Category: Uncategorized

I’ve pushed three new films live on SciCast: here, here, and here. Yes, I know the links to the experiment pages are broken, and that the ‘this link is broken’ page is itself broken, and that the draughts film doesn’t fast-start. We’ll get there. 🙂

I just hooked my phone to my PC for the first time in a few weeks, and tried the Nokia software update tool. Previously, it hadn’t recognised the phone model number, since Orange UK use a custom variation.

Good news: this time it agrees there’s an update, and is downloading 11.0.026. Which isn’t the most recent firmware going, but whatever.

Bad news: the N95 crashed during install, and on reboot appears to be unresponsive. Merde.

[update: leaving it a few minutes without power, it now starts up. I’m attempting to update again.]

[sigh]

[update 2: nope. Tried three times, and the phone crashes out each time, at the same point. Nokia forums are full of people having problems even with the latest 12.0.014. My advice remains: avoid the Nokia N95 at all costs, it’s a shambles.]

Flossie and I have a new game: send the other person a food parcel via Tesco delivery, and require them to blog what they concoct from the contents. The opening salvo in this trial of wit and culinary cunning was an unchallenging shot across the bows, but I sense a degree of escalation may be involved in my retaliatory volley.

So: a question to the web geeks here – how might we pressure Tesco to add an API to their service, so we can do BBC Backstage/Tesco TV dinner mashups? Or integrate Google Maps, get Tesco to deliver to GPS coordinates, and thus do on-demand drop-delivery geocaching.

Come to think of it, my phone has a GPS unit – why can’t I do operations like ‘I’m up this mountain and really fancy a bun. Tesco!’ ?

We took the decision to host SciCast films ourselves – rather than use YouTube, Revver, Blip, etc – 18 months ago, so it’s interesting looking back and trying to work out whether we were right. The list goes like this:

Quality. Old Flash video (Sorenson Spark) plain sucks. It was fine in its day, but that was 10 years ago. The frame-rate is poor, the compression blocks obliterate fine detail – even more significant when you’re making technical films – and it looks really rubbish when blown-up to project on a classroom wall, which is a typical use case for us. YouTube’s presumed move to H.264 will mitigate this, but see my previous post about high-quality stuff.

Rights. I believe YouTube changed their licensing package last year – I should explore this. But I’ve a sneaking suspicion that by uploading films you still grant YouTube some control. I can do that for films we own, but contributed SciCast films are under CC licenses, so I don’t have the authority to assign subsidiary rights except under Share-Alike. Which means contributors would have to upload directly. Which might be fine, but SciCast is also about editorial quality, not just sheer numbers. Tricky, and messy.

Advertising. Advertising around educational content is a thorny issue, and I’d rather duck it entirely. As I understand it, the UK ban on advertising ‘junk’ food to children applies to print and broadcast, though not the web. If that should change, not having control over advertiser focus would simply be untenable. YouTube have recently started running ads over video, too, which makes a lot of sense, but isn’t something that would be acceptable in a classroom environment.

Comment Moderation. I feel very strongly that a site aimed at children should be a ‘safe’ browsing environment. That means, crucially, pre-moderated comments – nothing gets published before it’s been read. You can now do this on YouTube, but the proximity to other content (as little as one click from something over which you have no control) makes it hard to maintain a buffer. So you end up building your own site anyway, and embedding YouTube videos – all you really gain is a clean upload mechanism. Which is significant, but not enough, because:

Blocking. YouTube is often blocked by school and/or Local Authority-level firewalls. Game over.

It’s easy to assume that YouTube ‘owns’ web video, and that doing anything else is swimming against the tide. I don’t hold to that, since I think video is easy and it’s what you do with it that’s hard. YouTube is the cable TV of the web – oceans of material with the occasional gem bobbing around. What I’m trying to build is more like BBC4 – targeted at specific audiences, and hence with a much higher hit-rate. I’m trying to build something that’s fun and useful.

…which is partly why I’m keen on a middle ground between a commissioning model (cf. broadcast), and a free-for-all (cf. YouTube). There aren’t many sites taking such an approach, but one prominent example is Current.tv.

We’ve been discussing – again – whether we should make ‘big’ versions of the SciCast films available for download, or just the embedded H.264 versions. By ‘big,’ I mean about 20Mb/minute, 640×480 resolution or higher, H.264 or WMV9 (VC-1?) – nigh-on broadcast quality, and in most cases high enough quality to reveal problems with the original camera rushes. Archival-quality big.

I’ve felt all along that it’s the right thing to do, and of course this is all Creative Commons-licensed so there’s nothing stopping me doing it on my own time and bunging the films into Archive.org. But yesterday I found myself saying something that ‘feels’ right:

We publish high-quality versions of the films because they allow people to do things with SciCast we haven’t thought of.

I’ve even got an example of this – somebody has a quasi-commercial idea that would rely on their being able to munge the films in specific ways, and they couldn’t do that without access to large-frame versions.

We should be pushing these ludicrous-quality versions up in the next week or so, though we haven’t yet worked out how to integrate them with the site, so it might be a bit of a kludge in the short term.

Excellent story about programming laziness relating to leap seconds, from the Risks Digest.

This article won’t be the last WordPress-vs-Movable Type 4 comparative review, but the comments thread is extremely useful. Credit to the respective product managers for showing up and answering questions too – to be fair, Movable Type staff have a long history of doing that.

…which should probably alert the wary reader to coming flakiness on this blog, when I finally get around to updating it.

In chat, Vinay and I just came up with a new web application protocol we’d like to see. Our options:

1. Do nothing about it.
2. Seek venture capital, then build it.
3. Build a slick website about it, write reviews of example apps, and see if anyone reverse-engineers the protocol. Thus saving us from having to write it in the first place.

I’m veering towards option #3.

Jeff Carlson has an excellent first look at iMovie’08, over at Macworld. Worth a read.

Having now had a little chance to play with the package, here’s my present thinking:

• It’s easier to lash a bunch of clips together in iMovie’08 than in iMovie’06. The new interface is very, very slick indeed.
• You hit the limits of iMovie’08 much sooner than you do those of ’06, mostly because it simply does less.

The issue for me is that you can’t cut dialogue-driven, classic ‘drama continuity editing’ style films in iMovie’08. You just can’t do it. You can do it in ’06, but it’s a pain in the arse. What I was hoping for was a revision to iMovie that helped mere mortals cut such sequences, letting them keep an audio track intact while trimming an edit from the mid-shot to the detail close-up, and back again.

Evidently, the famous diving-holidaying Apple video engineer either didn’t get that at all, or decided it was Way Too Hard. So, what we have is a tool that makes crashing clips together really really simple. This is, quite likely, what the world has needed for the last three years.

But as more and more people discover the joys of video, they’re going to hit the limits. What’s really interesting isn’t what iMovie’08 does, it’s what it doesn’t do, and where they’re headed with it. There’s no obvious reason why they couldn’t add a conventional audio timeline to the project window, and show video timeline length as at present. If they’ve really thought it through, then by iLife ’09 we might – finally – have the video editor for the rest of us.

In the meantime, there’s Premiere Elements. If you’re on Windows.

Man, did I call this one: Adobe Labs have released a beta update of Flash Player 9 that incorporates H.264 playback. Finally the video wars are over, and we have a winner: It’s not Flash, it’s not Quicktime, it’s not Windows Media, and it’s not DivX/XviD – it’s H.264.

For once, this is the right outcome. With YouTube encoding their entire back-catalogue as H.264 for iPhone and Apple TV, it’s been clear for months that they must be heading to the same codec on the main site. Now we know how – still in Flash.

What’ll be interesting to see is how they’re encapsulating the video. Done right, it might be viewable with either the Flash or Quicktime plugins, but Adobe’s initial demo embeds a Flash-only .swf file.

Still, this is great news. Around the end of the year we should start seeing a huge jump in quality for online video.

[update: They’re claiming it’ll play .mp4 and .mov. Wooohoo! See also this; .m4v (iPod/iTunes/AppleTV format), .3gp (mobile phone H.263), and so on.]

[update 2: here’s the full developer blah. Looks peachy so far.]