Chalkface Blog

The World Wide Web Consortium officially launches the Semantic Web tomorrow. The Semantic Web project has been a topic of hot debate for several years now and I’ve been trying to apply the well-honed arguments from that debate to disentangle my concerns about Curriculum Online.

The Semantic Web has the endorsement of Sir Tim Berners-Lee, the Web’s inventor. The idea is to clearly categorise each piece of information on the web (pages, links, names, products, places…the list is very, very long) so that each piece of information can be more correctly found and used. Its proponents imagine a world in which you can take for granted that you can always find the information you need quickly and easily, and in which computers combine that information in new ways to reveal knowledge that was previously hidden. The best exposition of the dream I’ve come across is this very entertaining piece of future-fiction.

The project also has its detractors. This article by Cory Doctorow focuses particularly on the difficulties of getting web authors to enter the required data in an unbiased way – or even enter it at all. If you really want to get into the theory (again from a critical perspective) I recommend you read Clay Shirkey. An oft-cited example of metadata’s failings is the triumph of Google over all other search engines. Unlike its competitors, Google completely ignores the hidden metadata traditionally built into web pages to help search engines. Why? It is too easy for webmasters to ‘game’ the system by putting half the dictionary in there, thus polluting search results.

So how does this relate to Curriculum Online (COL)? First because it’s easier to understand the COL dream if you think of the current system as the first step towards instant, easy discovery and use of a huge range of educational resources. Second, because it’s also a lot more obvious why nobody’s using it.

Curriculum Online suffers from several of the problems enumerated in Cory Doctorow’s article.

1. The amount of information about each resource that’s demanded is huge; the workload of generating this information is so great that many resource owners have simply not done it.
2. The system is easy to ‘game’. Spend any time on COL and you’ll see the same resources popping up again and again in almost whatever search you try.
3. Despite all this data, teachers don’t generally search by the national curriculum metadata. A random browse of the Chalkface search log revealed these typical searches, none of which returned any results on COL:
• examples of oxymorons in romeo and juliet
• causes of the russian revolution student examples
• pictures of people using legal drugs

I remain completely in favour of the goals of the Curriculum Online project; for one thing it’s role in policing the ring-fence around eLearning Credits is essential. I do think, though, that COL’s masters at the DfES could serve teachers better if they were familiar with both sides of the Semantic Web debate.

Today marks something of a milestone as Paperless School is finally getting a long-promised feature; recursive marking. Since the beginning, a teacher has been able to annotate an essay or portfolio assignment and send it back to the student. Now, the student can edit the work and send it back for re-marking. The cycle can continue for as long as the teacher chooses.

From the technical perspective, this has been a difficult feature to achieve. The system was built on the assumption that an assignment would go through a simple linear sequence of being set, opened, worked on, sent to teacher, marked, and returned to the student. Allowing recursive loops within this sequence required some of the deeper code to be rewritten. This in turn led to a debugging phase that took a full two months.

So how come we made the mistake in the first place? My conception was based on my own school experiences a generation ago. Essays were written, then marked. Then we moved on to the next topic. Only when we launched the GCSE Applied Business scheme did I discover that education is now far more of a collaboration between teacher and student than it used to be.

I’ve learned two lessons from this. Firstly, there is no substitute in product development for actually being out there, with a live product that people need to do their jobs.

Secondly, schools are unique. In the same way that school textbooks are a unique publishing format, schools’ publishing platforms will not and should not follow non-school models. Prior to commissioning Paperless School, I evaluated a number of off-the-shelf LCMS (Learning and Content Management System) products with a view to adapting one. Thank heavens I didn’t go that route.

Update: This was originally published in February. I felt it was worth redating it to coincide with the re-release of the recursive marking feature. First time around we were forced to revert to a previous version of Paperless School due to another bug in the same upgrade.