Mariage

TG will start work as a workstudy, digitizing EC's prose novellas, and doing some image markup.

LSPW and I spent some time setting up the XML document for her 1621 Sonnet de Courval task; we also had a discussion with CC about transcription policies, which gave rise to these decisions:

• Typographical ligatures will be ignored, because only some are available in Unicode, and they're actually stylistic rather than substantive.
• Ampersands will be kept (i.e. not expanded to "et") because they're easy and not a problem for the reader.
• Long s will be transcribed as long s, both in ligatures and where standalone.

Set up LW with access to the network folder, and added a range of obscure typographic characters to her favourites on the Character Palette. She's now working on transcription. If there's time, on Monday, I'll get the actual document set up with her, so that she can do transcription half the time and markup the rest.

Yesterday and today, had meetings with CC and the two new research assistants, introducing them to the project and the materials, and getting preferences and desktop setups organized for them. We'll have them both working on Sonnet de Courval, with SW working on the 1609 (which we'll treat as the base text), and PG working on the 1621. SW can do the base line numbering as a simple progression, and then PG can add line numbers later based on the 1609 lines, with additional lines sub-numbered (exact scheme still to be worked out).

In the process of tweaking some title capitalization, we noticed that elements below the paragraph level inside annotation divs were not being rendered at all. I eventually tracked down the problem to this line of code:

<xsl:apply-templates select="@* | * | text()" />

which I had to change to this:

<xsl:apply-templates select="* | @* | text()" />

in order to get the descendant elements of the paragraph tag to render correctly.

Also acquired some new texts from the BNL, and backed up everything to rutabaga.

CC came by with some text documents (PDF page-scans) along with one new image. I've processed the image (cleaned it up and produced three different formats) as detailed elsewhere in the blog, created an IMT XML file for it, and added some basic metadata; waiting for proper details from CC.

After much work and testing, we finally got a portable version of Mariage working alongside ColDesp on a thumbdrive. These were the tweaks we had to do:

• We had to set the Java Options manually in the catalina.bat and catalina.sh files respectively, like this:

  set JAVA_OPTS=-Xms128M -Xmx256M -XX:MaxPermSize=256m

  JAVA_OPTS=-Xms128M" "-Xmx256M" "-XX:MaxPermSize=256m

  All attempts to pass these values through command-line parameters in our own startup batch file failed. We wanted to make sure Tomcat had enough memory; the default 64MB isn't very much.
• Enhanced the Windows startup batch file, so that it will work both when it's in the root of a removable drive, and when it's in a subfolder somewhere. In the one case, the %cd% variable ends with a backslash, and in the other it doesn't. This is the batch file now:

  if EXIST %cd%jdk1.6.0_06 goto javaInRoot
  set JAVA_HOME=%cd%\jdk1.6.0_06
  goto doneSettingJavaHome
  :javaInRoot
  set JAVA_HOME=%cd%jdk1.6.0_06
  :doneSettingJavaHome
  
  set PATH=%JAVA_HOME%;%PATH%
  
  if EXIST %cd%apache-tomcat-6.0.16 goto tomcatInRoot
  %cd%\apache-tomcat-6.0.16
  goto doneSettingTomcatHome
  :tomcatInRoot
  set CATALINA_HOME=%cd%apache-tomcat-6.0.16
  :doneSettingTomcatHome
  
  cd apache-tomcat-6.0.16\bin
  echo starting...
  call startup.bat
  echo Waiting for the server to start...
  ping -n 120 127.0.0.1>nul 
  start http://localhost:8080/coldesp/
  pause

• Updated the version of Saxon 9 that was in the two Cocoons. 9.0.0.2, which was previously there, was throwing an error with some of our Mariage documents. We were unable to figure out why, but there are similar bug reports which suggest it's a problem with Saxon, and updating to the latest version solved it.
• Tested on Windows.
• Tested on Mac.
• Tested on Linux.

I now have two thumbdrives I can use, and we'll put the stuff on CD too for the trip, just in case.

The point of this is not just for the trip to Finland, though; in case of server emergencies, we can now run Mariage in any other Tomcat as a standalone, and repoint the domain to it. This is the way we should really handle all our projects in future, so that it's easier to update Cocoon for any given project without damaging the others.

Built a portable Mariage site by taking the shell of the portable Coldesp, emptying it out, and refilling it will Mariage data and file structures. It still doesn't seem to be working 100%; many pages (especially image files) die with an out-of-memory error (heap space). But perhaps that's the best we can hope for. It should work well enough to get the idea across, if we find ourselves presenting on the project with no access to the net or Lettuce. I'll do more tweaking tomorrow and see if I can make it more solid, perhaps by increasing heap space in the batch file that starts Tomcat.

The gallery pages were showing the engravings in no particular order, and even though dates can be problematic, we do have approximate dates for most of the images, so I've now set it up to sort by date. That makes it much easier to browse the images with a mind to the way they change over time, and how particular features are concentrated in specific periods.

I think dates should be added back into the main index pages as a sort heading. They used to be in there, and I think they were removed for cosmetic reasons, but I'm going to think seriously about how to put them back without affecting the layout too much.

I turned the outline into a full-scale presentation, and merged in blocks of the original abstract along with blocks of new text (since some things have changed since last year) as accompanying notes. Had to massage the CSS for the presentation again; more problems with font size in the slide titles at the beginning and end.

I've added screenshots at every point where the presentation shells out to the site, or to the IMT, just in case we don't have access on the day.

CC and I went through the images again looking at patterns and dates. We're beginning to be able to draw some conclusions about how they develop through the 17th century.

We also found some bits and pieces that needed fixing -- typos, an engraving with a title that didn't adequately represent it, and an XSLT issue.