The idea is that you create a workflow for posting to social media sites. The input can be email, application or bookmarklet, and it allows you to build datapaths for different bits of information – which can then be posted automagically to your various social media accounts (Twitter, Delicious, Flickr etc.)

The really clever bit is that you can pass the data through various other services on the way. In the first one I built, anything received (in this case at the email address associated with this particular workflow) is routed to Delicious and Twitter, but the Twitter posts are sent to TinyURL first.

This means that over at Google Sightseeing we’ll be able to use the Delicious feed to bookmark the original URL, whilst simultaneously posting a short URL for our Twitter followers.

In the example above, I have additionally routed the Twitter and Delicious URLs back into a email, which is received by whoever sent the original message – giving confirmation that both posts were completed successfully.

The developer has posted a video showing something even cleverer – he uses Evernote’s automatic OCR technology to create tags that are applied to the image as it is added to Flickr. Genius!

We’d like to see Wordpress support added, and also found that the bit.ly module is a bit broken, but if the developer keeps adding more services, and perhaps more importantly, more functionality – then Tarpipe could become an absolutely essential tool in the online arsenal.

According to our FeedBurner stats, Google Sightseeing recently hit a milestone in terms of number of subscribers; we now average over 100,000 unique subscribers.

Live updating figure

Unfortunately the number of people actually reading the feed on any one day is likely to be far below that, but it’s still seems like a pretty massive number to us. Because FeedBurner numbers are public¹, for any site that uses the service, you can use a free service called FeedCompare to chart their subscriber number against anyone else that uses FeedBurner. Which is how I created the following graph:

(The red line is Google Sightseeing, and the regular drops were due to a long standing bug that incorrectly returned a very small amount of subscribers)

This shows the number of subscribers reported by FeedBurner over the past 24 months, for Google Sightseeing (red), GEarth Blog (blue), Google Earth Hacks (green), and Google Maps Mania (purple).

The big jump in reported numbers back in February ‘07 was when Google took over and Google Reader result began to be shown in the figures, but what’s really interesting is what happened to the Google Sightseeing feed immediately after that…

As you can see, once the Google stats were included, our figures began a steady rise that has continued on exactly the same trajectory ever since. The figures for the Google Earth Blog and Google Maps Mania have remained about the same, whilst GEarth Hacks has had several large jumps – this probably makes sense, as when a site gets a big link from somewhere you might expect a sudden increase in subscriber numbers².

Given the relative flatness of the other graphs, why is it that Google Sightseeing has had such steady growth over the same period?

Although we can’t be sure, and of course the high quality of Google Sightseeing promotes people to tell their friends, but I think this may well be due to Google Reader’s suggestion feature.

So, if you read Google Sightseeing, how did you find it?

But, in the most recent versions you can get there with just one click, all you have to do is add the toolbar button!

You can find the button in the View menu, under Toolbars > Customize…. Drag it onto your toolbar, and inspect away, with two less clicks to get there.

The icon is awful, and doesn’t blend in with the new Firefox 3 theme on the mac, but it looks slightly better when you install the GrApple theme.

Tony presented the plugin (originally as a Flock extension) during last year’s Refresh Edinburgh event, and we were very honoured to see that he’d even built in support for Google Sightseeing sights to be easily loaded into the map.

Since then he’s ported it to a Firefox extension and added even more features. I’ve been using it for months and it got to the point where I’d forgotten it wasn’t a standard part of Firefox (which may explain why I took so long to write this post…)

You can install the plugin from its homepage.

–

OpenID is a “decentralized identity system”, which is like a single login that you control. You authenticate to a server which you trust, and then decide which details that trusted authority passes onto the site you’re trying to access.

The OpenID website recently relaunched with a Wordpress powered blog, which (of course) accepts OpenID logins for commenting.

If you’d like to add OpenID comments to your WordPress site, then I’ve outlined a couple of your choices below. Be aware that I’ve only tested these plugins with Wordpress 2.3 on PHP 5, so your results may vary!

WP-OpenID+ 1.0.1

If you search Google for “wordpress openid” you’ll find many abandoned projects such as Wordpress OpenID Plugin and OpenID For WordPress. Avoid these, as they’re way out of date.

WP-OpenID+ 1.0.1 is available from the official Wordpress plugin repository and does work, although it requires more effort than your usual Wordpress plugin.

The code appears to be a later version of Will Norris’ plugin which is in turn a “fork” or Alan Castonguay’s abandoned plugin. However, the supplied zip file is incomplete and this is what you must do to get it working:

1. Download and extract the zip file. You should have a directory called “openid”.

2. Within this directory, edit the file “openid-registration.php” with your favourite text editor and find lines 12-13, which should look like this:

   define ( 'OPENIDIMAGE', get_option('siteurl') . '/wp-content/plugins/wpopenid/images/openid.gif' );

   define ( 'WPOPENID_PLUGIN_VERSION', 1.0.1 );

   You need to change “wpopenid” in the path to “openid” so that the image references within the plugin work. Then wrap the 1.0.1 is quotes (like this: “1.0.1″) to prevent PHP falling over about the unquoted string.

3. This OpenID plugin relies on the PHP OpenID Library which, despite what the README claims, has not been included. You’ll need to download the latest release from the 1.x.x branch (not 2.x).

4. Extract the PHP OpenID Library files, and then copy the “Auth” and “Services” directories into your “openid” directory from Step 1.

5. Now you can upload this directory to your webserver and install it as normal.

6. If everything works then edit “openid-registration.php” one last time and find line 20:

   define ( 'WORDPRESSOPENIDREGISTRATION_DEBUG', true );

   Change “true” to “false” to switch off the verbose debug messages.

Well, that was a total faff! Be aware that, even once you’ve made the above modifications, the plugin may still have issues! When I installed the plugin here on rotacoo.com the settings were always reverting to the defaults, which was annoying.

If you’re also having issues with this version, or couldn’t be bothered with all the effort, then you might consider the alternative method…

Wordpress OpenID

Will Norris has continued development of his “Wordpress OpenID” plugin (formerly known as “Wordpress OpenID Plugin+“, and the basis of “WP-OpenID+” mentioned above) and is aiming to release it as “version 2.0″ sometime in October.

If you’re feeling a little crazy, and just can’t wait until the plugin is officially released, you can grab the latest code from the subversion repository¹. The repository code is quite stable, and is used on the official OpenID site.

Everything that is required is included in the package, so there’s no need to change any files before uploading to your webserver. Just make sure you checkout the trunk into a directory called “openid” (not “wpopenid”).

Is there some GMP in your PHP?

When you’ve successfully installed one of the above plugins you will find an “OpenID” page on your Wordpress options panel. This should have a status output that lists various configuration and debugging information for your particular install.

Make sure you check the output for whether or not you have GMP support complied into your copy of PHP. Although the plugin lists the lack of GMP support as a warning, it should almost be a fatal error – as commenting on the blog via OpenID takes a painfully long time without GMP support.

GMP is the “GNU Multi-Precision Library”, which is used by the OpenID scripts for arbitrary precision arithmetic (calculations on potentially huge numbers) within the cryptography.

Unfortunately most webhosts do not provide PHP with GMP support, and without it these calculations take a very long time.

If your site is hosted by Dreamhost then you’ll need to recompile your own PHP (which can be a very involved process) and then vote for GMP to be included by default.

Here at Rotacoo we run our sites on a Slicehost VPS running the latest Ubuntu release, but even that doesn’t have PHP-GMP by default (or available via apt). Instead, we had to recompile PHP, and nobody enjoys re-compiling things.

Conclusions

OpenID support in Wordpress is messy at the moment but it is possible, and will hopefully be a little easier once “Wordpress OpenID 2.0″ is released.

Let’s also hope that the plugin’s name sticks and doesn’t get any more confused with extra plusses, hyphens, abbreviations, etc.

Popular bloggers who aren’t comfortable re-compiling PHP are likely to have issues with the lack of GMP support in PHP installs. This will hopefully get better over time, as either webhosts include GMP in the PHP installs by default, or somebody finds a quicker way to do the calculations without it.

Google Sightseeing will be accepting OpenID comments soon, once we’ve finalised a few other changes.

• There are currently 1000 posts and 10,998 comments, contained within 197 categories
• We’ve posted 60 site news entries
• We’ve had to move host 4 times because we kept outgrowing them
• Our most popular category is Buildings with 196 entries
• Our most popular country is the United States, followed by the United Kingdom, Canada and then Germany
• Our most popular state category is California, with 99 entries
• We currently have 11,315 RSS readers through Feedburner
• Akismet has protected the site from at least 83,351 spam comments since we installed it last year

Visitors

In May of 2005, just one month after the site was launched, Google Sightseeing had 196,228 unique visitors. We only installed Google Analytics in November of that year, so the following chart only begins half-way through that month.

Daily Visitors, 13th Nov 05 – 31st Jan 06

Daily unique visitors as reported by Google Analytics (Click to enlarge).

In September 2006 we posted the infamous topless sunbather, which (thanks to a major Digging) was seen by over 60,000 people in one day.

The enormous interest in the story meant that over the course of the month we clocked up over 1.7 million page views, making it our busiest month ever (even despite our extensive downtime). The second huge peak was our followup story, the Top 10 Naked People in Google Earth (again getting picked up by Digg), although it’s only higher because we’d moved hosts by that time and the server was better set up to handle the traffic.

The graph represents a staggering 3,194,852 absolute unique visitors to Google Sightseeing since the 13th November 2005.

Browsers

When we first launched nearly 2 years ago, we inititally got discovered through del.icio.us, which was brand new at the time, and frequented by many “early adopters” – which might explain why in the first month, only 48.2% of our visitors used Internet Explorer, and 35% used Firefox.

Amazingly, the growth of Firefox has meant that today’s more balanced stats show Internet Explorer accounting for 55% of all our users, and yet Firefox still makes up 35%! Of the 55% who use IE, 31% have already upgraded to IE7.

Comments

We’ve had an incredible 10,998 legitimate comments on the site (most of them correcting our facts and/or spelling), and here’s the top ten commenters based on the email address supplied:

1. Alex (admin) – 306
2. rob – 215
3. cookie monster (under various aliases) – 197
4. James (admin) – 130
5. Tim – 120
6. Keith – 98
7. Luke – 87
8. gIMpSTa – 68
9. Peter – 66
10. Propaganda – 64

So that brings us to the end of this celebratory first annual report. Thank you to everyone who helped us get this far, including all those who blogged or wrote about us, everyone who linked to us, and especially everyone who contributed on the site through comments or submissions.

See you in another 1000 posts!