~ Git Push ~

Lots of people seem to be disagreeing with things I said in my last post. That’s to be expected; I’m wrong a lot. But some of the disagreement appears to be with things I didn’t actually say. Or said poorly and intended in a different way. Patrick Honner, as is his way, summed it up nicely:

@Lustomatical Generally speaking, reaction to your piece misunderstands GitHub, focuses too much on "lesson" as unit of collaboration.

— Patrick Honner (@MrHonner) June 9, 2016

I will take all of the blame for the part about misunderstanding GitHub. In retrospect, I definitely assumed too much familiarity with the tool. I was trying to avoid bloating an already long piece with a GitHub primer (of which there are several good ones), and that led to some understandable confusion. A thousand apologies.

I will take some of the blame for the part about focusing too much on individual lessons. I did talk about individual lessons a bunch, and those passages seem to have stuck. But I am also going to wag my finger at Dan (in a gentle and loving way) for this tweet/post:

Why Secondary Teachers Don’t Want a GitHub for Lesson Plans. Responding to @lustomatical's proposal. https://t.co/DnMjUgUCdJ

— Dan Meyer (@ddmeyer) June 8, 2016

I definitely did not call for a GitHub for lesson plans. I called for a GitHub for curriculum. And I laid out what I think is a careful line of reasoning about why those are very different beasts. Unfortunately, this turns out to be one of those situations where one word changes things in a big way. So it goes.

At any rate, I’m going to respond to a few general themes in the comments and try to clarify some of my remarks. I’m sure many people will still quite reasonably disagree, but hopefully with a more accurate snapshot of my brain state.

§ We don’t need another lesson sharing/download site.

I agree with this. Mostly. A good one (whatever that might look like) could at a minimum increase the density of quality materials over the average Google search. I suspect that’s why people tend to like virtual filing cabinets — at least in theory — because it comes with some sort of built-in quality control (this person, on average, has things I like). That’s not nothing. But that’s also not at all what I’m talking about.

For one thing, I don’t think “lesson” is the appropriate unit of account. I used lesson examples in the original post because I think that’s the level at which teachers spend most of their authoring/adapting efforts. But I also think those efforts, to be maximally useful, have to take place within a curricular framework. In particular, I’m not suggesting a (not so) new way to cobble together your own VFC. I’m suggesting that, since teachers take ownership of their curriculum as a matter of course — very often in collaboration with colleagues — an efficient system for facilitating and managing that process is probably useful. I think that sentence is the minimal faithful reproduction of my claim. On top of that, I think a GitHub-like system could have other benefits (some below, some in the original post).

@ddmeyer stuff like lesson study is so valuable, though. But lesson repository I guess feels like the worst of both worlds.

— Rachel (@Seestur) June 8, 2016

Many people are balking at the word “repository,” which has sort of a negative connotation in common usage that’s totally absent in the sense of something like GitHub. As in, my garage has become a repository for broken appliances, which isn’t awesome. And which also describes large portions of the internet. But a repository in this model is a living, breathing, collaborative thing. It’s my fault for not making that clearer from the get-go for people who don’t regularly use GitHub. This is not a holding tank. This is a system for integrating and tracking changes and that preserves important information typically lost to entropy.

But, hold on…

§ GitHub is confusing as hell.

It certainly is. That the word “repository” has a highly domain-specific meaning is but the tip of the confounding iceberg that is GitHub. It’s an intentionally leaky abstraction layer over a version control system called Git, which is itself a sieve-like abstraction over a particular data model.1 A quick peek at the documentation should be enough to make your eyes bleed.

Rest assured, that’s just as opaque to a large group of people who use Git professionally as it is to you. Git persists,2 in spite of its basically non-existent user interface, because under the hood it’s very clever and efficient.

There’s a thriving cottage industry making tools that sit on top of Git to make it easier to use, with varying degrees of success. I think GitHub actually does a pretty good job. In my mind, the relevant features are:

  1. Discussion lives in the same place as source material, even inline if you want.
  2. Great support for documentation, including wikis.
  3. It’s trivial to compare differences across versions.
  4. Tracking for issues, goals, and milestones.
  5. Ease of integration with other tools.
  6. Total transparency in terms of who is doing and has done what, and when.
  7. A complete history of every state of the project. Time travel, basically.
  8. Parallel development that doesn’t involve stepping on everybody’s toes.

None of the confusing stuff about Git or GitHub actually matters in this discussion, because I’m not advocating for the use of GitHub. But a system that implements something like that set of features supra (and, since somebody would have to build this thing, whatever else we might want!) would be pretty useful.

Then again…

§ Lessons/students/teachers aren’t software.

Nope! But, again, I’m not necessarily in favor of GitHub as the ultimate solution. More importantly, I don’t think disanalogies between curriculum and software are particularly germane. GitHub is a platform that’s good for managing changes in large, complex, (primarily) text-based projects involving a bunch of collaborators. That it’s mostly used for software is a historical accident. Yes, it is incredibly well suited to the idiosyncrasies of software (by virtue of said historical accident), but it strikes me as a huge missed opportunity not to think about how a similar system could be suited to the idiosyncrasies of curriculum.

Okay, but even then…

§ Other people’s stuff doesn’t work for me/my kids.

First of all, I mention in the original piece that modifying stuff is just what teachers do as a part of their everyday practice. But this seems like an argument in favor of a system that lets you modify materials in a coherent, sustainable way.

It’s possible I’m misreading this argument, and people actually mean that they write all their own materials from scratch. I’m inductively skeptical about that, but I’m sure there are superproductive freaks out there who could totally pull it off. If you don’t collaborate on a curriculum, then I can understand why you’d find a collaborative curriculum tool unhelpful. I don’t think that’s a common situation, and I absolutely can’t see how that’s a generalizable solution.

Anyway, irrelevant because…

§ This isn’t actually a solution to any existing problem.

@mpershan @ddmeyer @lustomatical In tech, we used to call this "a solution in search of a problem."

— CheesemonkeySF (@cheesemonkeysf) June 8, 2016

That’s fair. Certainly not all the problems I find problematic are universal, or even widely applicable. But I wish I would’ve had something better last year as the algebra team collaborated on planning and materials with our ad hoc, multi-platform, system with a bunch of almost-right tools. That was a big impetus for the last section of the piece. If that’s not a problem for you, I agree this wouldn’t be a solution for you. And that’s okay; just ignore it. But I also can’t imagine I’m unique in wanting something more.


§ Isn’t this just a way for publishers to exploit labor?

@Lustomatical @abusch38 Yeah. Question remains, tho. The value in either system mostly accrues to publishers. What's the incentive for T's?

— Dan Meyer (@ddmeyer) June 9, 2016

Well, almost certainly yes and no. Every human interaction provides an opportunity for somebody to exploit somebody else. Setting up systems that actually facilitate that opportunity is — I’ll go out on a limb here — a Bad Thing. Publishers certainly have incentives to crowd-source unpaid work, but I think there are incentives distributed along the path from publisher to individual teacher.

First of all, upstream changes aren’t mandatory. Nobody is going to force teachers to do any work for anyone other than themselves. And not even that, really: you don’t have to do anything different if all you want is to receive a curriculum like the good ol’ days. But a good upstream change will then make its way downstream to other users. So yes, an individual teacher might do some work that benefits the publisher, but also benefits the entire user base. Maybe more to the point for most, it would just be easier to contribute changes that benefit your local community — however you might define it. The incentives for community contribution are roughly the same as they’d be anywhere, and there’s evidence those incentives matter to people.

There are also folks in between who handle curriculum coordination at various levels of school hierarchies, but since I’ve never been in any of those positions, I’ll just gently speculate that some consistency and transparency about who’s actually using and doing what might be helpful.

§ This sounds like a lot of work.

This is a good objection, but one that’s highly dependent on specific implementation details that I haven’t even begun to wrap my head around.

In one sense, it’s trivially zero percent more work. Just fork your curriculum, never touch a thing, and life goes on. Or make/modify all the stuff you want, never bother to propose or accept changes for or from anyone else, and life goes on. You don’t have to play the larger game that’s made possible by the tool in order to use other good features of the tool. GitHub admits a lot of different styles and workflows, and I think part of its success is due to its robustness across a range of use cases. Something like that, geared specifically toward curriculum, seems to me worth exploring.

But if the participatory nature of this thing intrigues you, the costs3 of implementation matter a great deal. I suspect there’s a way to spread those costs over the system (for instance curriculum folks at the school/district level likely need to absorb a significant portion) that doesn’t make it untenable for individual teachers. If that turns out to be false, then this was a fun thought experiment and we should go think about something else.

Speaking of which…

§ Haha, I used to think that, too.

The argument I found most compelling came from Mike Caulfield on Twitter, who has actually implemented something like this before, and then significantly walked it back. I’ll let him speak for himself.

§ Final Thoughts

I hope that cleared some things up. If nothing else, it helped me hone my own thinking a bit, which is always good. I said in the original post that I think this suggestion is wrong in some specific and important ways, but probably right in general. I still think that’s true, but I don’t want to downplay the fact that the specific ways it’s wrong might be so great that the general rightness is irrelevant.

Here’s something I’m more confident in: we’re going to have to figure out our relationship to publishers and each other as high-quality, fully formed open curricula become available. There are some very smart people hard at work on this problem.

It’s coming. What do we want to do with it?

  1. This is the kind of sentence I was trying to avoid in the original piece.

  2. It’s not the only game in town, by the way.

  3. …interpreted in the broadest possible way.