(This is Arlene). It’s not often I indulge my need to complain in quite such a public setting, but I’ve spent a lot of time over the last few days working on additions to the Alaska’s Digital Archives and, well, I’ll say it, ContentDM is driving me nuts. Don’t get me wrong: I like the searchability of ContentDM and the ability we have to design the description that goes in and a lot of what it does makes our work very efficient, but some of the programming peculiarities make what should be simple tasks into extensive and time-consuming tasks. This is a great piece of software and it really fills a need for us. So why then, does it have these odd little idiosyncracies? Some of which could probably be debugged and fixed really easily?
Essentially I’m taking large groups of photographs with minimal information and loading them onto the Alaska’s Digital Archives as compound objects. There’s several benefits to this: first since we don’t have all the location detail and dates for each of the images, but do have them for the collection, we can do some collection level description and provide that and context to each image by connecting them together into a compound object. Last week I loaded an album with 256 images, this week’s project had 290. Follow those links if you’d like to see them and get an idea of how a compound object looks different than the regular interface. What really should have taken less than a half-hour per album (especially since I wasn’t adding any item-level metadata) took an average of 6-10 hours of my time. Per album. And all of that time was repetitive data entry or mouse clicks. Years ago when I was (briefly) a computer programming major in college, I’m pretty sure that one of the basic assumptions of programming was that if it was repetitive, it was a prime candidate for automation. Or is that an obsolete viewpoint? (I hope not or I’m really showing my age here.)
So, my ContentDM colleagues, if you’re reading this, here’s a list of my requests. My top 8 candidates for upgrade, so to speak. If this is something that can be controlled from our end of the system, will you please let me know so I can get in touch with my colleagues and our sysadmin to get these things fixed?
1. Stop assuming that the title is the file name. Does any institution title images with what they use for a file name? Really? Usually the file name is going to be an accession number, if the institution accessions materials on an item by item basis, or it will be some sort of an alphanumeric name with reference back to the collection call number or such. The title field is going to be a description, a caption, an actual title. We do use the file name as the identifier, so each individual item up on the Digital Archives site has a unique identifier, but please allow us to pull that data into our identifier field–or to choose what field we want it to inhabit–instead of having to rewrite the title field with every single record that we enter. The template allows you to pull the file name into whatever field you like, but then you have to erase the file type extension. Okay, so this isn’t peculiar to compound objects and it’s been an annoyance since we started using DM about 7 years ago, but it did add about an hour of data entry to my work on each of those albums since I had to touch each image in turn to move the file name (minus the extension) into the identifier field and put some sort of a title into the title field. Minimal description is supposed to make my work more efficient, not reduce me to doing repetitive data entry. Oh, and by the way, when you fix this so we can choose what field the file name inhabits, please allow us to set that as a required field in LIEU of the title field. One of our partner institutions has taken to ignoring the identifier field when doing these minimal description projects and leaving the file name/identifier in the title field. I wonder if that doesn’t maybe confuse our users sometimes, but I have to admit, it’s an awfully tempting workaround. Even if it does negate the whole reason for having structured metadata.
2. Why can’t any of the sorting mechanisms cope with a combination of alpha and numeric? When I look at the file structure on my computer drives, those present them in proper order so somebody must have figured out how to do this. But here we get abc1, abc10, abc101, abc11, abc12, and so forth. When everything in DM is defaulted to sorting by the title field (that needs to be optional too) this results in some very strange arrangements. This also feeds into my next issue in compound objects, which is ordering the materials within:
3. Please, allow us to move more than one item at a time when reordering the elements of a compound object. (actually, solving #2 would probably take care of this for me in this instance, just not in others.) The default sorting means “Image 90” to “Image 99” are located in slots 290-299 in your 299-item object because of the alphabetical sorting of the title field that occurs when you create the compound object. Add to that the fact that every time you move an image you get taken back up to the top of the list and then have to page back down to the end to find the next item to be moved, well, you can guess how long this takes. This sorting of images 1-290 in the administrative interface took me 3 hours of tedious finding the image to be moved, clicking on Edit, typing the location #, clicking on Submit, and using the scroll buttons to get back down to the next item that I need to move back up to the top of the item list. And worse, I’ve done it with materials that have actual titles that end up being sorted alphabetically, which means that I have to consult some sort of an item level listing to figure out which of the titles goes next in the list and then hope I see it as I go scrolling through at high speed. At least with a title labeled “Item #x” it was fairly easy to determine at the end whether or not they were in the correct order. Why can’t I just edit all of them simultaneously and put them in the places I want them? Oh, and that reminds me:
4. Please make keyboard shortcuts available for all buttons on the administrative interface. Every time I have to take a hand off the keyboard and move it to the mouse to click submit, I lose time. Again, not necessarily related to the compound object thing, but I was having to do this with every Edit, Move, type in location, Submit, scroll down to find the next item that has to be moved. That probably could have shaved about an hour off that rearrangement time last night.
5. In the edit function in the administrative interface, could you please give us the ability to search by strings as well as by whole words? It must be possible, you’ve figured it out for the find and replace function.
6. Give me the ability in the find and replace function to first choose a subset in which to work? I had need to add 2 words (Album #) to each title in the two albums last night. I hadn’t done it in the original load because I was getting tired of all the typing and forgot it would be necessary to separate album 1 (already online) from album 2 online when I went to build the compound object in the administrative interface. But since I couldn’t choose a subset of images in which to work, I took a deep breath and hoped that the word “Image” wasn’t used in any of my other 3000+ item titles and just did a global replace. And then, of course, since I couldn’t segregate each album out to do one as Album 1 and one as Album 2, I did them both as Album 2 and went through and individually replaced the 2 with a 1 on each image within album 1. (My workflow was logical in a weird way: since album 2 had more images than album 1 I reasoned that I’d have to rewrite the titles of fewer of the images if I chose to name them all Album 2 than if I’d named them Album 1.) And okay, this added another 3 hours of work.
7. And there is one other little annoyance that may be peculiar to only some of us with fairly complex ContentDM instances: please allow me to choose a default collection that appears when I first log into the administrative interface. And allow me to change that at will. Because of my work with several partners, I have administrative rights on 8 different collections and that number is going to grow. And my institution’s collection is number three on that list. So every time I log into the admin interface, I have to go to that collection field and change it from one of my partner’s collections to one of mine. It’s not so much the time it takes, it’s the remembering to do it at all. One of these days I’m going to rewrite some controlled vocabulary for a partner thinking I’m doing it only for my collection. That can’t be good.
8. And in terms of the end user: why can’t the individual records that make up a compound object look just like their individual compatriots in the rest of the database? We need to be able to see both the object and the metadata at the same time. Having to actually click through to see the metadata independently from the image? I don’t even do that, I can’t imagine users less adept with this sort of online database are going to be able to do that, assuming they even understand that it’s there to do.
Other than that, am really liking the new project client. I haven’t figured out yet how to pull materials from online back into it so I can fix things (like the 256 images that I uploaded last week, none of which have our band on them because I thought I knew how to apply a band but it turns out I didn’t and I didn’t discover that until after I uploaded all the images online) but I figure that can wait. Somebody will explain it to me eventually. That’s, of course, the benefit to being willing to tackle technology even if you’re not particularly technologically adept. (Note that I’m an archivist coming out of a history discipline, not the computer programmer I started out to be those many years ago). Eventually somebody gets annoyed with all my dumb questions and either walks me through how to do it or fixes it so they don’t have to walk me through it. After all, it’s for the benefit of all the users, right?
And apologies. I just realized that this is probably the longest blog entry I’ve written to date. Take it as a sign of how dedicated I am to the Digital Archives project (after all, at least 4 hours of work on this last night was pretty much on my own time.) I don’t get frustrated with things I don’t care about.