ICE for theses (ThesICE) , where we are we up to?

Danny Kingsley who is the Manager, Scholarly Communications and ePublishing at the Australian National University, emailed me to ask about ICE and theses. I asked Danny if she minded making her question and my reply public, and she’s OK with that, so I will use this as an opportunity to summarize where we are at with ICE for Theses, in my usual long-winded way.

Danny asked:

We are trying to plan some processes for electronic submission of theses and trying to work out ways of getting a thesis to an examiner that they can annotate without printing it out. The ICE process came up in discussion so I have a couple of questions:

  • Does the output allow for any annotation?

    [PS: Yes, but the annotation in ICE is designed for the drafting/authoring stage, we have not done the work to make an annotation system that could be used by examiners, with all the privacy issues etc sorted out. More below on options on how to proceed.]

  • What form can ICE take as input I needed to make sure my Word document was properly styled how much leeway is the system able to cope with (if a student puts hard formats into the document what happens)?

    [PS: Hard formatting inline and in tables is preserved, but when converting to HTML, ICE should discard any paragraph level format that's added, like making a whole paragraph italic or hand-formatted headings. For PDF all formatting is preserved. We do require the use of ICE styles for headings and bullets etc, but if ANU has a house style you could (a) adapt ICE or (b) write a conversion macro. ICE styles are also extensible so if new ones are required that's easy and cheap.]

  • Are there other inputs that the ICE system can take?

    [PS: ICE takes .doc, .docx and .odt for documents, but it also has a plugin architecture for other file formats so it could be used to render wiki markup (kept as text files), LaTeX (we have a LaTeX plugin already but LaTeX is not standardized well enough to reliably convert aribitraty markup to HTML), or anything else for which you can get commandline conversion software we can plug in. We have a growing list of things it can handle, like chemical markup language, text to speech for documents.]

In summary the rest of this post says that:

  • ICE needs support, which should be doable at ANU.

  • The released version is designed for use on the desktop, which may not be ideal for many candidates and/or supervisors.

  • There is a web-only version of ICE but it is not stable enough to release yet, we would have to scope this.

  • We don’t have solution for managing theses past the authoring stage, with annotation and structured comment by external examiners, I suggest a couple of leads below and note that eventually our work on The Fascinator should yield these features, but not for a while.

That’s the summary, now some more detail. I will look at this from the beginning, covering:

  • What ICE offers for thesis writing and why you might want to use it, or not use it.

  • What do we have now that could be used and what you’d need to support it.

  • Potential future developments.

About ICE and theses

ICE, the Integrated Content Environment was developed by a team at USQ under my direction for the purpose of creating web and print courseware from the same source files, using a word processor as the editor. It’s now a core system of the university, and it’s freely available as an open source product, but it’s a product that would need to be supported.

I put up an example of an ICE-formatted thesis here ages ago, my BA Honours thesis in computational linguistics from 1990 note that there are some issues with the images because it was done with obsolete software on the Mac, but you get the idea of what a thesis looks like in HTML rather than the standard approach of sticking PDF files on the web.

Since then we have converted a few more theses to ICE after the fact, and we have some people writing theses in ICE, but not a very concentrated effort.

Looking to the future, my group recently completed a project with Peter Murray-Rust’s group at Cambridge, to look at thesis workflows for chemical theses with live-links to data embedded in the document. We looked at issues like chapter-level embargo and suggested a repository architecture where there would be research office repository that managed examination processes and embargo, with the institutional repository continuing to handle open access. Have a look at the document Jim Downing and I presented about ICE-TheOREM, a joint USQ/Cambridge project at Open Repositories 2009. You can click through to the HTML view of the document, which was created with ICE, of course. You can also look at it as a slide presentation or as PDF all automatically generated from the same word processing document.

ICE consists of a few components.

  1. A toolbar, which works with a standard set of style names to give a consistent way to author documents. The toolbar can be used in Microsoft Word on Windows, or with OpenOffice.org on Mac OS X, Linux or Windows. The toolbar can work with templates, which can be set up to match any requirements that a university might have for thesis formatting.

  2. An application which can turn ICE-formatted documents into web-ready content, and automate the production of PDF. The application can run in a few modes:

    1. The standard ‘classic’ mode, which installs as an application on your PC (Mac, Windows, Linux). It shows you the stuff you’re working on (like a course, or a thesis) in your web browser, and it will backup and version control your content to a remote repository using Subversion.

    2. Server-mode where you interact with ICE as a web-site, uploading documents and data flies and media via a web form; this is potentially a really good model for research students writing theses, and it could possibly be used to but the software is not released yet, which means I can’t simply recommend it to ANU.

    3. As a service, where ICE runs quietly in the background. This is normally relevant only for embedding ICE in other applications, but it can be used if you want to just use the ICE Toolbar/templates and use ICE to convert your documents for the web. As far as I know there only person doing this is Peter Murray-Rust who uses ICE to blog. This service could be integrated with some other system like SharePoint or the Open Journal Systems, or Sakai.

Why use ICE?

Assuming that you have the resources to support ICE, which I’ll cover below there are a few reasons why an institution might want to use it for theses specifically.

Getting things done

So, if there is a reason to use ICE, in what circumstances might it make sense?

Using ICE-classic

Using ICE in the standard way would be a sensible approach to writing a thesis, and getting comments from supervisors, just as a couple of hundred authors at USQ use it to write book-length courses. This could work for on-campus students given some support staff, up to the point where the thesis goes off for examination. As I said, we don’t handle documents past the authoring process, we have button for repository deposit, but you don’t get to self deposit a thesis without having it examined.

Using ICE-Server

Using the ICE server could work well, except that it’s not stable enough for us to be comfortable releasing it. We are using it in internal trials with PhD candidates, but we’re not really dedicating enough time either to helping them or to fixing up the software, and I have to say that I’m not sure that we will fix up the software as it currently stands; it may be that the future for ICE will be to build its conversion services into our eResearch Desktop / Institutional Repository project, The Fascinator.

Using ICE with other services

ICE is designed to play nicely with other software, so one approach to getting a thesis written and commented on by a supervisor and examined would be to mash-up ICE with other software. We have looked at doing this with the Open Journal Systems, which has review features, if not annotation, and which in the past has been used at USQ for thesis review. That would be a modest development project, which I outlined early this year.

Another thing I have had a quick look at for this post, is digress.it, which is being used in a JISC supported project. It has similar inline annotation features to ICE with one huge advantage, it is built on WordPress which means that it’s easy for people to install, I got it up and running in about half an hour at my hosting service, and had my first document in there for comment minutes later. Now this is not suitable for examination, out of the box as the comments are in the open, but it could possibly be used for earlier stages of thesis writing, such as getting supervisor comments. A candidate could use the ICE template, post to a word press ‘blog’ periodically, with each new post creating a new entry for comment by the supervisors and potentially peers.

If you’d like to have a look at it, see my test site where I put up a recent blog post for testing.I was impressed by how easily I could get digress.it going, but I didn’t like the way it removed my document formatting, still, I have joined the developer list and the lead developer said he’ll look at a mode that does not strip formatting.

Another option would be to go ahead and build another one of my ideas, a general purpose annotation system to supplement ICE. We have a lot of the pieces, so between us I’m sure that ANU and USQ could make this happen given a business a case, and getting the resources.

Conclusion

Looking at where we are at with theses in ICE makes it clear to me that we have a lot of potential, but very little action. So, I’m going to try to get things moving again at this end, with what we have called the ThesICE project. This will involve:

  • http://jiscpress.blogs.lincoln.ac.uk Joss Winn

    Hi Peter, Thanks for your comments about digress.it Hopefully we can make it work well for you, through our work on the JISCPress project.

    Now this is not suitable for examination, out of the box as the comments are in the open, but it could possibly be used for earlier stages of thesis writing, such as getting supervisor comments. A candidate could use the ICE template, post to a word press ‘blog’ periodically, with each new post creating a new entry for comment by the supervisors and potentially peers.

    You could make the entire blog private so that only authorised users could read it. For single install WordPress blogs, I like this plugin:http://wordpress.org/extend/plugins/wp-sentry/

    WordPress has some privacy features built in, too. You can make certain posts password protected or private to only logged in users.

  • Danny Kingsley

    This request from us of Peter was part of scoping the options for electronic submission of theses. We have been discussing the possibility of creating a situation where the student submits their theses electronically and it is sent to the examiners electronically. We are trying to eliminate the need for printing out swathes of paper and then paying large amounts of money for postage. In a perfect world the examiners would have the option of annotating the thesis they are marking. Our student services department indicated that a large percentage of printed theses come back annotated.

    However it might be that we are a long way from electronic examination yet. I made an enquiry of a person here at the ANU who does research into examiners’ behaviour and her response has given me pause for thought. While there is no specific research on the reading of electronic theses. “We do know that some examiners read them on long plane trips, in bed or at the equivalent of a ‘place at the coast’ for weekend. We also know that they often/ even generally read in fits and starts over week or more. They read in the order of: abstract, acknowledgements, intro, conclusion and refs, then start back at the beginning. They take notes – mainly to themselves – as they go. Many tag pages with post-its to return to. Most read with a pen in their hand.”

    This indicates it is important to consider the *client* who in this case is the examiner. What do they want? How do they work? There is no point developing a range of fabulous tools if they do not fit with the working processes in place. I am considering undertaking a small research project to try and establish the answers to these questions – our solutions might be for a non-existent problem.