April 2013

ANVC Scalar logo

ANVC Scalar looks very promising:

Scalar is a free, open source authoring and publishing platform that’s designed to make it easy for authors to write long-form, born-digital scholarship online. Scalar enables users to assemble media from multiple sources and juxtapose them with their own writing in a variety of ways, with minimal technical expertise required. …

The feature list and the showcase look great. If this tool is even half as good as it seems to be, the world will be a better place.

There’s just one problem: the code isn’t open source.

I couldn’t find the source code linked to from their site. [Update: I eventually found a link to it, when I re-trawled the site one last time after having already written most of this post. The link is from their sign-up page, but the license stated there, the ECL-2.0, is not the same license as actually found in their source code snapshot. See below for details.]. There’s a contact form, which one could use to ask them for the source code, but hmm, that’s not how these things are usually done. The only message I could find about development was this:

Development Roadmap: Scalar is in ongoing development. This spring 2013 beta release provides broad public access to the platform via the Scalar servers (click the orange “sign up” button to get started.) While many authors have experimented with Scalar during our alpha phase, we are eager to roll the platform out to even more users. We look forward to hearing from you.

It’s perfectly fine to be planning to be open source and just not have gotten there yet, of course (though it’s usually a much better strategy to just be open from day one instead of waiting for everything to be perfect before going public under an open source license — the advantages of open source are greater the sooner in the development process you open up). But what the Scalar site says is that the software is open source: present tense. That’s only meaningful if there is source code released publicly under an open source license.

I finally resorted to Google: search://github anvcscalar/ (after github+scalar didn’t get useful results), and found their Github repository at github.com/anvc/scalar:

Congratulations on discovering Scalar, the next generation in media-rich, scholarly electronic publishing!

If you just want to create a Scalar project, the easiest route is to work from our servers. You can register and learn more at http://scalar.usc.edu/scalar/ . Using the version of Scalar that is hosted on our servers guarantees that you are working on the most up-to-date version of the software. During our beta phase, updates will continue to happen with some frequency as features are added, user feedback is incorporated and Scalar continues to broaden the horizons of electronic publishing. If you are technically inclined and decide to host your own version of Scalar, you’re free to customize and modify it in any way, but it’s up to you to download, install and troubleshoot updates as they become available.

We are also very grateful for all feedback based on your experiences using Scalar. We are especially interested to know where and how you are using it, innovative or unexpected uses of Scalar, requests for features, opportunities for future development, potential press, archive or scholarly society partnerships, as well as reports on any bugs or difficulties you may experience. Learn more at http://scalar.usc.edu/scalar/

So, is the open source code to Scalar really here? Well… sort of and sort of not. There is a complete source tree, and an “INSTALL.txt” file whose instructions look like they would get that version of Scalar up and running. But there’s only one significant code commit, from 11 days ago (March 30th): the initial import of a source code snapshot, under their own custom license that does not appear to be an open source / free software license as recognized by the OSI and the FSF. The restrictions placed by the license are not onerous, but I’m not sure the indemnification clause is compatible with the Open Source Definition, and clause 4 disallows anonymous and pseudonymous redistribution of modified versions, which I believe is also incompatible (as well as being a bad idea):

4. Any files that have been modified must carry notices stating the nature of the change and the names of those who changed them.

There are other potential problems with the license, but I won’t go into them here. The point is, this is not an open source license. So, stacking up the situation:

  • You can’t find the code from their site, at least not from the expected places. But if you’re persistent and use a search engine, you can find it.

  • They’re not doing development in the open. Instead, they’ve dumped one code snapshot out to the public, and it’s not clear at what intervals they will put out the next ones. There’s no public forum for development discussion, nor is there any public bug tracker. (Alternatively: maybe these things all exist and I just couldn’t find them?)

  • The license is not an open source license, though this appears to be more by accident than design — they clearly do intend to be open source. The best way would just be to use a recognized open source license, because even if they fixed the issues in theirs, people would still have to learn yet another custom one-off license. (There’s no need for them to spell out the trademark protections as they do because a copyright license does not imply trademark permissions anyway.)

It’s the point about not being under an open source license that makes them “officially” not yet open source. But the other points are important too. While you can be technically open source while doing development in a closed manner, why would you want to?

None of the above denigrates their technical achievement so far: the software is pretty exciting, and I hope these issues get resolved soon, because it would be great to have a tool like this available as open source software. I took the time to write about Scalar both because the project looks so interesting, and because what their current situation (open-source-wise) is not uncommon. We often see projects using the words “open source” without quite getting the tune. Fortunately, it’s easy to fix if they want to.

I just read Evgeny Morozov’s critique of Tim O’Reilly in the Baffler. It misses its mark pretty widely. I know Tim, and have worked with him on some of the things mentioned in the piece, and I don’t recognize the man Morozov thinks he’s writing about.

The article is profoundly intellectually ungenerous. If there are N ways to interpret something, Morozov picks the one that most matches his thesis, whether or not that’s the interpretation that makes the most sense in context. He also indulges in guilt-by-association and guilt-by-superficial-similarity. So what that Eric Raymond likes guns? So what if some things Tim says are similar to some things Ayn Rand said? Does that mean Tim O’Reilly is any closer to being a libertarian, second-amendment-quoting Randian? (He’s not — far from it. I’m not either, but some of my friends are. I wonder what Morozov would say.)

I’m increasingly disenchanted by Morozov’s apparent belief that he is a more careful and rigorous thinker than, well, everyone. This piece contained a great example of why. Morozov starts out by quoting O’Reilly:

Expanding on this notion of “algorithmic regulation,” O’Reilly reveals his inner technocrat:

I remember having a conversation with Nancy Pelosi not long after Google did their Panda search update, and it was in the context of SOPA/PIPA. . . . [Pelosi] said, “Well, you know, we have to satisfy the interests of the technology industry and the movie industry.” And I thought, “No, you don’t. You have to get the right answer.” So that’s the reason I mentioned Google Panda search update, when they downgraded a lot of people who were building these content farms and putting low quality content in order to get pageviews and clicks in order to make money and not satisfy the users. And I thought, “Gosh, what if Google had said, yeah, yeah, we have to sit down with Demand Media and satisfy their concerns, we have to make sure that at least 30 percent of the search results are crappy so that their business model is preserved.” You wouldn’t do that. You’d say, “No, we have to get it right!” And I feel like, we don’t actually have a government that actually understands that it has to be building a better platform that starts to manage things like that with the best outcome for the real users. [loud applause]

Here O’Reilly dismisses the entertainment industry as just “wrong,” essentially comparing them to spammers. But what makes Google an appropriate model here? While it has obligations to its shareholders, Google doesn’t owe anything to the sites in its index. Congress was never meant to work this way. SOPA and PIPA were bad laws with too much overreach, but to claim that the entertainment industry has no legitimate grievances against piracy seems bizarre.

Now, wait a second. Tim was spot-on. He’s pointing out the big problem of representative democracy: the distortion inherent in the transactional, seat-at-the-table model, the distortion that comes from having interest groups with deep pockets. They become, effectively, first-order constituents even though they’re not citizens. Tim reminded the listener that the explicit purposes of copyright law do not include pleasing any particular corporate actors — industry can be a means to an end, in these laws, but it’s not supposed to be an end in itself. If government can achieve the stated ends without making Demand Media happy, then it is free to do so. Of course, that’s understandably difficult for politicians in practice… But Morozov’s refutation isn’t about implementation details. It’s about the philosophy, the underlying purpose:

While it has obligations to its shareholders, Google doesn’t owe anything to the sites in its index. Congress was never meant to work this way. SOPA and PIPA were bad laws with too much overreach, but to claim that the entertainment industry has no legitimate grievances against piracy seems bizarre.

Read that carefully: Morozov is saying that, while Google is only strictly speaking responsible to its shareholders, Congress’s responsibility includes satisfying… industrial/corporate constituents? Not merely as a means, but actually as a first-order end??

Uh? Can he really mean that?

I assume he doesn’t, and that rather he’s just not thinking very carefully. Earlier in the essay, Morozov seemed just fine with the idea of “disrupting someone’s business model”. I guess he’s just not in favor of Tim O’Reilly being in favor of it.

(See how easy a Morozov-style takedown of Morozov is?)

Tim bounces a lot of big ideas around. Anyone sincerely looking for something to criticize could find something useful to say (and in many cases Tim would appreciate it, and even change his mind). Yet when Morozov gets close to one of these things, he shies away from making an effective criticism, and instead opts to make Tim’s ideas look bad through shallow, associative analysis, without saying outright what would be a better idea. Morozov provides no constructive analysis; he wants someone to be wrong, but he doesn’t particularly care what’s right. This is just Andrew Breitbart for intellectuals.

Big ideas have porous boundaries, but that isn’t the same as being meaningless. A good critic recognizes the useful big ideas, and after puncturing them helps define their boundaries better, or else counters with other ideas — puts something on the line, actually comes out and says something capable of being refuted. Morozov never takes the second steps. He plants seeds of doubt, but takes no responsibility for the crop that results.

(Update: O’Reilly himself seems to have had a similar reaction to mine.)