Caveat Lector

I was taking a break from moving this morning when I saw an immense tiger swallowtail butterfly land on a leaf outside the office window and spread its wings to bask. Real museum-quality specimen; David found the binoculars and we got a close-up look.

An hour or so later, Dream got up on David’s desk and started scratching frantically at the window. I figured a bird had landed in the wisteria, but no—Dream had found a little brown bat, tucked up to sleep between our ill-fitting storm windows and the inside windows. I thought I’d heard something thumping around in there last night; now I know what it was.

Our batly neighbor can’t get in, can get out, and where he is he’s quite harmless except to insect life, so I’m inclined to let him be. Dream was contemplating climbing into the air-conditioning vent in hopes it would lead to the bat, but he eventually gave up in disgust.

The bird life is mostly robins, sparrows, starlings, doves, and cardinals, but I’ve seen a yellow-shafted flicker. And coming home from a restaurant run a week or so ago, we found several hop-toads hopping about the grounds.

Shame about the condo management here, truly, because I do like the neighbors.

I know I’ve said this before, but it bears repeating: bloggers are good people.

Andrew came over earlier this afternoon to help empty the pod. We’d never met, but he lent us his time and quite a lot of energy to help us out. We owe him dinner in addition to eternal gratitude.

The more so because we miscalculated—that immense Psychic Shopping Networked couch of ours just would not fit through our door, twist and turn it howsoever we would. Of course, since our door is on the third floor of the non-elevatored building, David and Andrew endured an awful lot of work before we learned it was hopeless. (And afterward, for that matter. Getting it back down the stairs was no picnic either.) We found a spot to leave it where it isn’t a public eyesore, and we’re going to try to freecycle it.

But everything else is out of the pod, thanks to Andrew; now we just have to keep the jackbooted-thuggish condo management around here from trying to tow the pod before it’s picked up on Tuesday. (They are Tagare Management, they’re quartered in Woodbridge, and I would beware of them; they really do go out of their way to be unpleasant and uncooperative. I’m terribly glad we’re only renting here, and once we’re on the market, I refuse to buy a condominium. Townhouse or nothing. I’m just not putting up with this garbage.)

Another friend-I’m-about-to-meet is on her way now, to take me out grocery shopping now that I have a more-or-less functioning kitchen. (”Another kitchen box? Where am I going to put it all?” I wailed earlier today.) Bloggers? Good people.

And we’re pretty much moved in, thank goodness. I wonder how many boxes we can avoid opening altogether?

The POD arrived on time, and if the horrible management in this complex manages not to give us a hard time about it, all will be well.

(David did everything but stand on his head to keep them informed and cooperate with them. They were uncommunicative, rude, and unhelpful. There’s a condo just under us going up for sale, and ordinarily I’d be interested, but I will not live where these yobbos are overlords. Just won’t.)

Sleeping in a real bed is very heaven compared to air mattresses and thin foam pads. I didn’t want to get up this morning. Shame about that work thing and all. (Lest you suspect me, I’m at home, snatching a few blogminutes before work.)

Technology is not being good to us. Moving the DSL modem meant disconnecting it, and when we connected it again it lost signal. Hello, Verizon, please fix our line again? (I wanted to try hooking up the other modem, which I’m better at configuring, but I can’t find the thing. It was in the office. I swear it was. I don’t know where it is now.) And David’s Mac is refusing to boot properly. (Open Firmware, eh? Oh.) I think it’s the new video card rather than the move, which we can test as David wisely kept the old card. If I can just get the thing running, I’ll back up that which is most crucial onto CD, and raze the OS to the ground if I have to. Er, when I find the boot CDs, and all.

But having chairs, tables, sufficient unmentionables (don’t ask; I usually pack better than that!), and a real bed still overcomes the crankiness of high-tech. No question.

The pod is coming today, and the horrible weather broke last night, so we won’t be hauling furniture in dangerous heat. This is very good. So is no longer living out of suitcases.

Verizon sorted out whatever was flaking my broadband. I am now connected again. This also is good.

My commute to work becomes free the first of the month. Free is nice (especially for a commute in the DC area!), but being able to save quarters for laundry is even nicer.

The fly in the ointment is that my email is still hosed. If you really have to get to me, try textartisan at softhome dot net. I sense that this webhost is not long for the world… my world, at least. I loathe moving webhosts, but I do rather depend on my email.

DSL blinkenlights still not behaving themselves, though Verizon says yesterday’s problem is solved. I expect much phone angst.

To add insult to injury, my incoming email appears to have been routed to /dev/null. Am taking it up with webhost; will see what happens.

It was a good, bad, and ugly day in DSpace-land today.

The good: After the redirection fiasco (which I have adopted the 80% solution to, videlicet and to wit: auto-redirection from the homepage to the DSpace install), I was leery of touching the dratted server again. But I can’t redesign it until I can get a redesign to stick, so I copied all the JSPs off the main install, stuck ’em on the staging server, and rebuilt DSpace on the staging server.

Then I took a deep breath, and surfed over to see what was broken. Mirabile visu, nothing was; it just worked. Since nobody else was in the office yet, I executed a small victory dance. (Very small. If I dance big, I’ll damage a machine somewhere.)

Shortly thereafter I strolled over to the copyright office, which I’m terribly, terribly glad we have. I left with considerable reassurance about handling potential challenges, as well as several good workshop ideas and a decent new working relationship, which is always a good thing.

The bad: DSpace templates aren’t coded very well. I know, I know, I’m not a Java guru and I have no status to say such a thing. But they still aren’t. Take the navigation sidebar. There are two templates for it, one for the admin pages and one for the ordinary-user pages. This is fine. What isn’t fine is that the two templates use completely different code to get the same effect (those little arrows that you see on an out-of-the-box DSpace install)—the admin template refactors the code into a mini-function and the user-page template doesn’t. That’s just sloppy bad, and it hurts Java neophytes like me most, upending our expectations.

The DSpace developers could take a page out of MovableType or WordPress as far as templating goes. The actual templates shouldn’t have nearly the amount of raw Java code in them that they do, I suspect. Should I become enough of a Java guru to hack on this thing, I think that’s where my effort will go.

The ugly: The markup. Oh my goodness me. Tables inside tables inside more tables, and let’s bung another table in just for the heck of it. Enough to give a text artisan a major case of vapors.

I’ve managed to strip down the markup and juggle some page elements in the five most-used templates. Before I go any further I need to write down class and ID names for what I’ve done, though, or I’ll forget them when it’s CSS time. Dozens of templates to go, unfortunately… I went into my informal Gantt chart at the end of the day and gave myself another week to finish the staging-server redesign. I’m now thinking that seriously wasn’t enough.

I must say, the customer service I’ve endured since moving to Fairfax has been the worst part of the experience.

First there was Verizon (pause for collective groan), which managed to tell me an incorrect phone number (which I duly passed on to all my friends-relations-and-business-associates) and then allowed me to sign up for DSL on that wrong number without ONCE checking the correspondence between self and number.

So, okay, that’s been straightened out awhile, and today I got the email that said my DSL had been turned on. Yay. Unpack modem, plug in… the DSL blinkenlight won’t stop blinking. No engineer I, but I do know that’s a bad sign. But, hey, while I was working, David picked up a phone call that was a Verizon recording telling us our DSL was now on. So it must be on, right?

I do all the little troubleshooting things, and then I roll my eyes and call ’em up. “We’re having an outage,” says the calm recorded voice.

*headdesk* You morons couldn’t have stopped the startup calls in my area from going out?

And then there’s Apple Federal Credit Union. I love these guys. They’re funny, in a Keystone-Kops kind of way. First, we find out that they do not in fact dispense cash from anything but an ATM, so if you have an unusual request (such as, say, quarter rolls for laundry), you’re out of luck, bub. No customer service for you. Back of the line!

I emailed them about this. They confirmed it was policy. Yeah, okay.

My ATM card came the other day. Mine. Alone. No card for my husband. You know, the other name on the account?

*HEADDESK*

To hell with ’em. I’m moving us to United Bank. They, I am told, allow their tellers to dispense coinage.

MPOW needs someone to cope with all things library-web-related. You heard it here first. (I’ve known this was coming since before I was hired, but the official posting didn’t happen until earlier today.)

If you’re a librarian, you know your way around large-scale web development, and you’re looking for work, go to our job site, click on Search Postings, and enter the position number FA501z.

I think the job description is fair and comprehensive, but feel free to contact me out-of-band for more information about MPOW in general and this specific posting in particular. I do especially encourage newly-degreed librarians to give this one a whirl; I’m living, breathing proof that MPOW hires newbies and takes good care of them. (In fact, one of our liaison librarians got noticed by Library Journal this year for his mentoring abilities.)

Also, please pass the word to anyone you might know who fits the bill. Thanks!

To give myself a break from computer hassles, I spent the day getting ready for a meeting tomorrow, writing up talking points for the liaison librarians and looking for MPOW-produced journals. (There’s one very prominent one, a weighty law-review thing, and several smaller ones, one of which is reported to be considering the repository already.) I don’t know that I need to spend a great deal of time sniffing out potential projects; they seem to be coming to me.

And therein lies a problem: namely, DSpace isn’t as good as these projects will want it to be. It does what it does (ingestion, tracking, and bitstream preservation) quite well. It doesn’t do metadata well. It doesn’t do collection-specific user interfaces at all. Its search is so-so. And while it’s now possible to link directly to a bitstream (that’s a “file” for the rest of us) instead of eternally being routed through the bloody metadata, the mechanism is awkward and (in my judgment) brittle, likely to change and thereby break existing uses of it.

(I know, I don’t sound like a librarian when I say things like “bloody metadata.” I’m thinking in terms of user expectation here. If you see a list of article titles from your average database search-results page, and you click on one, what do you expect to see? Hint: NOT THE METADATA. Guess what DSpace shows you, even if you just came in from OAIster or something else that just showed you the very same metadata? Right. Total UI disaster. We librarians, sometimes we love our metadata just a teensy bit too much.)

The real elephant in the closet, though, is what they’re calling “complex objects.” See, DSpace has been implemented to think that each “item” is representable as a single bitstream file. (Jargon, bah. Speaking of which, the first thing that’s leaving MPOW’s DSpace implementation when I redesign it is this “community” business. We don’t have communities at MPOW, nor I daresay at most other universities. We have departments and research programs, how about you? “Community” is fine as an internal catchall placeholder for DSpace developers, but I doubt any actual DSpace implementation should be using the word.)

As I was saying… DSpace copes nicely with the idea that the same essential information can be represented in more than one file format. Got PDF and HTML of the same article text? Not a problem. Same item, two associated files, no big deal. To get all FRBRish on you, DSpace handles differing manifestations of the same work or expression very neatly.

What DSpace chokes on is composite information objects. Take this weblog page as a cheap and easy example. It consists of an HTML file, a CSS file, and a JPEG file (the left-hand sidebar background). There is presently no way to tell DSpace that these are all parts of the same information object and need to be served up together.

I didn’t realize this until my lunch chat with a faculty member last week, when he asked me about archiving whole websites. Hadn’t even thought about it. Thought quickly—and fortunately, I seem to have good instincts, because I said “I’m not sure DSpace can do that terribly well” and I was completely correct.

Okay, so there are workarounds. Sometimes. External CSS can be pulled back into an HTML file. Background JPEGs can be considered “pretty noise” and discarded. When I got back from lunch, though, I was immediately confronted with an HTML journal article containing an information-rich JPEG graph. No discarding that, not no way, not nohow. If you’re not me and you don’t abominate the format, you just print to PDF. Me, I’m considering wrapping it all in a gzip archive and calling that a bitstream. Ugly, but more or less functional, and better for content remixing and reuse.

This problem is actively being worked on. The top contenders for complex-object metadata appear (in my somewhat cursory survey) to be METS and a piece of MPEG-21 called DIDL. (Don’t ask me to spell out the acronyms. Just don’t. Yes, I can, but no, I’m not going to.) MPEG-21 DIDL is apparently cooler and prettier, but it’s also got intellectual-property issues all over the place. If the DSpacers are sensible (and I think they are), they’ll stick with METS.

In the meantime, I’ve got some expectations management to do. And more computer hassles starting tomorrow, when I try to turn my staging server into a clone of the actual live install. It should be as simple as replacing a single folder full of JSP templates and rebuilding DSpace, but heck, I thought getting rid of the port number in the URL would be simple, so what do I know?

Mike picks up on a lot of things I didn’t about Charlie and the Chocolate Factory. I can’t believe I forgot about the lockout (that was dead obvious, wasn’t it?), but I could never have come up with the sharp analysis of the children that he did. The comments are sharp too, saving my own.

Another thing does occur to me in passing: the way in which the children and their parents are set at each other’s throats by Wonka’s framing of the visit as a zero-sum game. The Beauregarde-Salt moment takes on a new sort of intensity, as do the sundry oneupsmanships among the parents. Note that Wonka himself is off-limits as a game piece, having set himself above the game; any attempt to approach him directly (I am thinking of Mr. Salt and Ms. Beauregarde particularly) fails.

That ruthlessness deriving from the game rules sounds all, all too familiar.