Mariage

Got some metadata from CC, and created the XML file. Did the initial transcription of the legend (it'll have to be checked, of course), and added the metadata. Still missing a modernization of the legend (which I'm not equipped to do) and the teaser.

A new engraving has arrived, so I've gone through the process of cleaning it up and prepping it using the Gimp instead of Photoshop. This is the procedure, updated from the previous PS procedure:

• Archive the original image in a backup folder (backup, parallel to mariage_5).
• Open in Gimp. Previously we would have converted to RGB, but I see no reason to do that; I left this at grayscale. Save into the scans folder.
• Using the magic wand, select some background grey or speckle, and delete. Play around with this step until you've got a relatively clean image, without losing detail. You might also use Select / By Colour if that helps without deleting anything important.
• Clean up the edges of the image, along with any egregious surface marks in the picture.
• If there's a BN mark with a ruler, it can be moved in closer to the image sometimes, to keep the size down.
• Crop the image to remove excessive whitespace around the outside.
• Resize the image using Bicubic to create one 800px wide. Save this image at around 90% quality, for an image size of two to three hundred K, with the wv_ prefix, into the scans folder.
• Undo the resize to get back to the original image, then start selecting a thumbnail area; once the selection has started, press the Shift key to constrain it to a square. Copy the selection to the clipboard.
• Click on File / Create / From Clipboard to create a new thumbnail image.
• Resize this image to 200 x 200, then save to the images folder with the thumb_ prefix.
• Copy the main image file and the web view )wv_) file from scans to exist_db/mariage/anthology/images.
• Create a new XML file in exist_db/mariage/anthology/images linking to the full size image file in the same directory, and add all the metadata.
• Update the server directories and db.

I haven't created the XML file in this case yet, because I don't have any metadata at all about this image.

[EDIT BY MDH: I've looked at these and can't find anything wrong in any of the programming code (XSLT, CSS or XHTML output); furthermore, the extra line breaks don't occur in Opera or Google Chrome, so I conclude this is caused by a Firefox bug, and I'm marking the task as Completed.]

Just a reminder to MH to fix the following lines contained in argument tags that include an extra line break when viewed online: Lines 10327, 10358, 10396 and 10534. Thank you!

[This is in the FOREST NUPTIALE.]

1) Lines 10328-10338 (extra line break occurs between "phes" and "In-" in XSLT):

<div>
<argument rend="float: left; margin-left: -6em; text-align: left;">
<ab rend="font-style: italic; font-size: 70%; text-align: left;">
Philoſo-<lb/>
phes <ref target="../../back_matter/references/references.xml#indien">In-<lb/>
diens</ref>.<lb/>
</ab>
</argument>
<p rend="margin-top: 0; margin-bottom: 0; text-indent: 0;">
ſans la liſte des Philoſophes <ref target="../../back_matter/references/references.xml#indien">In-<lb/>
diens</ref> en font pluſieurs cantons de<lb/>

2) Lines 10359-10368 (extra line break also occurs between "phes" and "In-" in XSLT):

<div>
<argument rend="float: right; margin-right: 5em; text-align: left;">
<ab rend="font-style: italic; font-size: 70%; text-align: left;">
Philoſo-<lb/>
phes <ref target="../../back_matter/references/references.xml#indien">In-<lb/>
diens</ref>.<lb/>
</ab>
</argument>
<p rend="margin-top: 0; margin-bottom: 0; text-indent: 0;">
deſquels ſont rentés, és autres, il faut<lb/>

3) Lines 10394-10406 (extra line break occurs between "des" and "Bra-" in XSLT):

<div>
<argument rend="float: left; margin-left: -5em; text-align: left;">
<ab rend="font-style: italic; font-size: 70%; text-align: left;">
Solenni<lb/>
tez au<lb/>
iour des<lb/>
nopces<lb/>
des <ref target="../../back_matter/references/references.xml#bracmane">Bra-<lb/>
mins</ref>.<lb/>
</ab>
</argument>
<p rend="margin-top: 0; margin-bottom: 0; text-indent: 0;">
ieunes,hommes & femmes: le iour<lb/>

4) Lines 10535-10544 (extra line break also occurs between "des" and "Bra-" in XSLT):

<div>
<argument rend="float: right; margin-right: 5em; text-align: left;">
<ab rend="font-style: italic; font-size: 70%; text-align: left;">
Femmes<lb/>
des <ref target="../../back_matter/references/references.xml#bracmane">Bra-<lb/>
mins</ref>.<lb/>
</ab>
</argument>
<p rend="margin-top: 0; margin-bottom: 0; text-indent: 0;">
quels couchent auec les ſœurs du<lb/>

I've set up a user on the db for LSPW to log on with, so that she can upload her documents while I'm away. I'll document it in the usual place. All docs are now owned by admin but with group "ras", so both admin and LSPW can update any document. The system collection remains admin/dba.

As our new document markup projects get under way, we're going to need to provide options for them to view and proof their work in the context of the site, without the documents themselves appearing in the various TOCs. I've set up a system for this based on the use of revisionDesc/@status.

• If a document has status="published", or it has no revisionDesc/@status at all (which is the case for our current published docs), then it will appear in the regular TOCs on the site.
• If it has a different status value, it will not appear in the TOCs.
• You can view the documents with a given status on any of the TOC pages by appending the URL parameter "status=whatever".
• The search system only looks at documents with no status or published status.

This will allow us to look at documents in development without "publishing" them officially.

Created a template with explanatory comments for marking up front matter. In the process, I realize that we should probably move a lot of existing content from the beginning of the body into the front -- dedications, prefaces and TOCs all really belong in front. A preliminary look at the XSL and CSS suggests that this will have no effect on the rendering, thankfully, so it should be a straightforward change. I'll put it on the list of things to do.

Working on creating a template for the <front> element, I discovered that the <argument> tag would be appropriate in <titlePage> and we should use it. However, it wasn't permitted by our schema, partly because we'd never used it in that context, and partly because it was only added to model.titlepagePart earlier this year (our schemas dated from 2009). Also, we had two separate schemas, with a sort of splinter schema used by the Agreemens texts and Complaintes, while the main schema was used for all the others. Time to generate a new schema. These are the steps:

• Copy all the XML files from the anthology into a temporary directory. Also copy in a new file which includes the new feature that you need to make available in the generated schema (in this case, <argument> inside <titlePage>).
• Open a terminal in that directory.
• Run oddbyexample.xsl like this:

  java -jar /home/mholmes/apps/apache-tomcat-6.0.26/webapps/mariage/WEB-INF/lib/saxon9he.jar -it:main -o mariage2010.odd /home/mholmes/WorkData/tei/1.8/xml/tei/stylesheet/tools2/oddbyexample.xsl   corpus=`pwd`/
  

• Upload the generated .odd file into Roma, and add title and description.
• Save the .odd file again from Roma.
• Generate the schema.
• Validate all the existing files against the new schema, before replacing the old one with it.

In the process of validating, I found a couple of errors in files which I've corrected (duplicate @xml:id attributes in two image markup files, and one mistyped attribute name in Sonnet 1609). I also discovered that I seem to have a broken Saxon jar in my cocoon_with_exist setup (which explains many things), so the path above is to a known working one inside the Mariage Cocoon.

Created a heavily-commented header template, and EGB used it to create a header for her document. By Monday I'll have one for the frontispiece, which will of necessity be more exemplary than prescriptive, because the structure and layout of frontispieces differs.

Metadata display functionality has been added to the images and the editorial articles.

I've implemented the metadata box in the images slightly differently; it's position: absolute; z-index: 1000; to make sure it pops up on top of the image, rather than displacing the image and making all the zone offsets wrong. I also had to override the default float: left; setting of the popupCloser for the box; in text, and here, it's float: right. Maybe that should be universal, for consistency.

I've implemented a metadata display system that works like this:

• In text documents, a menu item "Métadonnées" is added.
• Below the menu, a hidden div contains metadata.
• Clicking on the menu item hides and shows the div. There's also a little X in the top right of the div which closes it.
• The div contents are generated from the teiHeader.
• Each node or text element in the header is processed. In the case of a node, a div is created, and the node's local-name() is used to look for a possible caption in a specially-configured variable in the project_globals.xsl. If a caption exists, it's used as a label.
• In the case of text, it's simply output as text.
• The encodingDesc node is filtered out, because anyone interested in that information will most likely prefer to see it in full, in the XML itself, while to others it will be confusing.

We have a couple of problems that should be resolved ASAP:

• Our metadata is currently a blend of English and French. It should all be converted to French, I think.
• Some metadata is irrelevant or unnecessary ("An Electronic Edition", or the <extent> tag), and should be removed.
• Some headers are more complete than others.

However, this should encourage us to go back and look again at our headers.