PT’s blog

Last week we held a workshop at USQ where I invited a bunch of developers to come, gave them a vague theme and set them loose. The process seems to have worked well. I figured that if invited good developers, and asked them to bring a laptop then we would end up with some code and integrate a few things that have been going on in their own little silos.

I was right.

See the summary that I wrote with Daniel de Byl over on the ICE blog – and if you want to find out more use the feeds to subscribe and see the forthcoming screencast demos.

I recently reported on an application called FieldHelper, that I saw at Clever Collections. FieldHelper inspired me to buy a data cable for my GPS (a yellow Garmin eTrex). The cable’s not a great success, I went for a USB version off eBay and it only works on Tuesdays and only on the Windows box at home, no luck with the Mac so far.

Anyway, I took the GPS on the regular Sunday morning ride with the Toowoomba Bicycle Users Group. This group is friendly easy-paced and lycra-optional, not a racing pack. The original plan was to try to use FieldHelper to synchronize the time data on some photos with the GPS track so I could show them on a map with little flags, but the camera I took thinks it’s main calling in life is as a telephone and it doesn’t seem to have added a proper timestamp. Still, I was able to import the tracklog via GPSBabel, load it into Google Earth, export as KML then upload into Bikely. I know that’s a bit complex, but we’re doing this in the name of automation – our aim is to make it all much easier.

Here it is, an interactive map.

If you click on the title at the top of the map it takes you to Bikely where you can do stuff like see the elevation profile of the ride (click on Show / Elevation profile to see that we climbed 471m and apparently ended 8m below where we started, which may be to do with me editing the log so it doesn’t start and end from my house).

To make this happen, Ron Ward added a feature to ICE, along the lines of the hack I did to include CML in posts.

As this is the first time out, I had to do a few steps:

1. Visit the map at my Bikely account.

2. Click on the Share / Display this map on your blog or website link.

3. Copy and paste the HTML code they gave me into a new file. (And edit the code to work around a Firefox bug with empty elements.)

4. Copy the file up to my website. (I only had to this ‘cos of a limitation of ICE that will be fixed soon, in the future you will just save it in ICE).

5. Capture a screenshot of the map and paste it into my document, to serve as the print rendition.

6. Link the screenshot to the online version of the map.

When that’s all done and I look at the document in ICE it includes the interactive map in my web page while using the screenshot in the PDF. This is a really useful generic mechanism for including inline web stuff, could be used for lots of things where you want a live version for the web and a dead version for print.

Too complicated?

What we’ll do to make this work better in future is use the forthcoming ICE plugin system. All you will need to do is drop the tracklog from the GPS into ICE and it will make both a print-renderable map and the code you need for a live map. The same plugin system will work for stuff like CML, and countless other data driven visualization tools.

Some time ago, around the time there was some discussion on why repositories use PDF and not HTML there was this thing called Lemon8-XML announced – at the time there was very little information available. I tried to get on the mailing list but nobody got back to me. This week we’re having an invitation-only workshop at USQ, Connecting Repositories to Authoring Tools, and I went back to have another look. Turns out that there is now an FAQ. And a test site – which ICE collaborator Ian Barnes knew about.

L8X comes from PKP, same place as the Open Journal Systems (OJS).

The bit of the FAQ I’m interested in is, of course, about using styles.

Q: does it rely on authors choosing styles?


A: not at this point, although we have considered developing (or adopting) a standard set of styles to help provide the document parser more information to work with. We would be glad to hear from anyone willing to help collaborate on a standard style template for authors.

http://www.lemon8.org/pages/docs#faq

We at the ICE project would be delighted to work with the developers from PKP on a standard style template. We have lots and lots of experience at developing styles and cross platform, easy to use interfaces to apply them. We also have pretty well developed code to go from our generic styles to XHTML. It should be possible to adapt this to target the XML format used by L8X either by writing a whole new converter, or by converting via XHTML-plus-microformats.

I retried registering my interest and I’ve now made contact with MJ Suhonos, who is looking forward to talking more, as am I.

If we were to provide ICE-style templates for OJS then one potential workflow scenario could run like this:

1. The hard working Journal manager makes a template available using ICE-like styles.

2. A careless author ignores or misuses the template.

   This is normal for journals. There is typically lots of reformatting to do at the end of the process. I know from the questions people frequently ask about ICE that many people believe that authors cannot be induced to use styles properly, but our experience at USQ is that if you provide a feedback loop so people can see their document converted to HTML, then they do use and even like templates. See my post, Why ICE works for my take on this.

3. Careless author uploads a document to OJS for the first time.

4. OJS sends the doc to L8X, which tries to find the structure as best it can.

   L8X has a four-tab interface where you can see how it went:

   1. Extracting metadata

      (another thing we’re working on this week is how to embed metadata in ICE documents so it is easy to extract it reliably – if we can get people to use styles then this becomes much easier)

   2. What it thinks the structure of your document is.

      (We find in ICE that showing a table of contents, generated from headings gives people really strong feedback.)

   3. Extracting citations

      (In ICE we’re working on Zotero integration, so that citations are formatted automatically – this should make it much easier for L8X to recognize them)

   4. And a preview in HTML and PDF

5. L8X then round-trips the document back into a word processing doc using ICE styles and passes it back to OJS. The quality of the document has just improved!

6. OJS uses the styled doc from then on. If the author adds content, or metadata about another author then there is a good chance that they’ll do it correctly, if the template makes it easy. We’ve already got a toolbar that tries as hard as it can to format things using styles, we could extend that to have buttons such as ‘add another author’ just by using autotext.

(I tried out the Lemon8-XML demo site, using the styled Word version of my paper from Ausweb ‘06. There were several problems, which you’d expect from alpha software like this – I have reported the bugs on the wiki.)

I’m a bad, bad blogger. I find it hard to write reviews of conferences as I go, or even immediately afterwards, but I’m going to try to improve on that in 2008.

This is far from liveblogging, but here are some notes of my impressions from the Clever Collections conference in Melbourne last month. This post is expanded from my brief report for the RUBRIC board.

Daniel de Byl and I ran a workshop on ICE the day before the conference. We had a small, enthusiastic group who seemed to get at least something out of the day. We’ll see how many of the attendees keep in contact. Past experience is that without the kind of institutional support we provide at USQ, individuals don’t tend to stick with ICE (either that or they are very very self sufficient and don’t need to ask us or the ICE mailing list any questions).

One surprise was the talk by Professor Noshir Contractor, From disasters to WoW: enabling communities with cyberinfrastructure. He looked at the way social networks affect how people do research, and how understanding networks could make research more effective. There were some striking examples of visualization tools that showed how research articles and researchers are linked in various ways, going way beyond the simple search and browse used in IRs at present. One example showed two distinct networks of research and researchers working in the same area, but with little overlap (research articles have their own social networks apparently, through things like citations and keywords). I think the example was from the NSA in the USA, looking at how they could merge the networks somewhat.

Dr Ashley Buckle delivered a compelling argument, The Australasian Repository for Diffraction Images, in favour of open access to research data citing recent cases of scientific fraud in protein crystallography.

I was impressed by a whole session with APSR folks talking about their work on repository interoperability.

RIFF & the Australian METS profile

RIFF is the Repository Interoperability Framework which we’ll be exploring in detail this week. The METS profile, developed by the National Library is a real improvement over the situation a couple of years ago – it should now be possible to create packages that can drop easily into both Fedora and DSpace – something we couldn’t do at the start of the RUBRIC project.

AONS2

A utility for managing file formats in a repository and reporting on preservation issues.

Fieldhelper

A repository client designed for field researchers such as anthropologists or archaeologists. It has some interesting features, such as the ability to group items and add metadata to all of them with a cute drag and drop way to apply metadata.

Have a look at the screenshots. I think the interface show some promise but the application needs a lot more work and a good deal of field testing.

I was inspired by the demo by Steven Hayes in which he automatically geo-tagged a set of images and packaged then for ingest into a repository to buy a cable for my GPS. Stay tuned.

Two of the presenters, Scott Yeadon and Leo Monus, are coming to USQ in Toowoomba this week to a workshop we’re holding on interoperability between repositories and researcher-tools such as ICE and Zotero, as well as workflow tools like the Open Journal System. More on that workshop/unconference soon.