Peerticles & peer publishing
If a mimi is a personal data-space that can last a lifetime and acts, at minimum, as a journal, an archive and a set of references then a logical question arises. Why would you ever want to write anything anywhere else?
Computers are supposed to make tasks simpler and less onerous, and it is this promise that leads people to regard entering their name, phone number and email on yet another form as intensely irritating. A frequent and reasonable response to repeated requests for such information is, “Why am I doing this? Isn’t this something that should be taken care of automatically?”
Even more irritating, then, is the prospect of publishing a conference paper on a conference site, then publishing a copy in your personal data-space, to find that you are now, by default, responsible for the problem of how to keep the copies manually synchronised. At some point in the future, you correct a reference on your site and then have to remember to log into the other site and paste the correction in there. Or you will forget to.
The success of RSS has achieved two things. It has familiarised people with the idea that information can flow to them automatically and reliably. It has demonstrated that this process need not be technically very complicated. RSS, though, has been conceived as a way of keeping you up-to-date with information that changes rapidly: sports headlines, the latest uploads to YouTube, my opinions, and so on. This is parallel to, but not the same as, what I am referring to here as peer publishing.
What will peer publishing be used for?
Imagine an author (whether of academic papers or science fiction, it doesn’t matter). Imagine that this author is fully convinced about the benefits of publishing under a creative commons licence, and sharing her work with other people. Either she does not expect to make money from publishing her work or, like Cory Doctorow, she has come to the conclusion that the appearance of free versions of her work online will not detract from her ability to sell printed versions.
This author writes an article or story whose concerns lie at the juncture of two topics (cultural diversity and visual learning, for example). The compiler of an online anthology about cultural diversity wants to include it in the anthology. At the same time an organiser of a conference on visual learning wants to put it on the web as a conference paper. And the author wants to a) post it in the archives of her mimi, and b) keep a simple but effect control over updating and correcting it.
She peer publishes the article in her mimi, using means I will outline below, and (in a manner similar to RSS) the two other sites subscribe to the article. By subscribing they get the full text, graphics and other media, displayed on their site using their own specific style sheets, so that the article fits in with the look of the site. They also get the article updated automatically whenever additions or alterations are made to the original article residing in the author’s mimi.
There are examples of sites that might gain immediately from peer publishing. I posted a couple of weeks ago about Learning Spaces, a book of essays available online from Educause. If you look at the site the articles are all available for download. Either the authors have been given carefully moderated rights so that they can update their papers without interfering with the others, or authorial changes have to flow through a webmaster who deletes and replaces files as they are sent to him. With peer publishing in place this would happen automatically without any author needing writing or uploading permissions for the Educause site.
Students would benefit greatly from this approach. Where teachers peer published material online and students subscribed to it, then it would update automatically. Where students subscribed to external reference material they found on the web (several chapters from Learning Spaces, for example) then their reference material would update itself for them. Moreover, by encouraging students to gather their source material in the archives of their mimi many of the difficulties of suspected online plagiarism would disappear. Gathering and annotating existing work would be an overt part of the learning process, and could itself be taught and credited.
Students would also derive great benefit from peer publishing their curriculum vitae and portfolios, and the peer publishing of such would also serve to encourage others to search and package portfolios into employment anthologies for prospective employers.
How will peer publishing work?
Like instant messaging, file sharing and other peer-to-peer phenomena peer publishing contains both cultural and a technical issues.
Culturally, there are certain prerequisites to participating in such an idea. Most obviously, peer publishing will only work were the desire to make information public, to share it, is stronger than the fear of losing control of it by making it public. (If I worried about somebody stealing this idea and becoming rich at my expense, for example, then I would not be publishing the idea here like this.) This rules out individuals and institutions for whom the application of industrial-strength copyright is a primary goal. By implication it suggests that peer publishing will work best where copyright is maintained through Creative Commons licences that retain some rights while waiving others, or where copyright is not an issue at all. Much has been written about this, and there is still much to be written.
Technically, peer publishing will consist of three elements: an agreed xml schema; a plug-in at the creator’s end for creating a subscribeable peerticle from an article; and a plug-in at the user’s end for turning an arriving peerticle into a styled post on the user’s site.
If we presume that we are building a mimi around a kernel of WordPress (as I am attempting to do here), then the article will begin life as a standard WP post. The author will have an additional option when the post is saved: the option to create a peerticle for peer publishing. If this option is chosen then a php-based plug-in will take the contents of the post, add additional information, save this as a plain text file in a peerpub directory on the server, and then place a button at the end of the post that (like an equivalent RSS button) gives the reader a URL to use to subscribe to the peerticle.
The text file that is stored – the peerticle – will contain a fragment of xml. This will contain the article itself, marked up in xhtml, with all the relative links converted into absolute paths. This will ensured that any images or other media referenced in the article can be retrieved and displayed when it is subsequently served from the subscriber’s site. It will also contain a standardised set of metadata containing the author’s name; the date of creation; the original tags and categories; the original permalink for the article and the base url of the originating site; the kind of Creative Commons licence under which it has been registered; and possibly more.
The xhtml will itself need to be well-formed, and in an agreed format, using standardised classes and ids. This will be necessary so that subscribers can confidently apply a stylesheet to the incoming peerticle. Without such an agreement the system will be useless for anything more complex than unmarked text strings.
Subscribers will need a plug-in to search for updates at regular intervals; fetch those peerticles that have changed; parse the resulting xml; and convert the result into a standard WordPress post. Preliminary examination shows that there is a plug-in that is capable of doing this: inlineRSS which is downloadable from the iconophobia site. This will import a file and apply an XSLT to it, before sending it to WordPress to be processed as page content.
I am using this site as a working laboratory to develop the ideas and techniques needed to create a working mimi. Initially it was powered by a WikkaWiki, and then later by WordPress, as I experimented to see which would make the best foundation. When I transferred the content from wiki to blog I left many of the longer essays relatively unstyled, since I knew that CSS would play an important part in the development process.
The first thing that I will do then is to go through the existing CSS and make it as rigorously logical as I can. I will attempt to use standard xhtml tags, with easily understood classes. Once I have done this then I will explore the possibility of rewriting the text editor that WordPress uses in order to incorporate drop down shortcuts for applying the styles.
Whether or not this is possible in the short term I will hand-make some peerticles by manually styling some existing essays, bundling them into an xml fragment with version one of the required metadata, and saving them as text files. This should enable me to test inlineRSS, and to write the relevant XSLT document to “unbundle” the peerticle, so that inlineRSS can pass it to WordPress for display.
If the handmade peerticles work then I will have a proof-of-concept. I will then be able to exactly what the authoring plug-in needs to do and (hopefully) find a student who will find this an interesting thesis project.
I am deliberately exploring the idea of a mimi in a bottom-up way, by building, discovering problems, thinking about them, and then moving forwards (or sideways, or whatever). Because of this I am starting each stage from where the last stage left me, trying not only to solve the big problem, but to find out what the big problem actually is, by solving smaller problems as they occur.
This means that I am working with WordPress and text files at the moment, but that is not where I see the limits of peer publishing. If this can work with WordPress it should work with any other web-based software capable of receiving data, transforming it and applying styles to it. That means any blog, wiki or content management system worth looking at should be capable of peer publishing and subscribing to peerticles.
If I don’t see this as restricted to WordPress neither do I see them as restricted to text. If peerticles are xml documents, and the xml can contain absolute paths then peerticles can act as wrappers for anything that can be retrieved using a URL. The xml schema will simply need to be able to cope with wrapping pointers to different media in a standard way.
And that is a small problem that will occur at a later date, when these have been safely put to bed.
This paper directly relates to The mimi: a key tool for pedagogy, a thesis for an MA in epedagogy and visual knowledge building.