darcusblog

With all the talk of the open standards and free software revolution, the following assessment of Zotero may seem anachronistic:

First of all, Zotero only works with Firefox 2. Yes, not even with regular Firefox as most people use it right now, but only version 2. At the moment, only a release candidate is out, no official version 2 yet. Tilburg University (my employer) has a policy to only support MS programs, so only Internet Explorer is supported. Which means I cannot recommend Firefox to people with only limited computer expertise.

The second aspect of Zotero that needs work, is its integration with Word. The one reason people keep using EndNote, even though they don’t like it that much, is because it is easy to create citations and bibliographies in a text. The integration with Word (again, a Microsoft product which is supported on campus) makes it easy to write an article, insert citations in the proper places, and have EndNote create a good bibliography, using the correct style for a certain publication. Want to submit the article to a different journal? EndNote will change the style of the bibliography for you, in a few seconds.

But it’s actually not. Despite the fact that Firefox is free and better than Internet Explorer, and despite the fact that Zotero is also free and better, this sort of argument is not uncommon at all. It shows just how much weight the MS monopoly places on the market and on innovation. Now guess how hard it is to convince anyone to use OpenOffice.

OTOH, there’s something so incredibly small-minded about all this hand-wringing. The point about integration with Word is not exactly wrong, but a) it will happen (it’s on the roadmap), and b) there’s something so defeatist about the tone. Zotero is free software; if you’re worried about integration with Word — or any other application — then stop sitting around and do something about it! Write a letter or schedule a meeting with the campus IT people who set the brain-dead policies that they will only support Microsoft solutions. Talk to the people who make budgeting decisions and ask them to consider finding a way to direct money away from proprietary solutions into open ones. Talk to your campus computer science department and tell them you want their students to actually help create better solutions. Your institution will benefit in all kinds of direct and indirect ways; you’ll get better and cheaper software, and your students might learn valuable skills that will help them in the future.

If all of us simply worried about what it would take to change the world, nothing would ever get done. Zotero, and CSL, and OpenDocument, and OpenOffice, and RDF is going to do just that in this little corner of the scholarly world.

Got a question recently from a graduate student questioning whether to go on to complete a PhD. The question reflects the usual anxieties about the difficulties of the job market, the rarified world of academic life, etc. Below is my slightly edited response …

Well, first, I suspect most people who do make it through ask themselves that question at one point in their graduate career or another. I think it’s more common for people who have gone straight through from undergraduate work. Time away from academia tends to one a little more perspective on it.

What is that perspective? For me, academia makes sense for people who really like doing the work; whatever that may be: theory, field work, teaching, etc. That’s not to say there aren’t times when they also hate the work. I certainly find writing, for example, a trial. But ultimately, we’d rather be doing this than practically anything else.

The downside is that you give up the structured 9-5 world. You basically never stop working. A vacation, then, is just a slow-down; something that opportunistic politicians simply don’t get. You think about ideas all the time. Whether that’s a glass half-full (as it is for me) or empty is a matter of perspective (again). From my perspective, what better a job to have than to get paid to think and write and teach about the things that I care about?

Not all intellectuals are snobs, and probably most of them aren’t in fact.

The job market? There’s no denying it’s tough, particularly if you want to end up at a more research-focused institution. Very good people often spend a few years in temporary positions before they land a tenure-track job. But it’s not as bad in geography as in some other fields, and your options open up if you consider more teaching-oriented jobs. Again, though, that’s a matter of perspective. If you don’t like teaching, then that wouldn’t be much fun.

The value of a PhD outside of the academy? This I’m not really sure, but I’d speculate that it’s probably not worth the hassle. The reason is that by definition a PhD is very specialized training, and what you would gain would probably not be very portable. Any training you get with respect to research skills and such you’ve probably already gained with an MA. [this varies by field of course]

So you’re left with questions, among them what would you do if not academia? The bottom line is that you can succeed if you want, but you have to dedicate yourself to it fully. It reminds me of what an architecture professor of mine once said when I had decided to leave an architecture graduate program: “Architecture is really hard even if you love it. It’s impossible if you don’t.” The same might apply to academia.

Recent enhancements in CSL have been a collaborative effort between the members of the XBib and Zotero projects. Zotero developer Simon Kornblith helped a lot to refine the language to make it more powerful and more consistent.

The obvious question is now that there’s finally a shipping product that uses CSL, what next? To my mind, we need a nice 21st century analog to the BibTeX tool makebst. Makebst was a script that prompted users with questions and then created a finished BibTeX bst style file. But the interface wasn’t exactly elegant, and the styles aren’t easily distributed.

It’s been my goal all along that we exploit the web and its network effects to quickly build up a rich collection of citation styles, which can in turn be implemented by any interested project.

So what would MakeCSL look like? It would be a web 2.0 app: nice clean AJAX-ified GUI, smart category features, including category-based feeds to keep updated on the latest styles. To create a new style, the user would find the closest existing style, click a “create new style based on” link, and then be guided through a wizard-like collection of questions, with output preview. That style would then be instantly available for other users.

Now, we just need to figure out how best to build it, and where to host a repository, and so forth.

Zotero went live yesterday, so now everyone can try it. They’ve made a variety of changes since the private beta, most of them good. For example, they addressed my name issues, sort of. They still privilege Western name forms.

On the downside, they’ve not addressed my concerns about dates. It’s still totally unclear how the system processes dates, and it makes it essentially impossible for me as a user to know how to enter them. Moreover, entry for dates is thus inconsistent with entry for names, which is essentially the same UI problem (how to handle flexible content in structured ways).

Finally, they’ve taken a step backward on what had been a fairly tight set of reference types. Here’s the new “add new” UI:

As I said in a private note to Dan, it’s important to first have good generic fallbacks before adding a laundry list of specific types.

Right now, instead of adding generic types like “Document” (ahem, which I asked for) and “Communication”, they’ve added really specific ones like “Instant Message” and “Blog Post.” That approach is not going to scale well. The bottom line is that when you get specific with types, you’re then targeting particular users and communities, and if I have 10 or 20 types I never use, that’s not so good.

[Aside: one way to address the UI problem of the long list of secondary types is to add a type to the main select list once a user selects it from the "more" sub-list a first time, or to allow the user to configure it in the preferences.]

As I said previously, having a long list of flat types is not good design, even in user-oriented GUIs. That Endnote uses this sort of typing is a design flaw, not something to be emulated.

A related comment is that using labels like “Magazine Article” is problematic. It should be “Article (Magazine)” so that all article types are grouped in the list, and it’s easy for me to find quickly.

Oh, they did a really nice job improving tagging support. It now has auto-complete, and adding multiple tags quickly is much easier.

I mentioned previously that I had started looking into how I might be able to challenge an Apple patent application before it was reviewed. It turns out this is not so easy; the patent office will actually refuse to look at public commentary! Somehow that seems deeply perverse.

One patent expert suggested I simply help make the prior art easy to find when the application is reviewed. To wit, here is what I’ve found. It seems the ideas have been implemented or discussed in at least the following:

1. Microsoft’s CF_HTML clipboard format
2. the Macintosh application Palimpsest, produced between 1994-1996
3. X11 systems?

The above is about generic copy-and-paste of metadata and content, but we at the OpenOffice bibliographic project have been talking about the more specific issue of citation metadata for awhile. Consider this excerpt from a 2003 document by David Wilson:

For Quotation Tag handling to become more useful, cut and paste functions should be modified to provide metadata about the source of the text. OpenOffice could do this so that cutting text from a OpenOffice document would include the source metadata – the file name and the details from the file properties, together with the page number or number ranges of pages. This metadata would then be available for the automatic population of the Quotation Tag’s data fields….

It would have nice if my web browser and OpenOffice writer had automatically collected the web page source details so that I did not have to manually have to cut and past these in.

For this process to become practical other applications would need to enhance their cut / paste functions to include metadata as well. A new standard for cut / paste functions would need to be developed.

In turn, this document cites a 2002 mailing list post:

As one example, suppose I want to cut and paste from my web browser to somewhere. If I paste into a plain-text text window, I would like a pure-ASCII rendering of the section of the web page I cut; if I paste into a spreadsheet, I would like separate table cells in the web page to turn into separate table cells in my spreadsheet; and if I paste into a rich-text editor (say, a web-page editor) I would like all the tables, formatting, inline images, and other metadata from the original web page to be inserted.

Finally, I wrote this in August of 2004:

what I really want is to be able to have a button on my web browser that allows me to highlight an excerpt in the XHTML file, click it, and have the DocBook citation code pasted to the clipboard. Maybe with Javascript?

Other than this, then, it seems there’s nothing else I can do but simply hope whoever reviews the application at the patent office finds this page, and comes to the obvious conclusion that it ought to be rejected.

Dan Chudnov has the latest in his Library Geeks podcasts; this time an interview with the Zotero team. They talk about some of the history of their work, as well as the technical design of Zotero. Oh, and they mention CSL, which they are using for citation formatting.