collaboration will be building on the existing OpenID infrastructure, which is open and decentralized. Of course, this doesn’t replace Passport, it will just mean that Passport will be one of many identity services that one could use to log in to various OpenID-enabled sites, or that one could use an established third-party OpenID to authenticate against a CardSpace site.

And if that wasn’t enough, Simon Willison announced a new service called idproxy.net, which allows you to use your Yahoo! account as an OpenID credential. So, if you already have a Yahoo! account, you’re half-way to having an OpenID that you can use (with the trust level you have in Yahoo) to login to OpenID-enabled sites, via idproxy.net.

And last, but certainly not least, there’s the new videntity.org OpenID provider. What makes this one ubergroovy is that it has support for FOAF, vCard, and microformats like hCard and XFN. And it can both import and export multiple formats for the contact info. So, not only does it act as an OpenID authentication service, it can also serve as a general, multi-format identity provider. This service can store a lot of different profile information, and you have full control over which bits are public and which ones stay private (to be provided to other sites only with your explicit permission).

To get a better idea of what I mean by that, check out my profile at dougal.videntity.org. What you see there is just the information that I’ve chosen to make public — in my private profile I have additional information entered. Interesting to note here is that the links in the “Ernest knows” section were automatically imported from the XFN information here on my blog. It also imported some of my other info from hCard data on my site. I could also have imported information from my FOAF profile (if I hadn’t let it get so out of date).

If you haven’t really been following OpenID, or if you’re still confused about what it is and what it means to you, why not just try it out? Pick any of the numerous services out there, create an account (quick, while you can still get ‘myfirstname.someservice.com’!), and then use your new ID to login to an OpenID consumer site. Like this one. And if you have a WordPress blog, you can use the OpenID Delegate plugin and use your own URL as your OpenID. For example, if I want to login to an OpenID site, I just enter ‘dougal.gunters.org‘ as my login ID. I am then redirected to myopenid.com (which is currently my chosen delegate) for the actual authentication.

If you still want to do some more research, you might want to check out Planet OpenID, which aggregates the latest news.

ShareThis

The plugin serves two purposes: it lets you set your own title, instead of the “links for 2006-01-30″ format that del.icio.us uses, and it adds a rel="tag" attribute to the del.icio.us tag links (because I like microformats, and one day somebody might make a service that utilizes this information to aggregate tags in the wild across multiple tagging services).

The default post title is set to “Daily Links”. If you want it to be something different, change the variable indicated in the plugin code. I didn’t think it was worth the trouble to create an option screen for one little text string that you aren’t likely to want to change, once you’ve set it. It’s GPL licensed, so share and enjoy!

ShareThis

There are two additional features that I wish it had. First, it would be nice if you had some control over the title of the post. As it is, it uses something like “links for 2006-03-15″. This is redundant, because the date is already on the post, and I’ve got them going into a category named “links”. It would also be nice to be able to have more control over the semantics of the links. For example, I’d like the links to use the rel-tag microformat. I might just need to whip up a plugin for that…

ShareThis

This seems to be primarily aimed at electronic media, but I’m hoping that the primary authors might be willing to expand the scope a bit. I’ve already contributed a couple of links on that page and added an “Offline Media” section, which I hope others might fill in. If things move too slowly, maybe I’ll just go ahead and come up with some individual microformats that cover what I think I might need, toss them out there, and see what evolves.

ShareThis

I’m working on a movie review using the hReview format. I’m hoping that we’ll see one of the big search engines start correlating microformat data soon. I’d guess that Technorati might be the first, but I’m sure that Google and Yahoo! are looking at it, as well.

via: Eric Meyer

ShareThis

• hCard Creator
• hCalendar Creator
• hReview Creator
• hFolk 0.4

Basically, microformats are just organically grown rules to add application-specific semantics to existing XHTML documents. This is mostly accomplished by adding class attributes to your existing markup, and occassionally by adding additional markup (span & div) to indicate a hierarchical data structure. This description leaves a lot to be desired, but an example is worth a thousand explanations…

In the first restaurant review of the hReview example linked above, notice how some fairly normal HTML is enhanced by adding the hReview classes. By adhering to this standardized format, specialized web crawlers could take note of it, and aggregate the information into an uber review search engine.

That’s the incentive for people to use these types of microformats. Imagine if a Googlebot (or Technorati, or whoever) saw an hReview compliant review of a new movie on your site. It could correlate your review with reviews by other authors about the same movie. It could then present average scores, excerpts of reviews, links to the reviews, and other interesting statistics.

Maybe in my Copious Free Time, I can convert some of those hThingie Creators into WordPress plugins, and maybe even come up with some microformat proposals of my own.

ShareThis