Talk with Luiz Amaral about his research (natural language processing and student modelling) and possible ways HCMC could support him. Also spent time with him and Ali talking about how his research and language teaching is supported in the CALL.
January 23, 2007
hist.js file contains all the course listing information. I'm not sure whether that file is edited manually or whether some other file automatically creates it. Couldn't find such other file on site, but wrote Clifton to ask.
Wrote list of issues for data prep people to consider in advance of meeting scheduled for Feb 1. Mostly how I'm normalizing the data and the implications of those transformations.
January 22, 2007
COMPLETED...on January 29.
John wants to get his students to use the Image Markup Tool to do some basic markup of images for the Victoria's Victoria Website. He's asked me to come into his class and do a 90-minute session on January 29 (2.30). I went through the app and its features with him, pointing out that it's really intended for people who understand TEI markup, but he still thinks it's worth having a go, so he'll send me an example picture soon, and we'll plan out some basic activities to get them going.
Certain courses (e.g. HIST 265A) appeared on the web page as offered, when in fact they should have appeared as Not offered. Problem was a test in the javascript which tested parseInt(string) rather than the string itself. The parseInt test returned true when HIST 265A and 265B were compared, when it should return false. So, where 265A is offered and 265B is not (or vice versa) both appear as offered. Bug fixed.
January 19, 2007
Karin's really struggling because the site design she has to implement for the IAUPE site is a 1990's-style retro thing, that would way-back-when have been implemented in nasty tables. Implementing it in modern, accessibly, clean XHTML and CSS is proving virtually impossible, especially with IE bugs. She bounces CSS ideas off us when she gets stuck.
Did a large number of Regular Expression searches on the three data files to extract all the subject codes embedded in the documents (defined as \r(\w{1,16}/)
Returned a list of those to John Lutz, along with a count of records that have multiple subject codes.
Recommend he get his work studies to rationalize and normalize those codes (unless the idiosyncratic coding in there now is part of the data to be maintained).
- publication dates at start of text field to be moved (rather than copied) to the publication_date field
- cemetary plots at end of text field to be moved (rather than copied) to the cemetary field
- records with missing publication dates will take the most recent previous publication date rather than the earliest subsequent publication date
- try to identify intances of page numbers at start of records and prepend those with the string "page "
- try to figure out a way to extract compound subject codes from start of record and put them into the subject field with a space delimiter
- compile list of subject codes found in the three data files
- John to ensure that students are using consistent subject codes