Ideas.... » XML

OAI-ORE LITA Presentation with Herbert Van de Sompel

Rosalyn Metz — Sun, 25 Jan 2009 22:18:37 +0000

Yesterday I went to the LITA sponsered talk about OAI-ORE. I went because I wanted to see the man (the myth, the legend) behind SFX and OpenURL, as well as bX, Ex Libris’ new recommender service. Basically, if it weren’t for Herbert Van de Sompel‘s brain I wouldn’t have had a job for the two years prior to working at Wheaton, so you know…I had to go.

So anyway, on to the real meat of the post. What is ORE? Well that’s a great question; let’s see if I can sum up what I’ve learned.

First, Herbert did a quick crash course on what RDF was. It consisted of this:

resources exist at a particular URI
RDF is about explaining resources
there are things called RDF triples, they consist of a subject, predicate, and object
if you don’t have a URI for a resource you can have something called a literal

Below is an image of RDF triples:

Kind of an abstract idea, but if you’re really interested in learning more check out W3C’s page on RDF, perhaps reading the RDF primer first (of course I recommend this, but I have yet to do that myself).

Ok so now that we have a basic understanding of RDF we can move on and gain a basic understanding of ORE (supposedly).

The ORE Primer states that:

Open Archives Initiative Object Reuse and Exchange (OAI-ORE) defines standards for the description and exchange of aggregations of Web resources.

Brilliant, so now instead of just describing one resource we can aggregate resources together, but then what? OAI-ORE uses the same http 303 redirect guidelines that linked data does to redirect users to something called a resource map. This map describes how the different resources in the aggregation are related to one another.

What’s great about the resource map is that it lives in what Herbert calls the two web worlds (web 2.0 and the semantic web). The resource map can be written in Atom XML, RDF, and RDFa (RDFa is like RDF except in XHTML rather than XML).

So now we can have aggregations of resources in a machine readable map rather than sitting out there as seperate resources seemingly unrelated to one another.

Herbet went on to say that Dr. Robert Sanderson at the University of Liverpool actually went through and created resource maps for everything in JSTOR. You can find out a little more about this project via this Code4Lib listserv posting.

So that’s OAI-ORE in a nutshell. A very small, simple minded nutshell.

Share/Save

Installing Tomcat and XTF on my Mac

Rosalyn Metz — Mon, 20 Oct 2008 18:55:50 +0000

Friday I decided to be a complete and utter geek and install Apache Tomcat on my Mac. I know, I know you’re asking me “But Rosalyn Apache comes with a Mac, why do you need Tomcat?” Well that’s a good question. And the answer is I wanted to install XTF which is an XML publishing tool developed by CDL (California Digital Library), and XTF requires that you use Tomcat in order to run it (or you can use Resin…but I’m an Apache fan).

So now you’re thinking, “How did you do it?” I wish I could have thought it up on my own, but saddly I didn’t. I looked at a few pages and then finally settled on Basil Borque’s section (under Update 2006) on this wiki page: http://wiki.apache.org/tomcat/TomcatOnMacOS. Bravo to Basil for posting the directions. I didn’t bother with the firewall though because my laptop is behind a firewall by default. Also I wanted to still run Apache so I left Tomcat on its default port 8080.

Once I got Tomcat up and running, I then moved on to XTF. This process was ridiculously easy. CDL has wonderful XTF installation directions on the SourceForge XTF Wiki. If you’re on Wheaton’s Campus (and my computer is currently connected to the internet) you can view XTF here.

Share/Save

Creating a Constrained TEI Schema

Rosalyn Metz — Sun, 12 Oct 2008 18:02:31 +0000

To create a schema you can use Roma. This gives you a graphical interface that allows you to include or exclude elements, include or exclude attributes, and constrain values for attribute values.

A few pointers for using Roma…

hit submit every time you make a change. there is a bug that sometimes it doesn’t look like you’ve saved your changes, but in reality you have. and if you don’t hit save, then you’ve definitely not saved your work.
give your schema a name that will help you understand what its purpose is for. you can also supply a description of its purpose. one idea is to create a schema and then come back later and modify it as you delve deeper into analyzing your text.
save your schema as Relax NG schema (compact syntax). this files gives your xml code a place to go to double check that you are organizing your data properly.
save your documentation in html so that you can read what the elements and attributes you can use in your customized schema are.
save your customization (aka the ODD file). this is the file that creates the schema (ie .rnc file). if you have any problems with your schema you might want to look at the ODD file to make sure the changes you made are properly reflected in the schema document.

creating a schema allows your coders to reduce the options they can choose from and hopefully reduce the amount of mistakes they can possibly make.

Share/Save

TEI Publishing

Rosalyn Metz — Sat, 11 Oct 2008 21:28:02 +0000

Ways to Publish Your TEI Documents

CSS — just to make it look pretty
XSLT — to transform it into another thing (like XHTML)
XML Databases — to query it. a good opensource DB is eXist.
XML publishing systems — allows people to install software and publish documents. an example of this is XTF and TEI Publisher.

Share/Save

More TEI Markup Notes

Rosalyn Metz — Sat, 11 Oct 2008 16:32:17 +0000

Encoding Parallel Structures

This is useful for mapping one thing to another, in the below example we are translating french to english. Essentially you input an xml:id attribute the corresponds to another xml:id attribute. Below is an example of this.

Nos péchés sont têtus, nos repentirs sont lâches;
Nous nous faisons payer grassement nos aveux,
Et nous rentrons gaiement dans le chemin bourbeux,
Croyant par de vils pleurs laver toutes nos taches.

Our sins are stubborn, craven our repentance.
For our weak vows we ask excessive prices.
Trusting our tears will wash away the sentence,
We sneak off where the muddy road entices.

More Complex Parallelism

Its also possible to link to multiple elements, so in the below example mapping the french to multiple versions of the english.

So essentially you’re aligning parallel versions of a translation to one another. The poem lines are then given ids so that everything matches up. If you wanted to you can also link to another file rather than an id within the document.

Choice Elements

Allows you to show different spellings, or just choose between options.

…with them, bycause they woulde
not be

bo?de
bounde

also for an other wo
m? at theyr pleasure, whom they
knewe not, nor yet what matter

So you can have mark up that looks at the original reading and then you can see the corrected version. So if you want to view the spelling mistakes or whatever you can, but if you want to only see the corrected version you can.

Revision Processes/Editing Processes

It’s possible to look at revised manuscripts and step through the revision process. So if you can determine what was added and deleted you can include it in your mark up

After anthe ~~unsolv’d~~ argument
~~TheComing in,~~ A group of little children, and their
ways and chatter, flow in ~~upon me~~
Like welcome rippling water o’er my
heated nerves and flesh.

There is also a more complex version of this that allows you to express the different versions of the text (assuming that you can determine it from the original). In the below example you can see how a person was trying to determine what it was in the original document.

Johnston etc 1764 Mr Nikle
s Brown
&Co to me George Beverly juner
to ten Rum Barels at Four pound &per; Barel — — — £40

Slides For this Talk

The slides for this talk can be found on Brown’s Website

Share/Save

TEI interface development at Hamilton College

Rosalyn Metz — Sat, 11 Oct 2008 13:58:00 +0000

First Generation of Civil War Letters Project

Hamilton encoded administrative documents and letters from the civil war. They used TEI and Dublin Core metadata with their documents (for historical reasons they used Dublin Core).
They then created a PHP interface. The Dublin Core description was used as the “abstract” for the record and the text appears when you view the “full record”. The pages themselves were transformed using XSLT into HTML and then a CSS was used to style the page.

They created drop down boxes for person’s names, organizations, place names, and geographic features.

They also included images of the page. A link was encoded into the TEI to link to the images of the actual documents. The TEI elements that you might want to use to do this would be a page break; below is an example of code that would do this:

Second Generation of Civil War Letters Project

These letters were encoded and then loaded into a database called eXist-db. Since the db is already indexed it makes it easier to find data.

The Dublin Core records were replaced by ContentDM. So the records are in ContentDM, his search interface is in PHP, again with drop down boxes pulled from the eXist-db. The dropdown boxes then links to canned searches in ContentDM. The “search” then taxes you to records in ContentDM

If you want to view the full text of the letters, a link in the ContentDM record takes you outside of ContentDM to a webpage. The page it links to is TEI transformed by XSLT into HTML and stylized using CSS. There are links in the TEI that point toward the digital image of the original text.

Below is a link to the current digital collection:

http://elib.hamilton.edu/hc/hcbrowse.php?id=col_spe-civ

Share/Save

TEI Conference: Basics of TEI

Rosalyn Metz — Fri, 10 Oct 2008 17:07:13 +0000

Powerpoints for this session can be found on Brown’s Website.

TEI allows for…..

long term storage. its in xml so it should last
allows you to analyze information
if you want to share your data it needs to be meaningful for others. if you encode in TEI others can figure it out.

TEI is…

a standards organization
maintains, develops, disseminates the language
humanities documents of all kinds (internationally)
membership consortium that pays annual fee to support the TEI and activities surrounding TEI
social! it allows people to get together to discuss humanities
a set of guidelines that covers the grammar and vocabulary of the markup language

A system that you can take and apply in the ways you need to apply your data.

Different types of disciplines use different standards, ie. digital libraries might need different information than linguistics. However they are all valid in the TEI language. Some projects using TEI are listed on Wikipedia. There is a logistical way to normalize the different ways people use TEI. The MONK Project has created conversion processes to make the different types of TEI to work with their TEI.

Another thing to think about is the Resource Description Framework (RDF) developed by the W3C. Information on RDF can be found on the W3C Website.

Share/Save