De-activating the plugin disables all the articles connected to it, as shown here, changes the title of your post, removes the tags, all the content (not just the feed content, but whatever you had to say about it, too) and the extract.

I wasn’t thrilled about the amount of control it was exercising on my blog’s content in the first place: dictating the title of the post, the tags and the excerpt. It also reverted these when I tried to change them. So I disabled it, whereupon it’s added insult to injury.

So, nice idea. Horrible in practice. Sorry.

Your old stuff — the stuff you wrote before, even your best stuff — mostly turns bad. It always did, but the Internet remembers. The churl.

Most people don’t bother about it. I, however, am foolish.

I’ve recently started using the Broken Links Checker plugin on this site. It finds the articles and sites you’ve linked to that don’t exist anymore. I did it because had a feeling that there was a need for some curation of my old articles:

• It seems like a bad service to readers to send them to content that you know isn’t there. If you click on a link that says ‘Ten things about X’, and you only get five, because the rest of it has disappeared, then you’d be disappointed, I’d suggest. Probably a bit annoyed with the person who sent you.
• I’m told that Google regards broken links with a stern eye and downgrades you accordingly. I want to be found (still need a new job, people!) and so this seems like a squandered resource.
• There’s a sense of personal and professional hygiene to this. They may have link-rot, but dammit, I don’t.

Anyway, I ran it and it found about 300 broken links in old posts.

Drat.

In a lot of cases, the broken link didn’t matter – it was just a case of extra information that wasn’t essential to the heart of the piece. Nonetheless, I get annoyed when I click on something and it doesn’t work; I expect you do, too. It wouldn’t be right to just leave it there.

In some cases, it ruined the whole article:

• Check out this research report – I think that… [Except the research report isn’t there any more and so readers have nothing to go on].
• Cool video from XYZ – pass it on… [Not so cool when it doesn’t exist anymore or has been removed].

So what to do about this?

Maybe, in an ideal world, I’d go back and either (a) find the Internet Archive cache of the old file and re-link it or (b) rewrite the post to explain exactly what the report said or what was so cool about that video, so seeing it didn’t matter.

But that isn’t going to happen: if I had extra time to spend on this blog, it would be to create more new posts, not fool about with stuff from four years ago.

So back in the real world, my options are (a) delete the post; (b) brief note of explanation; © ignore it; or (d) unlink the link.

I’ve mostly gone for (d) unlinking. In some cases, I have deleted: hey, check out this cool video you can’t see.

Shouldn’t you delete the post when the evidence or source no longer exists?

No. Because there’s this whole permalink thing to blogger culture. If you wrote something, then it should be there forever. We made a break with the ever-breaking links of other media outlets and decided that these records are set in stone. Links disappearing every five minutes was a bad phase for the Internet and we made the right decision. I agree with all of that, except if it means that something useless is there forever, because I was linking to a source that couldn’t care less about that whole idea.

And also, I have sinned enough. I have a confession to make. I changed the permalink structure of this blog a few weeks ago, rendering almost all inbound links useless. [Short version — I got some bad SEO advice that killed server performance — see this for good advice]. Mea culpa. If I knew more about WordPress and search when I started, I would have done it better.

Hehe. You were so dumb in 2006.

Another consideration. I certainly was (am). A lot of my early posts are naive and sometimes stupid to my and your 2010 eyes (not saying that never happens anymore). Should I wipe them to make me look cleverer? No. That’s OK, in a way. The blog is also a personal history, and stupidity plays a major part in that. In my case, anyway. If this was a company blog? Hmm. Well, maybe I’d make a few edits, especially if the old guy had left.

So the broken links are displayed with the <del> attribute, mostly.

picture credit: zen

And when I say ‘essential’, I obviously mean ‘used by me’ and ‘jolly handy’.

Comments and Spam

Akismet

Owned by the publishers of WordPress itself, you’ve already got this one if you have a WordPress blog. Akismet does a great job of detecting spam. Partly a crowd-sourced effort, since every spam comment rejected by users goes into the service’s database. Needs some backup, though, since you tend to get a lot of stuff to check on the ‘suspected spam’ list every day. Definitely keep it installed and active, though.

[addendum: I am trialling WP-SpamFree, as recommended. A previous version created a couple of false positives, so feeling cautious and checking the logs.]

Back-Type Connect

Helps to join-up the dots when discussion of your post takes place on Twitter or FriendFeed, or even another person’s blog who cites your own post as an influence.

Subscribe to Comments

Lets readers who leave comments elect to receive emails if further messages are posted to the current comments thread. A bit of a no-brainer. Can’t imagine why it isn’t built into WP in the first place.

Housekeeping

Broken Link Checker

The web is ever-changing and the stuff you linked to yesterday may not be there tomorrow. Keeping dead links alive is poor service to readers and weakens your credibility with Google. This plugin flags up broken links to readers with the [del] attribute and creates a handy to-do list (*yay*) at the back-end for you to re-link or delete.

FD Feedburner Plugin

This redirects requests for your RSS Feed to the Feedburner version, allowing you to collect stats from readers who prefer the RSS edition of your work and adds the results to Google Analytics reports. This version, rather than the official one, seems (at the moment) to work better.

Front-End Editor

Obviously, the moment you notice that glaring spelling error comes exactly 1.4 moments after you press the ‘Publish’ button and are ready for a well-deserved cup of tea. Going back into the admin interface is a pain. This (frankly amazing) plug-in lets you simply double-click on the post from the front — like the name says — and make your changes right-there in an AJAX-ey mini-edit box.

Google XML Sitemaps

Allegedly entices Google, Bing and the rest to more accurately and frequently spider your site for new content. Looking at my logs, they’re round here all the time anyway, especially those buggers from Microsoft. I keep it activated as much as a rabbit’s paw than anything else.

Redirection

This blog has been going for nearly four years now, if you include earlier variants on Blogger etc. I realised recently that I had something like 20 categories, with most posts belonging to 2–5 of those. Slimming it down meant merging and deleting some categories. This plug-in stops visitors (and search engines) getting frustrated by redirecting them to the new address.

WordPress Mobile Pack

A suite of plug-ins and themes that mean mobile users get a mini version of your site that will look OK on their phone’s browser and load quickly. My mobile traffic has risen by 5% since installing it last month and now equates for over 10%. ‘Nuff said.

WP Super Cache

Makes WordPress a lot faster for your visitors by serving them up pre-prepared pages. Should you get a traffic spike (haha) then your server will not disconnect you as quickly. This is good.

On Page

Similar Posts

There are literally dozens of similar/related post plugins (i.e. the bit that says ‘you’ve read this, why not look at these?’ at the end of posts), but this one is doing great service. Its rivals seem to either find very poor matches or take up so much processing power that they either fail or result in really slow page loads.

Sociable

The ‘share this post on social media sites’ thingy that appears at the end of articles. This one has the benefits of (a) looking nice and (b) not loading images and scripts from another server – something that will slow your site down considerably.

[addendum: I have switched to Light Social. Not as pretty, but it does render valid CSS and is even quicker.]

WP-PageNavi

Makes the handy-dandy this page out of X / go forward / back navigorator at the bottom of most pages. Looks cute and is apparently good for letting Google sniff-out your content.

Sidebar Widgets

delicious-plus

There’s a dozen or so delicious sidebar widgets available. This one seems to be the most reliable. Delicious (like Twitter) only accepts a limited number of calls to the source per hour, so you want a widget (like this) that doesn’t go to the source on every page load, but rather caches it away for an hour or so.

flickrRSS

Can be tricky to set up – or it was for me — you need to set the cache folder up on your server to be read/write. I’m not a photographer, but the benefits are definitely worth the effort, nonetheless – the official widgets from flickr will baulk half the time and delay page loads for the rest. Also allows flexible styling through CSS, as not seen here.

Linkable Title Html and Php Widget

Bit of an unwieldy name, but really useful. Lets you create widgets from XHTML and Javascript that stay in style with the rest of your site, without CSS, rather than sticking out like a sore thumb. And also allows PHP queries within a widget – like the Featured Posts widget above-right (“show last ‘X’ entries with the tag ‘Y’”, in that case). Okay, a bit nerdy. But you will want it at some point, I promise.

Twitter Stream

Again, there are a billion variants. Most poll the Twitter API on load which leads to delays and frequent failures. This one caches for a little while and doesn’t look like cat-sick.

WP-EasyArchives

Nifty looking archives with collapsible bits. Who wouldn’t want that? OK. Maybe not strictly essential.

WP-RecentComments

The recent comments sidebar widget that looks less-bad than the default. I’m of the opinion that letting visitors see that people do actually comment on your site encourages them to comment themselves. And that is – most often – a good thing in itself.

SEO Plugins

Nothing.

I’ve tried a few search-engine optimisation plug-ins and may do again, but my current thinking is that, unless you have adopted the worst WordPress theme ever and have no understanding about search and your content is terrible, you’ll be fine with what WP offers out-of-the-box.

[addendum: a couple of readers recommended All-in-One SEO. Digging around a little, I discovered the fast-developing young pretender Light SEO, which does the same thing but with less of an overhead. Bit hard to tell how it’s doing, of course, but it’s made an improvement to my title tags at the very least.]

Any recommendations?

It also lets you test your site using ready.mobi – this is available to anyone, with or without the plug-in. You may find it instructive to give a go on your own site(s). Apparently, I still have some work to do to make twopointouch standards-compliant.

Part of the suite measures the number of hits from mobile browsers. I found the results rather shocking:

7% of your traffic is currently from mobile users.
You’ve had 93 desktop hits and 7 mobile hits in the last 11.8 minutes.

Really? In that case, I have two things to say:

1. I’m really sorry it took me so long to install a mobile alternative to the full fat site. Why didn’t you say something?
2. I hope it looks OK. Let me know in the comments or on twitter if you have any issues or ideas. Can’t do anything about the quality of the writing, I’m afraid.

I also had a go with Mippin, which does a similar service. Upside: no configuration required. Downside: no configuration allowed.

I’ve produced an experimental social media news page using Feedly Mixes. You can embed this sort of thing into any site you like.

As you can see, it grabs and mixes up the content from selected RSS feeds – a list of sites covering the subject, as chosen by me. It then ranks the articles according to whether I deem a particular site important. Articles that I tweet or share in Feedly will also be included and take precedence on the page. Then it uses Google Reader’s statistics on how many people read, saved and shared articles to bubble up leading stories. Finally, it uses the age of the article as a fourth filter. The content is refreshed every 3–4 hours.

I’m not sure that my particular implementation is especially useful. The sort of person likely to visit this site almost certainly already reads the sites I have selected.

However, I think it could be a really useful tool in other circumstances. If you work in a particular domain – let’s say it’s construction – then it really is child’s play to create a page that gives a digest of the day’s news from the publications that deal with that area. That page could be set as everyone’s home page in your firm, or a link on the browser’s favourites bar, making sure the whole business is up-to-date with the latest news.

It’s better than the sort of pages you get at Alltop because you decide and curate the content sources – the construction page on that site is a good example of why you want this – a lot of the sources are US-based, which won’t be very relevant if, like me, you’re based in South London. There’s also no way I can see to flag particular stories as important.

A slightly more poetic use for Feedly Mixes might be the Interesting Pictures feed from flickr or ffffound. Maybe a personalised web comics page or a selection of new short stories.

NB: WordPress doesn’t deal well with the iFrames used to display the Feedly Widget on pages – you’ll need to install the Embed iFrame plug-in to get it to work.

I’ve been making some changes to this site recently: upgrading to the newest version of WordPress, choosing a new theme and fiddling with the layout.

One thing I discovered is that, unless you’re running a very minimal installation, it’s quite easy to run out of memory, even with only a handful of plug-ins. Versions of the software have become slightly larger since WordPress 2.6 or so, plus plug-ins have become more ambitious, not to mention the increased usage of rich content like videos and AJAX transitions.

Every time I tweaked something, I ended up with blank pages and an error message like this:

Fatal error: Allowed memory size of X bytes exhausted (tried to allocate X bytes) in /home /Your-Username/public_html/ the particular folder on line X

I searched around for quite some time for advice, and tried each of the following, before ending up at a working solution. They didn’t work for me, but I’ll leave them here since they might prove useful for someone else. Please be very careful, though: some of these ‘handy tips’ will break your entire site.

1) De-activating all but the most essential plug-ins. This is probably good housekeeping advice all-round, but didn’t actually make a lot of difference. The basic install was using 26MB of the 32MB allocated; with a full suite of plug-ins it was just over 28MB. Add some users, and you’ll quickly run out of memory — the extra 2MB from disabling plug-ins didn’t make enough difference.

2) Alter the value of php_value memory_limit in your .htaccess file. Doing this broke the whole site: YMMV.

3) Optimise your database tables using phpMyAdmin. Saved about 200KB. Woop.

4) Adding the line “memory_limit = 64M” in my php5.ini file. Made no difference one way or another.

What actually worked was this.

Open wp-config.php and find this bit:

/** Sets up WordPress vars and included files. */
require_once(ABSPATH . 'wp-settings.php');

Change it so that it reads like this:

/** Sets up WordPress vars and included files. */
define('WP_MEMORY_LIMIT', '64M');
require_once(ABSPATH . 'wp-settings.php');

Touch wood, that’s done the trick. The site’s still as slow as molasses, of course, but at least I’m not running out of memory.

I’ve recently installed a relatively new WordPress plug-in from the good folk at Backtype.

This is what it does: it scans the web, including social networks like Twitter and other blogs, for mentions of your post and draws those mentions in as comments on the post. This is a good thing in many respects. It helps to give readers a sense of the whole debate, not just the point of view of people who manage to stop-by your site and leave a comment. It links to all its sources.

So why am I fretting?

In some ways, it feels like cheating. When people measure website engagement, the ratio of comments per post is a key indicator. Someone who gets 100 comments on every post is clearly more *cough* important than someone who gets 1 comment, when it comes to blogging and such.

My recent post on the #PRDebate event that we produced at NMK a couple of weeks ago ostensibly has 17 comments. Yet I know only a couple of those were from people who came to this site — the rest are collated from Twitter and other people’s blogs.

Have I stolen those comments in some way? In a way, I have. If other bloggers are competing to be the most popular, then they lose if they don’t use this plug-in. It seems like they’ve got less engagement than an entirely-equivalent-in-every-other-way blog that does use it. Which isn’t true and so that’s not really fair.

However, until people start complaining, I’m going to carry on. Conversations about blog posts are distributed nowadays – you’re more likely to get a reaction on Twitter than your own page; people reference your post on other people’s blogs. Blogs are less important as destinations as people dip into the flow rather than visit sites. Creating a resource, and multiple resources, to let people get the whole picture is a valuable thing. The engagement metric based on comments/posts is in some ways flawed since if people are discussing the post elsewhere, then that’s equally (more!) important than them discussing it on your own site.

Be delighted to hear others’ thoughts on this – comment below or via your own blog, twitter or anything else, it seems.

It takes more written words than it’s worth, so here we go with videos from people who have more talent than me and have taken the time.

…with Blogger, which has become a great platform for casual blogging nowadays, and is certainly the easiest place to start, much under-rated…

and WordPress — much slower, deeper video, but a much more powerful platform, IMHO…

There’s a little bit more to it than the videos show. But not that much.

In my opinion, buy a domain name from anyone (e.g. Bluehost, the service I use — they have been fine for the last two years and are dirt cheap) and map it to your wordpress.com account, or even better install wordpress on your own hosted server. (If you’re still with Blogger at this point, no problem — your host can map to that, too. WordPress can also import all your blogger posts if you want a fresh start).

Most hosting services, including Bluehost, but also GoDaddy and most of the rest, make that absurdly easy. Look for ‘Fantastico’ in their feature list. That’s a service that will allow you to ‘auto-install’ a load of website software, including WordPress. No technical skill required.

If that last couple of paragraphs sound like a foreign language, then a simple wordpress.com or blogger.com account will be a start. Just go to the address and open the account. It is really easy, as the videos describe. It’s also easy to just have a go and then delete the whole thing: then there’s no embarrassing past to be unearthed by someone. Just have a go and then delete the whole thing.

Once you’ve got some technical skill, or got someone in who has, you can customise to your heart’s content. Priorities: (1) more, better content; (2) that it doesn’t look like shit. Contrary to what the whole advertising and marketing world thinks, content is more important than appearance on the Internet. Get great content and no-one will care about the appearance. Get mediocre content and a great appearance and no-one will care, full stop.

Not looking like shit is an important, secondary priority. First choice (and you went for the hosted service option above, right?), choose from and implement one of the thousands of free themes at http://wordpress.org/extend/themes/. Second choice, get one of those and customise it bit yourself with Photoshop and a decent CSS editor. Option3: get your designer to create a custom job. If they can’t work with CSS, fire them (seriously).

Even if you’re not planning on blogging anytime soon, go into these services and make sure you have a decent user name, not like the user1238237@hotmail.com you ended up with when you were late to the party with web email. If people are searching for you on the web, they’re searching for ‘yourcompany’, not ‘yourcompany12921134’. You’ve probably already got .com and .co.uk names, but have you got the delicious, flickr, youtube, stumbleupon, etc. names? Gotta get them all. If you can.

Reserve good user names on every online service you can think of, even if you aren’t planning to use them straight away.

Look forward to comments, or get in touch if you want me to sort this sort of thing out for you.

The newest thing is native support for tags, as well as categories. Everything is miscellaneous, of course (check this fantastic video  of a presentation by David Weinberger on this theme (57 minutes) if you haven’t caught onto this idea yet), so tags ought to supply a folksonomy for posts on the site, as well as the taxonomy it’s always had. Windows Live Writer apparently supports this natively now through the Keywords field at the bottom of the editing window. Except it isn’t quite right, since I am the only person who can add or edit those tags. I wonder if there are any plug-ins out there yet that allow this to be carried to the next level — to let users add their own tags to articles in a way that sits inside the system rather than through widgets and third parties like del.icio.us?

On the theme: it’s back to Cutline again, for the time being, since it’s one of the few that supply native support for tags. Yes, it’s easy to hack existing themes to provide that, but I really can’t be bothered.

Things to do:

Haven’t quite finished with the masthead — there’s still a couple of generic pics in there and I think the title of the site needs to be a few points bigger.

Add tags to previous posts. I used to use the very popular plug-in Ultimate Tag Warrior. It used too much CPU ultimately whenever this site became moderately popular, and I had to drop it. Nonetheless, a lot of the posts on this blog have already been tagged. Some sort of import for that would be more than welcome.

Database may or may not be screwed. Oops. (Proper IT person: “Of course you backed up your database before performing such a drastic operation?”; Me: “Uh… yeah… sure I did.”). Updating old posts to add tags results in a MySQL error, though the edits are carried out nonetheless. Have to google it and see where I may have cocked things up.

This is what it looks like, for people in the know:

WordPress database error: [Table ‘twopoint_wrdp1.wp_post2cat’ doesn’t exist]
SELECT cat_ID AS ID, MAX(post_modified) AS last_mod FROM `wp_posts` p LEFT JOIN `wp_post2cat` pc ON p.ID = pc.post_id LEFT JOIN `wp_categories` c ON pc.category_id = c.cat_ID WHERE post_status = 'publish' GROUP BY cat_ID