HCMC Blogs

A British academic, tired of correcting the same spelling errors in his students' work year after year, is now recommending that universities simply accept common misspellings, and treat them as if they were normal spelling variants. He has provided a list of ten common misspellings that we should accept to get us started, including "occured" and "thier".

The gentleman in question is a criminology lecturer at Bucks New University (which used to be known as "Buckinghamshire College of Higher Education" -- is "Bucks New University" a variant spelling?).

I'm currently locked out of three of my projects, because I'm a member of too many groups on TAPOR. Among them are two I don't even recognize, "jr" and "solid". "jr" has no home folder at all, so I have no idea what that is; "solid" appears to have been created in July, and I was added to its group, but there's no documentation on the blog to suggest what on earth it might be.

Single-handed in the office, so I'm doing long days this week.

I now have JPEGs of all the TIFF files in the three sizes required, and everything is burned onto a set of four DVDs. I started uploading the smaller ones to the Coldesp site on Lettuce, but the upload stuck half-way through; I think that's because we've run out of space, and I've written to sysadmin about that. I would put the images into the coldesp account in home1t, but I'm unable to do anything in that account; according to groups, I'm a member of the coldesp group, but it looks like it's the 17th group in my list, and I remember some problem with a limit of 16 groups. There are a couple of groups earlier in the list which I don't even recognise ("jr" and "solid"), so I'll have to investigate those. Meanwhile, I'm copying the downsized versions to my local drive E, and I'll follow up tomorrow with the full-sized versions.

After reading some comparisons, and looking at the documentation of both, I'm now considering JQuery, for these reasons:

• It's supported on web.uvic.ca, so other UVic coders will be more familiar with it.
• It has better, clearer documentation.
• It has some slick UI effects that will be useful in a multiple-form interface.
• It's not much slower, despite being simpler to use.
• It does enough with AJAX to cover my needs -- not as much as prototype, but plenty.

So I'm now working through the JQuery docs too.

I've more or less settled on using prototype.js as the basis for GR's RomanLaw database system partly because of the sophisticated AJAX Request object it has. It can automatically evaluate JavaScript code which the server returns in response to a request, so the application can more easily handle the addition of new items to existing tables on the fly, while editing or creating a new record in another table. This is the sort of sequence that will be required:

1. User creates a new Inscription item (destined for the inscriptions table).
2. User wants to tag this with a new keyword, because the required keyword is not in the listbox built from the keywords table. User has already control-clicked several keywords in the list (there's no limit to keyword tags for a given inscription).
3. User clicks on Add Keyword, and gets a little sub-dialog.
4. User completes sub-dialog and submits; AJAX sends request to server, including both the new proposed keyword, and the list of any ids of existing keywords which have already been selected by the user (in a listbox).
5. Server stores this information, then creates a new keyword item in the table. It responds with a new list of keywords for the listbox (ids, strings, and selected status), in alphabetical order including with the new item and its id. This list is in the form of some JavaScript which clears the existing listbox items and replaces them with a new set. Previous selections are preserved, and a the new item is selected; it's also scrolled into view (assuming that's possible).

This all looks feasible using prototype.js right now. I'm currently reading all the docs to get a good handle on how best to use it.

R from ETCL reported that the server she had been using, chou.hcmc.uvic.ca, had apparently gone down. Searched all the documentation and blogs for any info on Chou, but nada; R remembered its being in our server room, so we went in there and started up a black, unlabelled machine on the top shelf. Chou came back to life at this point, so we concluded that this was the one. It would be good to have a bit of info on Chou. I was able to SSH in as root, and list the users; there are lots of them, including ME, JD and R from ETCL, and several names I don't recognize at all. Looks like this machine might get quite a bit of use, so we should have some info about it for when Greg's away.

Quick tweak to Medieval Studies site for SA.

I'm booting up both labs first thing every morning because there are classes in both at 8.30, and they seem to be showing up. Today all the Lab B machines booted fine, and the only oddity was in Lab C, where STN 18 was very slow indeed.

The existing documents are over 20MB each, but JPEGs at good quality are only between 3 and 4 MB, so I'm building a set of those using PS batching. This batch has had to be restarted a couple of times, because I had trouble setting it up so that it saves the file only once, but closes it when it's done; if the Save operation is part of the Action, then it leaves the file open in PS, but if Save isn't part of the Action, and is configured in the Batch setup dialog, you're not really sure whether it's saving with the correct parameters and settings. What actually happens, then, is that it saves twice, once as part of the action and then once at the end of the Batch process. Finally I twigged that I should put the Close operation at the end of the Action, then remove it from the Batch process. Doh. That's now running, and logging errors to a file instead of stopping on error, which will at least make it more straightforward to fix any oddities at the end.

When that's done, I'll need to get the data off the Mac where it's sitting, so I'll burn DVDs as appropriate.