Scandinavian-Canadian Studies

Typo in the Ostrauskaite text.

Finally found a fix to make the sidebars long enough in IE6. Basically, you set it like this:

var heightForSidebars = document.body.scrollHeight;
        if (document.body.offsetHeight > heightForSidebars){
            heightForSidebars = document.body.offsetHeight;
        }

This sets the height to the height of the document, or the height of the window, whichever is larger; when it's a long page, the former (scrollHeight) is larger, and when there's little content, the latter may be longer.

On the actual ScanCan site, the problems are slightly more complicated, because there are lots of interactions during the JavaScript that lays out the page. If the content of the scrollbars changes during the layout, then we need to check their height to see if either of them is now taller than the heightForSidebars we've already calculated, and take the largest.

I also had to add a call to the CheckCompatibility() function in the ajax.js file, so that when the AJAX code writes material to the page, the page is laid out again, since its length will often change.

The results are good enough for me to feel confident in suppressing the warning for IE6 users. The site basically seems to function correctly on IE6 with this compatibility code.

• Look at this post for details of work and decisions required on the biblio.
• Fix the the three biblio items in the Godin paper which have been the subject of email discussion -- see the emails from April 4.

New abstracts came in for Godin, so I added those, and she'd found a typo in the text, which I fixed.

Copied the vol 15 cover to a new file, and double-checked all the sizes and crop marks -- they were indeed correct, so we can keep them for the next volume. Changed all the dates and vol numbers, but left the old TOC in place as an example. We'll replace it later when the new TOC is available.

This was an interesting process because it raised a number of issues that need to be addressed for teiJournal; we need a consistent, clear and simple approach to marking up quotations, citations and references, and working on ScanCan makes these problems more obvious.

The three articles I've marked up recently have a few online sources in the reference lists, something which didn't happen in Vol 15. This has raised a lot of questions, which I put to John. The online citation code is basically unchanged from ACH, so it's not consistent with Chicago, and we'll need to update it. I spent a lot of time reading the relevant bits of Chicago -- it's both complicated and vague, so there are lots of possible ways to approach the problem, but I think ScanCan references fit into the type B (bibliography) or R (Reference list in author-date style), which are pretty consistently similar. Chicago 17.237 is relevant for most of these informal online sources.

What we'll need to do is:

1. Decide on a consistent policy for styling them.
2. See if they can be made to come out that way just by marking them up in a particular configuration without changing the existing XSLT. If so, document the markup format.
3. If not, decide on how to identify them (perhaps using the type="WWW document" designation on the title, which we used in ACH), and then figure out how to trap for that designation and invoke a different template or templates.

As I was marking up the Godin article, which has lots of direct quotes from the speech of study subjects, I noticed that we have always marked such speech up using a <q> tag inside a <cit> tag. The TEI Guidelines (both P4 and P5) suggest that a <cit> tag should only be used if the quotation is "together with a bibliographic reference to its source"; otherwise, we should just be using a plain <q> tag. I began the process of correcting this, but then realized that it's actually very problematic because so much of our rendering code depends on the presence of the <cit> tag. At this stage, it would be very complicated to rewrite it.

A simpler solution, therefore, might be to rewrite the P5 transformation so that when the P5 XML versions of the articles are rendered, the problem is corrected, with the external <cit> removed, and its rend attribute (which we often need to specify that quotes are rendered as blocks) transferred to the <q> tag. However, the more I looked into this, the more I realized that we have <q> tags whose associated <bibl> tags are elsewhere in the same paragraph; in other words, it's not always easy to tell whether a quote has a bibliographical reference or not. On balance, it's better, I think, to live with the unnecessary <cit> wrapper. We're slightly at variance with the Guidelines, but it's only a minor issue, and it makes processing much easier.

So this post amounts to public documentation of that variance, which may have to be addressed in the distant future.

Our XSLT for PDFs has followed Chicago in rendering the <soCalled> tag (usually used for "scare quotes" etc.) with double quotes, but our XHTML code, inherited from the ACH project, was still rendering them as single quotes, following instructions from Peter for that project. Fixed the XHTML so that it's consistent with the PDF.

This one has multiple sections, and lots of use of the <mentioned> tag for linguistic discussion.

The paper lacks bio and abstracts, but the rest is complete. In the process, I found and fixed a couple of bugs in the bibliography rendering, and also found and fixed a couple of typos in Waerp, which I was using as a model.