Various updates to EADitor ahead of new release

[Ethan Gruber]

Hi all,

The American Numismatic Society will be releasing their finding aid collection, "Archer," publicly in mid-late October.  This is an ANS implementation of the EADitor general distribution (the code is virtually identical except for the CSS file), and I plan to release a new EADitor beta to coincide with the release of Archer.  I have written a blog post outlining recent changes to the application: http://eaditor.blogspot.com/2011/09/brief-description-of-recent-updates.html

Ethan

[dan freidus]

Ethan,

Is installation (and maintanence) of Tomcat on a Mac pretty
straightforward (for someone who doesn't do much via command line
interfaces anymore)? I'd be doing this so I could run Numishare and/or
EADitor witht he hopes of having a single database that includes
records of coins I own, coins I've seen in private collections, as
well as records I've downloaded from institutions such as the ANS.

A single platform for both collection management and managing research
projects would be great. I'm just unclear on how much work it would be
to learn, get running and maintain.

Thanks,
Dan Freidus
fre...@umich.edu or
danf...@gmail.com

[Ethan Gruber]

Hi Dan,

Tomcat runs on a Mac, but whether or not Numishare is the best solution for your collection really depends the level of access you want people to have.  Numishare and EADitor are applications designed to run on servers that provide access to collections to a wider audience online, so they may be heavyweight apps to run on a personal computer for cataloging personal collections that only you will search and browse.  We aren't currently working on any smaller-scale applications for collections management.

Ethan

[dan freidus]

Hi Ethan,

I was trying to figure out if it'd be worth running these as tools for
collaborative research. Allowing colleagues to upload records from
various collections they examine (institutional, private, or auction
records) to a database with a common format would be great. If it ran
on a server so instead of users sending records to an administrator
who maintains a single user database, it would be easier to
colloaborate with more than one or two other researchers. It
shouldn't be too hard to have it be on a server but require a password
for certain actions (e.g. adding records), right?

Dan

[Ethan Gruber]

Hi Dan,

Sure, you can have one or multiple accounts for people to log into on the server in order to upload information.  I'd recommend we discuss this further in direct email or on the Numishare google group (http://groups.google.com/group/numishare), since the EADitor group consists mainly of librarians and archivists who probably are not interested in emails about coin collections.

Ethan

[Dale Patterson]

Good morning everyone - 

I am in need of some advice and guidance concerning the installation and functioning of the EADitor.  I have downloaded the following  files from the EADitor site: editor.-1109.zip ; orb eon.war and solr.war.  I have placed them on my Ubuntu server at home for testing.  I am running **** on an older machine.  I installed using the Ubuntu guide on the Wiki and making the changes from 'localhost' to my server's IP.  Everything seems to have installed, but here are a few of the challenges I've encountered.  

First let me say I am attempting to access EADitor through my MacBook and have the same challenges whether I am using Safari or Firefox.  I have an older Windows based laptop and have similar problems. 

First it is incredibly slow.  They may just be my server, but I am running Tomcat6 and another version of eXist on this same system and it isn't this slow.  But again, that my be my server.  I have frequent timeouts errors. 

Second, and more problematic.  I can't upload my existing EAD documents, which are DTD based and validate in oxygen into the system.  In Safari, after I select the document I get some fake path statement pointed to a non-existent "c:" drive.  In Firefox al I get is the file name - no path appears - and then an error statement that the file is invalid.  Ok, so maybe I need to hand convert all 254 files to EAD schema.  I can deal with that later, if I must.  

So, I decide to create an EAD from scratch.  It is a pleasant experience, except when I get to the component levels.  Maybe EADitor is using different terms here, but what I mean is the <c> element within the <dsc>  I can only make one level.  For example, if I have two series, I can create a <c> level for each series, but then there are no links, plus marks, or whatever, for creating <c> levels underneath those series so that I can enter file folder  descriptions. Nothing.  What am I missing, or better yet, where am I missing it? 

And one final thought.  At some point on the Wiki, I think, I read where there was a link in the EADitor to change its style.  Don't see it.  

So, any assistance or direction on what I have obviously missed would be most appreciated.  Thanks so much.  And especially to you Ethan, for realizing that an EAD should not be chopped up into bits and stuffed into an alternative database.  This is the right direction.  

Thanks

Dale Patterson 
--
Dale Patterson

[Ethan Gruber]

Hi Dale,

Since you are running Tomcat on an Ubuntu server, I am assuming you installed Tomcat through the package manager.  Are you running Sun's Java or the open source version (openjdk, I think it's called)?  Have you modified the Java options in /etc/default/tomcat?  I've found that Tomcat running EADitor needs at least 1024MB of RAM devoted to Java to run well.  I set the Java heap to 1600MB on my desktop development computer and our web server, and forms take about 10-15 seconds to load.  If you're getting PermGen or Heap errors, check out Orbeon performance tuning: http://wiki.orbeon.com/forms/doc/developer-guide/admin/performance-tuning#TOC-Tuning-the-Java-Virtual-Machine-JVM

I'll take a look at the upload mechanism.  There are two stylesheets than an uploaded finding aid is passed through.  The first makes some changes to the guide--moving unitdates within unittitle up a level to occur directly in the <did>, changing all numbered components to the plain, unnumbered <c>, and a few other minor changes.  The second stylesheet is the dtd2schema.xsl that can be found online.  There may be glitches in the former stylesheet that cause the transformation, and upload, to fail.  I have attempted to upload numerous finding aids from the Virginia Heritage Project into EADitor.  Ones from UVa almost categorically fail, but I can successfully upload finding aids from most other institutions.  There may be a glitch in UVa's encoding practices that triggers the transformation to fail.  I have not done significant testing of the upload mechanism, but will try to address the problems before the next release.

As for creating subcomponents, I have tweaked this code significantly recently, but the changes aren't available in the downloadable package.  I am aiming to release the next beta before the end of the year.  In practice there will be two ways to access the subcomponents.  When you add top-level components into the <dsc> and save the guide, the table of contents in the column left of the form will be updated.  The table of contents contains a nested list of all components, and you can click on the link to load that component individually into a new form.  You can repeat the process of adding subcomponents.  The other way to access components is through the finding aid list in the Admin page.  If a finding aid has components, a link labeled "expand" appears to the right of the title of the finding aid.  This link shows/hides a nested list of components.  You can click the link for one of the components to load it into the form.  Eventually I would like to introduce a simpler way of creating components lists.  Perhaps I will have time to do that before the next release.

I have started to document EADitor in a wiki:  http://admin.numismatics.org/wiki/doku.php?id=eaditor:eaditor  There is much left to do, but this is a good start.

Ethan