Week beginning 24th August 2014

Another primarily DSL related week this week, with pretty much four out of my five days spent working on the redevelopment of the new website.  Last week I started migrating the ‘History of Scots’ section from the old website to the new one and thankfully I managed to complete this task this week.  I seem to have given myself repetitive strain injury whilst getting it done but at least that’s it completed now.  There were times when I never thought I’d get it done so it is quite a relief to see it all there.  All this retagging text took up rather a large percentage of the week but by midway through Wednesday I was able to move on to other, more interesting tasks.  I fixed the issues relating to the scrolling to highlighted search terms in the entry page.  Previously we’d set it up so that the entry page scrolled down to the first instance of the search term, but occasionally this would then be obscured by the ‘fixed’ heading that appears once the page has scrolled down a bit.  Adding in a margin ensured that there would be enough space to prevent this from happening.  I also updated the highlighting script to make it only activate when the search type is ‘full text’ or ‘quotations’, meaning a headword search loads the page at the very top now.

Other DSL tasks this week included updating the ‘News’ box on the homepage in preparation for the launch and adding in an error message to be displayed in the (hopefully unlikely) event of the entry XML failing to pass through the XSLT.  I also tackled two of the larger outstanding items on my ‘to do’ list, namely adding in facilities to enable users to perform an advanced search in the full text while excluding quotations, and adding a predictive text search to the bibliography search form.  The latter required some extending of the jQuiery UI autocomplete widget so that I could pass multiple variables to my AJAX script (to enable users’ selections of source type and search type to be passed) and also some reqorking of the AJAX script I had previously created for the quick search predictive drop-down so as to enable both text and an ID to be passed.  This allows the correct bibliography page to be immediately loaded as soon as the user selects a bibliography item from the list of possibilities.  It all seems to be working pretty well, although I’ve had to strip out the XML tags that were being returned via the API as the autocomplete widget can only display plain text without further extending.

On Friday I devoted quite a bit of time to continuing updates to the Historical Thesaurus database.  I thought I had solved the ‘duplicate t7’ issue last week but alas it turned out that there were lots of other subcategories that needed to be updated.  I managed to fix the bulk of these but there are unfortunately still some issues with the data that may need manual intervention to be sorted.  Thankfully it’s not a massive number of categories (a few hundred out of 230,000) but it’s still disappointing that the renumbering process didn’t proceed as smoothly as I’d hoped.  Also this week I updated the search so that categories are now returned even if they don’t contain any words, which is something Marc noticed that it wasn’t doing last week.

There are still quite a few tweaks I would like to make to the HT website, but these will have to wait for another time.  I didn’t have enough time this week to rewrite my PDR form, which is something I really must tackle next week.

I gave a bit of advice to a couple of people this week, firstly Wendy Anderson on a new possible project she’s putting together (I can’t really say much more at this stage) and secondly Stevie Barrett in Celtic who wanted me to give some feedback on the technology they are proposing to use for an upcoming project.