I have been in Cambridge for 10 days working on Scholarly HTML with Peter Murray-Rust and team, plus assorted remote participants. I think the emerging consensus says that Scholarly HTML is a set of practices and conventions for using the web for scholarship way that will reduce friction for people and machines. The nascent format Scholarly HTML is the pivot point for tools that will be used to create documents, package research objects and consume content.
Scholarly HTML (should it take off) is relevant for publishers who create content for the web – making it more useful to their readers. It’s relevant to authors who want to produce the highest quality meaningful documents they can and reduce the time and effort they spend re-working content and re-entering metadata, and it’s relevant to scholars who are trying to deal with vast amounts of material out there on the open web, and behind various paywalls and barriers, including the often impenetrable prison of the PDF format.
In this post I wanted to report some progress on simple tools for creating Scholarly HTML. I will focus on Chemistry. I ask:
How do I refer to a molecule described using Chemical Markup language in a way that a reader will get a nice view of the molecule, while a machine can tell that what I am linking to is chemistry?
How do I do this declaratively, so that I am not baking-in to my document any particular behaviour? I will describe some progress towards tools below?
What tools would help?
Before getting to the examples, a bit of more general discussion.
We are in the process of defining some core guidelines, and setting out conventions for dealing with scholarly semantics such as citations. The techniques we’re documenting have been drawn from lots of previous work, particularly RDFa and micro-formats. Martin Fenner has a blog post setting out some of the benefits – I have minor quibbles with the examples he has chosen be he articulates some of the benefits clearly.
There has been some push-back for example:
TAC_NISO Todd Carpenter
TAC_NISO Todd Carpenter
If all Scholarly HTML is to take DCMI terms, existing ontologies & layer upon #HTML5, what’s the “there” there? Value being added? @rtennant
I don’t think we’re duplicating work but trying to make it easy to find, packaging it into a single set of guidelines and providing a locus for tool builders. For example, this morning we have been talking about how to take the Object Reuse and Exchange spec and provide a single, simple way to use that in an HTML page so that page can aggregate resources into a scholarly object. Frankly ORE needs all the help it can get to drive adoption.
A lot of what we are doing amounts to providing guidelines. Writing guidelines that tell people how to apply existing RDFa techniques in a particular way is not ‘duplication’ it’s documentation. Maybe we could think of this effort not as “HTML for D—ies” but “HTML for Scholars”.
The way I see it we are taking the opposite approach to the one taken by Perl where Larry Wall’s mantra was “There’s more than one way to do it”. On the web, there certainly is, but for Scholarly HTML to succeed I think we need to say to people:
There’s one good (enough) way to do it.
Here’s a meaningful way to do it.
Here’s a simple way to do it.
We meed to make things easy for people to copy and paste examples into their blogs, easy for developers to code tools, and easy to document. So, Scholarly HTML is not RDFa, which is hard and very broad and very complicated, but it does use fragments of RDFa that should make sense to RDFa processors (at least when RDFa 1.1 arrives with support for full URIs in properties and relations etc).
So, to the example.
I want to include some chemistry in my blog. Now, remember this is just one tool-chain. We know it won’t be suitable for everyone, and we’re hoping that our friends at Microsoft Research and the WordPress blogging tool community can make all this much easier for users. With that in mind let’s go through two cases which would be suitable for early-adopters who want to embed a molecule in their blog in a way that should interoperate with other systems, and not cause problems in the future. Two cases:
Where I am using plain-old WordPress or similar,.
Where I am using a Scholarly-HTML-ready publishing system (which I am here).
We start out the same in both cases. Here’s what you do:
Grab the Crystalize.me bookmarklet. Drag the link to your bookmarks bar.
Go to some Chemical Markup Language you want to talk about. Anything I am likely to say about it will be meaningless, of course as I was kicked out of chemistry 101. As I am sponsored by the Murray-Rust group I get my CML from Crystaleye. http://wwmm.ch.cam.ac.uk/crystaleye/summary/acta/c/2010/01-00/data/dn3132/dn3132sup1_II/dn3132sup1_II.complete.cml.xml
While the CML is showing in your browser, click the Crystalize.me bookmarklet. Don’t click it now, or you will create a bad thing something claiming to be chemistry that is actually a blog post. The result is something like this (plain-text ‘cos if I link to it the tool-chain turns it into CML): I have http://ontologize.me/?tl_p=http://www.xml-cml.org/convention/crystallographyExperiment&triplink=http://purl.org/triplink/v/0.1&tl_o=http://wwmm.ch.cam.ac.uk/crystaleye/summary/acta/c/2010/01-00/data/dn3132/dn3132sup1_II/dn3132sup1_II.complete.cml.xml
If you resolve the above URI, you are looking at a page at my ontologize.me site. From here you can copy and paste a bit of HTML to include in your blog post if you know how to do that.
<span rel='http://www.xml-cml.org/convention/crystallographyExperiment' resource='http://wwmm.ch.cam.ac.uk/crystaleye/summary/acta/c/2010/01-00/data/dn3132/dn3132sup1_II/dn3132sup1_II.complete.cml.xml'><a href='
'>Insert text here</a></span>
Or of you’re using a scholarly HTML aware tool like the Beyond The PDF demonstrator you can create a document with the link in it, and the Scholarly HTML preprocessor in the tool will unpack the information encoded in the link, and insert a span just like the one above.
So now you have an HTML document with some semantics embedded in it. If you happen to have a browser plugin that understands the semantics, or you have a penchant for reading the source of web pages then you might be able to do something useful with it.
I think a more sensible use case is that you use a delivery system that knows about Scholarly HTML.
So, using a test plugin, for example this blog can now show you chemistry. Like so: http://ontologize.me/?tl_p=http://www.xml-cml.org/convention/crystallographyExperiment&triplink=http://purl.org/triplink/v/0.1&tl_o=http://wwmm.ch.cam.ac.uk/crystaleye/summary/acta/c/2010/01-00/data/dn3132/dn3132sup1_II/dn3132sup1_II.complete.cml.xml
You can try out the very simple plugin if you can work out what to do:
hg clone https://bitbucket.org/wwmm/schtml/
What does all this mean? Why is it important?
What I have demonstrated here is a way of embedding some semantics in a scholarly document in a way that builds on existing practice and serves as a example for conventions for other disciplines. Instead of embedding a JMOL viewer in my blog post by pasting in a lot of script-code, I can cause a JMOL to be shown in appropriate circumstances. I will post a more generic demo of reference and citation handling soon and there is much more to Scholarly HTML to talk about, including packaging for scholarly objects.
We have a few things happening here:
We have a format: a scholarly HTML convention designed to work in any web-processing environment or browser and to be as interoperable as possible, including supporting that important kind of interop, preservation, AKA “interop with the future”.
But we need better tools: the tool chain I have shown here is probably viable for people who really really need to do this, or like playing with code, but we can do better. I can imagine tools built into Word and WordPress and so on that make this easy and transparent for authors.
And finally, yes Todd Carpenter this is partly a branding exercise, like the Linked Data movement but we think it is for a good cause. If we can get this right we can help the web be what it was meant to be from the start, a platform for scholarship.
Copyright Peter Sefton, 2011. Licensed under Creative Commons Attribution-Share Alike 2.5 Australia. <http://creativecommons.org/licenses/by-sa/2.5/au/>