PDF application packager for Euclid

I would like to propose a new project: a PDF application packager for EUCLID. It seems like a larger task than would ordinarily be handled in a support request, but I can redirect it there if you'd prefer.

The University is now using EUCLID for PhD applications, and it hasn't been a total failure so far, but it's still far from being properly usable. One big problem is that there is not currently any way to pass around applications to appropriate people for comment or decision making. There's not even a way, as far as we can tell, to get a usable link to a specific application, so that we could send it as a URL. Instead, one has to send a UUN, plus instructions for logging in to EUCLID (if necessary) and how to access that particular student. Then once you get to the right page, it has a bunch of links that each will launch a different browser or viewer for each of the different types of things that might be in an application (transcripts, reference letters, degree certificates, etc.). That makes it extremely difficult to page back and forth quickly through an application getting a general idea or looking for specific information.

This completely unsuitable interface is a huge issue for the DTC, as we receive many dozens of applications each year, which are each reviewed by 5-6 people. In the past we've dealt with it by having our administrator print out copies of each of the applications, but this was a huge waste of paper. It also took her an inordinate amount of time, because each of the bits had to be printed separately. It also often ended up with EUCLID web pages in an entirely unreadable font size, with lines cut off, etc. But of course this is not a DTC-specific problem.

When I have needed to pass around an application electronically, I've been able to do it only by manually downloading each part of the application, converting each part to PDF using standard conversion tools on DICE (e.g. printing HTML files to PDF in Firefox, saving PDFs directly to disk, converting .jpg and .png to .pdf using ImageMagick 's convert, etc.), and then joining them all into a single PDF using the pdfjoin shell script. The result was a single PDF that was very easy to page through and evaluate, easy to print, and even allowed zooming in when EUCLID's HTML code resulted in unreadably small fonts. But it takes a very long time to generate one such PDF manually, and moreover our admin staff don't usually have the UNIX knowledge to be able to do this, so I've had to do it all myself.

Given that EUCLID appears unable or unwilling to provide a usable interface, I propose that we write our own script to automate the process of generating a combined PDF: given a UUN, it would look up a given student's application on EUCLID, download each of the components, convert them to PDF, and join them all into a combined PDF that we could download, pass around by email, or even print out if we needed to (e.g. before an interview). We could then have our internal listings of students (e.g. in the Theon reports by institute) provide a quick link that generates this PDF on the fly, allowing very easy browsing of candidates and insulating us from EUCLID and its horrible ways.

I can demonstrate each of the required steps by hand, so building the script is mainly a matter of automation rather than writing new code, but of course such a script would need to handle various error conditions (such as finding a file in a format it didn't know how to turn in to PDF). There might also be issues with authentication for EUCLID. So it's not a completely trivial task, but I do think it's quite practical to do, and also that it will have a huge payoff for our ability to evaluate potential applicants (both in saving our time and in allowing us to make better decisions).

Here's a quote from Don Sannella, head of the Graduate School, about this proposal:

"On converting applications to pdf: I don't know of a better way to do it than you describe, and it would obviously be a good thing to have. All the tools are available for EUCLID to do it for us, but this is one of many things that we will never get from EUCLID so either we do it ourself, with all of the problems of maintenance etc. that brings, since the only interface we have is the webpage that EUCLID generates and it is subject to change, or we do without."

He's right, there are maintenance issues, but when considering those then please consider all the time that our admins and our academic staff are currently wasting on interacting with EUCLID, and I think there will be a clear case for using CO time to develop and maintain such a tool. There would be an even stronger case for EUCLID to simply present a usable interface, as that would improve things for the entire university, but I think we have to give up on that idea at least for the next five years or so.

Jim (Bednar)

-- AlastairScobie - 01 Jul 2011

Topic revision: r1 - 01 Jul 2011 - 14:10:00 - AlastairScobie
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies