Archive for May, 2008
I just discovered the benefits of creating custom Firefox search plugins the other day, having hooked up a Firefox search plugin for our internal MediaWiki which I shared with my colleagues. The search plugin integrates with the MediaWiki apis which can be used to provide suggestions. This suggestion functionality is pretty neat, since it allows you to quickly scan what pages are available on your wiki (without hitting the search button) and then select what you’re looking to go directly to your chosen page. A quick click-to-install link can also be provided for your team. Follow the steps below to find out how.
- Create a base-64 encoded 16×16 pixel image to appear next to the search engine input box so it’s always clear which search engine you’re searching. You can use an online utility such as Antonin Foller’s
- Create an OpenSearch description XML file and include your base-64 encoded image, e.g.
<OpenSearchDescription xmlns="http://a9.com/-/spec/opensearch/1.1/"
xmlns:moz="http://www.mozilla.org/2006/browser/search/">
<ShortName>my wiki</ShortName>
<Description>my wiki</Description>
<Image width="16" height="16">data:image/x-icon;base64,
(put your base 64 encoded image here)
</Image>
<Developer>Ian Homer for bemoko</Developer>
<InputEncoding>UTF-8</InputEncoding>
<Url type="text/html" method="get"
template="http://wiki/index.php?title=Special:Search&search={searchTerms}"/>
<Url type="application/x-suggestions+json"
template="http://wiki/api.php?action=opensearch&search={searchTerms}"/>
<moz:SearchForm>http://wiki</moz:SearchForm>
<moz:UpdateInterval>8</moz:UpdateInterval>
</OpenSearchDescription>
where http://wiki is the base URL to your media wiki installation.
- Host this description XML file on a server of your chosing.
- Create and host an HTML page that links to the absolute URL of your hosted description XML file. The link must use the AddSearchProvider JavaScript API as demonstrated below.
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en" dir="ltr">
<head>
<title>search engine loader</title>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<script type="text/javascript">
function addEngine(engineURL) {
if (window.external && ("AddSearchProvider" in window.external)) {
window.external.AddSearchProvider(engineURL);
return false;
} else {
alert(error_opensearch_unsupported);
return true;
}
}
</script>
</head>
<body>
<div id="wrapper">
<a onclick="return addEngine('http://host/searchmywiki.xml')">
install my wiki search engine plugin
</a>
</div>
<body>
</html>
- Open this HTML file in Firefox and click on the link to install the plugin … and of course circulate the URL to this HTML page with your colleagues so they can do the same.
… now even less excuses for people to say that they “can’t find it on the wiki”.
I have been using MythTV since version 0.17 a couple of years ago. Back then the installation process was for the brave/foolhardy only.
I built the computer, plugged in my TV card (a Hauppage PVR-350) and installed Linux. Then the fun really started. I eventually found Jarod Wilsons Myth(TV)ology pages with some excellent instructions to follow. After a lot of fiddling and playing around, I managed to get a TV picture displayed – a joyous moment!
The comments from my family were around the lines of ‘why not just buy a TiVo??’ – at the most frustrating times of my install, this was a good point, but when done it felt much more of an achievement than just handing over the credit card.
The system worked well, with an uptime of over a year until I decided that I needed to make some ‘tweaks’. The tweak I had in mind was to get a DVB card so I could watch and record UK freeview channels.
“Just a quick upgrade” I assured my wife. She was not convinced. At 2am when I finally gave up and went to bed, I wasn’t convinced either – it was not going well.
Next day I discovered Mythdora, which seemed to be a complete ‘Myth-in-a-box’ installation. Now, normally I don’t like these types of solution as I like to do it my own way, but I was getting desperate now. I downloaded the image and created my install DVD. I decided on a fresh install rather than an upgrade and just preserved my recorded programs directory.
What an amazing process – install, reboot and, lo and behold I could view and record the Freeview channels. I was a convert. There was still a lot of fiddling to do, but nothing to the level I had been used to. Only problem was that my remote control wasn’t working, so we had a keyboard in front of the screen. Amazing how you get used to not having to get up to change channels like we had to before remote controls.
I was happy with this arrangement apart from the remote control. Everything worked perfectly. Unfortunately every forum I tried insisted that the remote control worked on their setup – so I was getting nowhere with that.
A few weeks ago I noticed that Mythdora were beta-testing the next release, Mythdora 5 which included a new version of MythTV and a new kernel. My only hope for remote control happiness was to upgrade to a new kernel as the video4linux drivers had been improved, so I signed up as a beta tester and duly received my link.
I downloaded and installed the new software and, incredibly, my remote control worked first time! So I now had a fully functioning system.
Nearly….
I have a VIA EPIA M10K motherboard which has a VIA graphics chipset (CLE266). MythTV will use the XvMC extensions to pass off some of the MPEG2 decoding the graphics chip. This means that CPU usage when showing a recording or watching live TV is about 20-30%. Unfortunately this doesn’t appear to be working in the new version. If I choose a playback profile the uses XvMC, then the frontend crashes (saying that xv-blit is not available). I can use the ’slim’ profile, but the CPU usage is about 80% for playback now and the nautilus process wakes up at regular intervals, using CPU and causing the playback to pause.
This is not a huge problem to anyone else except me. I know it used to work, so will not be satisfied until it works properly again. I have tried many forums (fora?) to try and find the answer to this problem, to no avail. Many problems look like the same issue, but none of the answers solve it. It must be something to do with the way the frontend has been built, but it looks like XvMC is compiled in.
So I have a mostly functional system, with TV, Radio, News, Weather, Internet Streams, DVD, Picture gallery, access to the PVR over the Web etc. Was it worth it? Yes, if only to be able to say I don’t have to use MS media center!
If anyone passing reader knows how to get XvMC working, then I will be a truly happy man!
Once I have a stable version, then I will work on making the web interface mobile. Imagine being able to control the system from your phone…..
I’m really pleased with our new website. It gives the clean design and warmth that gives you an idea of what bemoko is all about. We’re all really up beat about the next few months and we’ve had some fantastic feedback on what we’re doing. Thanks to all those involved so far.
If you haven’t had the chance to meet up with us, feel free to get in touch and say hello. We’re normally at the momo london events and I’m personally at the MEX conference in a couple of weeks. Really looking forward to that especially having been awarded one of the momo scholarships.
I’ve been pretty impressed with the Google Reader which I started using a few weeks ago, especially since it bridges over to the mobile service so cleanly. The mobile site hasn’t got the most flashiest of interfaces, but it is real sticky and genuinely useful. Now why do I think it is so good? Because it does what I want it to do:
- View news articles from feeds I find interesting. I can see the next 10 headlines on the radar and drill down when I need to. I am even getting to the point where I prefer reading these headlines on my mobile as opposed to on my mac. I know I have the luxury of screen size of the E61i, but whether I’m sitting back on my sofa or waiting in line, I can quickly get my news fix.
- Mark articles as interesting and share them. I even share them with this blog (see in the sidebar of the homepage of this blog).
- Filter news based on my mood. I’m either looking at whatever (I’ve got loads of time and am just browsing), feeds I really rate (I’m just quickly make sure I’m not missing out on happenings), mobile related (Let’s find out what the buzz is) or obscure feeds (let’s check through some of the more offbeat news)
So I’ve got into the stage of configuring my feeds and reading trends at my mac and reading the news on my mobile, playing each medium to it’s strengths.
It’s not all rosey though. Some feeds provide lots of content, but others just provide a sentence relying on a click through to the main site. The click through experience is appalling, whether or not I go through google’s own transcoder. Most sites are not mobile ready and just take too long to load and the google transcoding UX is hit or miss as to whether the content I want is on the initial page of the transcoding. At least google give me the choice of which one I prefer – I’ve opted for the transcoded version (for the time-being), but to be honest I don’t click through often, since the experience is typically poor. Also, if the feed doesn’t provide a decent amount of content, then I’m starting to remove them from my reader. I actual would encourage feed providers to include with a decent body of content with the ad (text or image) embedded. I don’t mind seeing the add in the middle of content that I’m enjoying reading. As a service provider myself this does throw up some challenges for analytics 
.
I took a quick look at the iPhone AJAX UI which gives you an idea of UIs we may all head towards. It was only partially compatible with my E61i though. Furthermore we can only move towards such UIs if it ready does improve the UX – i.e. not slow and not unnecessary eye-candy that distracts from my primary ask – I want to read relevant news.
Good work Google & thanks.
At last Vodafone is providing more attractive data tariffs. It had to be coming, and the sooner the better. There’s still better rates on the market, but it’s great news for me (I’m with Voda UK) and the 17million other subscribers. The barrier of entry for mobile web usage in the UK is dropping.