Archive for the 'Wordpress' Category

WordPress 2.7 – wow!

Just upgraded to WordPress 2.7 and I’m impressed. A new, clean and fast interface with extensive AJAX and even Google Gears integration. Yummy!

(0)

Apologies for black out

Apologies for the black out of this site for some readers. I was caught out by a failure with Bad Behaviour (the WordPress spam/robot blocker I make use of). You should be able to read and post comments again now.

(2)

Blogs, references and hyperlinks

Those of us who have been to university generally appreciate the importance of good referencing [Wikipedia]. (This is especially true for those of us who must admit to having been rapped on the knuckles for being a little sloppy in this regard!) One thing that has been bugging me lately is precisely how to reference sources in a blog.

Mention referencing, in the context of cyberspace, and many will instantly think of hyperlinks [Wikipedia]. The ease of creating a hyperlink has, unfortunately, created a tendency to create too many hyperlinks, bad hyperlinks, or both. A classic example is Tim Roll-Pickering’s blog. Whilst Tim has an engaging writing style he does use a lot of hyperlinks and these often lack context. His worst sin is creating meaningless hyperlinks like “click here” or “this”. Chromatic explains why this constitutes a grave error in Good Hyperlink, Bad Hyperlink [Onlamp.com].

Of course a hyperlink need not be just for the purpose of referencing. Web pages can, and should, be interactive documents and hyperlinks present a powerful way not only of referencing sources but also of allowing a reader to ‘drill-down’ and explore individual concepts and ideas covered. There’s a good Boston Globe article [Cool Cat Teacher Blog] on this.

Back to referencing: I stumbled across A quick guide to referencing [The Blog Herald], and it seemed fairly reasonable. The trouble is that it deals only with the easy part – when to create a hyperlink for referencing purposes. We’ve already dealt with bad hyperlinks of the “click here” ilk. I’m troubled by this quote from the author:

The link can be either on the name of the author, or on the short reference “a good question” or the long part “how much Yahoo! paid for Flickr”. There are no strict rules on where to base your link on an indirect reference.

Whilst this does accurately summarise the current state of affairs on the Web it is also problematic. Many search engines use the anchor (the text associated with a hyperlink) to help rank the page cited in their search results. So, whether you base your hyperlink on the author, reference or quotation does have an impact. I think, as far as possible, it should be the reference that is hyperlinked. I know some of my geekier friends will tell me that we needn’t worry as search technology improves but the fact is that we are not going to get a semantic web overnight. Moreover it is important to have some consistency to the user experience of working with hyperlinks.

Recently I’ve been guilty of taking this “hyperlink the reference” approach too literally. You may have noticed that many of my posts have involved a long reference in square brackets. Take, as an example, the post Why are only Microsoft’s competitors allowed to innovate? The use of hyperlinks as references in there is clearly wrong. It is unsightly and detracts from the subject matter. In retrospect the real problem is failing to understand that whilst a hyperlink need not always be a reference, a reference need not always be a hyperlink (at least not to another page). When we have long things to cite, perhaps the best way to reference, without compromising writing style is to make use of good old-fashioned footnotes in web pages [SmackTheMouse.com].

Interestingly I have found a footnotes plugin for WordPress [elvery.net] although I have not yet had the time to investigate it. I’m a little concerned that with no current standard on how to markup footnotes in HTML use of the plugin may limit the forward compatibility of my blog database with new blog software.

One last thing. I’ve long noticed a practice amongst some posters on Slashdot to put the name or domain of the web site you are linking to in square brackets after the hyperlink (when the link is to page that is on a different site). I think that this is good practice, especially as isn’t always obvious from the anchor the source that a hyperlink is pointing too (especially if it is written in the form of a reference). Whether to use the site’s name or name is subjective but should wary depending on length and easy of understanding.

I would appreciate any thoughts anyone else has on the matter. Hard and fast rules may not be in the spirit of cyberspace. Nevertheless it might not be a bad idea to develop a collaborative set of guidelines for the blogosphere. Any takers?

Getting WordPress to talk OpenID

OpenID is a very good idea.

Most of us have to remember far too many usernames and passwords for far too many insufficiently important web sites. This is particularly true now that so many people have blogs (and so many of them insist on logging on in order to avoid comment spam). Even if we always use the same username and password it is a pain to have to register each and every time. This means that we are less likely to comment on a blog, for example, if it requires registration.

OpenID is a solution to this problem. It allows you to use a single online identity across many web sites. As an example, you can use an OpenID from any site that complies with the OpenID standard to login to post comments on LiveJournal. Your details secure because you don’t actually give your password to LiveJournal to login to. Instead, once you enter your OpenID on the site, it generates a link back to your own OpenID site (for the sake of argument let’s assume it’s your blog). Your blog will then check that you are logged in (and prompt you to log in, if not). It will then ask if you are happy to trust the referring site (LiveJournal) and, if so, will send a message back confirming your identity.

In summary your registration details are kept secure and not shared with anyone – but you do get a single identity that you can take with you when you visit different web sites. You can even associate an avatar with your OpenID which can be used on those web sites.

So whilst it is a good thing, why is almost no-one using it? The main problem, as always, is lack of support. LiveJournal, as far as I’m aware, is the only major public site supporting OpenID. And the support in self-hosting blog software packages is pretty primitive too. My own efforts getting it to work with WordPress have proved frustrating.

After spending many hours analysing lines of code (and Apache access logs) I finally managed to get Singpolyma’s OpenID plugin for WordPress to work as an OpenID server. Frustratingly it came down to the value of a single parameter – see my comment on Singpolyma’s blog if you’re interested.

I’ve now posted my first comment on another site (Bonnie’s LiveJournal) using OpenID, So now that’s working okay, I’ll have to look to modifying K2 to support using OpenID for comments on this blog.