Twitter is your IT support
This post isn’t intended as a criticism of anyone, rather an observation on a trend I’ve noticed from several others also.
I’m running my block of the Masters course H817 as a MOOC. It’ll start this March, and one of the things I wanted is a DS106 style blog aggregator. That is, I want the contributors to register their blog, and for posts they tag appropriately to repost automatically in the course blog.
Now, the sensible way to do this seems to be to install WordPress and use the FeedWordPress plug-in. For reasons I won’t go in to, I haven’t been able to get this done at the Open University, despite trying since last July. I’ve spoken to people at others unis and it isn’t isolated to the OU, it seems to be this low-level, experimental type of IT support is increasingly difficult to find.
Do you know who I think the culprit is? The VLE. As universities installed VLEs they became experts at developing enterprise level solutions. This is serious business and I have a lot of respect for people who do it. The level of support, planning and maintenance required for such systems is considerable. So we developed a whole host of processes to make sure it worked well. But along the way we lost the ability to support small scale IT requests that don’t require an enterprise level solution. In short, we know how to spend £500,000 but not how to spend £500.
With time pushing on, I got some local support in IET (thanks Will), and investigated the best options. It became clear that hosting externally would be the easiest option. I asked on Twitter who the best hosting solutions were and got a great set of responses. Alan Cann suggested siteground and sent me a friend’s introduction offer, so I signed up with them. In 10 minutes I’d followed their WordPress installation, installed WP, found the FeedWordPress plug-in and installed that too. The blog is here: h817open.net
My next problem was to get registration to work. I looked at what DS106 have done, and in response to a request on Twitter, Alan Levine pointed me to this solution from Martha Burtis. That was a bit beyond my coding skills. So I created a simple Google Form and embedded it in my blog. I then asked Martin Hawksey if he knew how to take a Google form output and turn it into an OPML file as FeedWordPress allows you to create a list of syndicated sites from an OPML file. He pointed me to this service, which does just that.
He and Tony also suggested some solutions which will help with autodiscovery (ie finding the feed for the blog if someone just enters their main URL). Martin lists some options here. I haven’t quite cracked this yet, so it needs people to enter their feed url, which isn’t ideal, so I’ll keep working on this.
So, I think I have a working solution. If you have time, could you try it for me by registering your blog here and then posting something which has the tag #h817open. Only posts with this tag should appear. You can just do a test post, but perhaps more interestingly you could respond to the idea of whether “openness has won” or not with a post “Openness has (hasn’t) won because…”
This took me about an afternoon. The problem is that it’s quite a flaky solution, the OPML service could disappear, I don’t really know if it will work for all blogs, and there’s no support. So it could go down and I’ve got no backup. More worryingly I am at the edge of my expertise here, so I can’t really do much more. So a university supported proper grown-up solution would be preferable. On the other hand, one of the things I frequently say is that the best thing we can teach our students is how to be good members of a network. And this is a clear demonstration of why I believe that. I think this tension between the innovative possibilities of new technology and the understandably cautious nature of institutional IT services will mean that we’ll all need to do more of this DIY tech set-up and experimentation. And the only way to do that is to have a good network who can help.
One last plea – I joked with Alan that I needed DS106 out of a box. I think I’m serious though – it would be great to have a step by step, idiots guide to installing and setting up a DS106-like environment. The rest of us don’t have Alan and Jim’s tech skills, so getting to the starting line is difficult. I know they’ll say you should invent your own way, but they done so much great work that I don’t think they realise just how much expertise they have. A simple installation that let the rest of us get started, would mean we could all go off in different directions then. So any of the DS106 crowd up for it? And I do mean a simple guide, it has to be Weller-proof.
Couldn’t agree more. We’ve talked about it at length many times, just a matter of resources to build it. I think ultimately it could be a plugin of its own. Install the plugin, put in any necessary details, and you’re up and running.
if I read between the lines, you seek a system that would include a way to have Feedwordpress and the signup form integrated? Having been mothering this for a year, and currently with the ETMOOC site, some is easy, and some is black art.
Mechanically it is not a challenge, expect that FWP stores all of the information in the wordpress Link structure, which is no longer being automatically present in WP 3.5 (i can be done via a plugin). It’s an arcane way to store the data; we asked the developer about doing mods, but enver heard back.
The device we use at ds106 requires the use of a paid WP plugin, Gravity Forms, to handle te data. Its likely possible to code this w/o that plugin, it just makes things really easy (and does email confirmation).
That said, getting the feeds is easy if people are submitting blogs solely devoted to your class, that was the reason for the class I set this up for with Nancy White, that we asked students make new tumblt blogs- an entire blog is easy to autodiscover the feeds.
It gets more convoluted when you let people use a more general blog, but use a tag or feed. Oh my it gets messy. WordPress category/tag feeds are easy to ID, but they are not the first ne auto discovered. Blogger labels (equivalent of tags/categories) are not autodiscoverable. They have to enter it manually.
For ETMOOC I tried to make this explicitly clear, a long doc which very people seem to read
I have evidence that 15% of educators do not read 😉
For class, it is feasible to verify the feeds, but you have to know something about the various ways they work.
So in the end, setting up a ds106 signup form is not the hard part, its dealing with the vagaries in the data.
I’m willing to take a look once I can get past the startup of two open classes, and we can see about that Weller test.
I half been nagging a few people to see if an idea like this has got legs – developers and academics dating. I’ve been half emailing nagging a few academics to get involved.
So far I’ve made – http://wordpress.org/extend/plugins/buddypress-activity-graphs/ – to try and visualise group activity – something very handy for a MOOC I think.
But on a broader scheme it’s about pairing academics and developers to make cool open source tools. Sounds like it has legs to me, but it’s early days.
Am genuinely keen to try this out though, so feel free to get in touch
@timmmmyboy – yes, the DS106 plug-in is what we need. You need to get some whizkid intern for the summer to pull it all together. Isn’t there an MSc in Computing or a PhD we can make this a task for?
@Alan – I’m trying to just do general blogs with the appropriate tag. At the moment I only have the one tag in my aggregator blog and have gone for the option in FWP that says it has to match a tag. Martin Hawksey has some autodiscovery, so that might work. The problem is I think FWP doesn’t automatically keep checking the OPML feed, I have to manually tell it to check it. Which is no good at all. So man, do I need that Weller-proof system.
@Pat – that sounds interesting, these kind of peer support tools will make MOOCs fly I think, as we move away from supported versions. Be keen to explore this.
I don’t see this as being tied to MOOCs, just tied to open. Happy to work on anything open, but then I am just one coder, and I am sure there are more. I’ve always been a lucky elearning developer in that I’ve always dodged the VLE sys admin bullet.
I see it, if we excuse the dating metaphor as basically academics saying “I’d like to take X and do Y with it). This could then be published on a website, and people chip in with answers / agree to work on stuff.
Hopefully this would solve the disconnect mentioned above, but mainly make more educational content available to people.
I’ve already half written a plugin for LJ (I have use for it too now) but happy to keep it going on elsewhere – not sure how best to create it?
Know next to nothing about coding but happy to try out the technology.. Do I need to open a blog to register to the blog above?
Always good to hear Martin Hawksley is still busy!
I’d be glad to help with the set up.
I’ve been nose deep this, having processed over 370 submitted URLs for the ETMOOC hub, and I can tell you the concept sounds reasonable at a conceptual level. In practice, with the variations in how different platforms publish feeds and combine with the human factor – I have seen URLs submitted as http// http//: http://// URLs that pint to single posts, URLs that point to pages, URLs that say “none”…
I’m ready to outline about maybe 30 different things you might have to test for.
Auto discovery sounds automatic, until you realize a few things. A WordPress URL can have 3 or 4 possible feeds (blog rss, blog ATOM, comments, category). Its one thing to run auto discovery on one URL and pick a feed (FWP has autodiscovery built into it). But its trickier to figure out in an automated script which one you use. Blogger URLs for “labels” (what it calls tags”) lack anything to autodiscover, the only you feed you get is for the whole blog, not what you want.
I am nto sure why you are fiddling with the OPML. You should should copy the URLs and paste them into FWP.
What I am doing for ETMOOC is adjusting URLs to be actually Feeds or ones I know can autodiscover the right feed as the first one. If I get a URL like http://bigtoenail.wordpress.com/category/beer if I leave it up to autodiscovery, the first URL is the root RSS feed not the category, but if I add/feed on the end, that is the one it picks.
So I cam cleaning my list of web sites to URLs I know FWP can manage. I then copy a list of these URLs, and let FWP multiple process them. I have let it rip on 50 URLs with no problem. We are up to over 350 feeds it is running now.
I’ve not tried the tag matchin approach, the way I have gone about it is to assume we want to syndicate everything coming in; all from a simple blog or a category/tag feed from someone who has marked their own content.
I’m sorry but its not quite “just a plugin” solution.
Thanks Alan – that sounds like a very hand-crafted solution. I was after something much more automated (coz I won’t be full time supporting the MOOC and will be away for some of it). I was looking at the OPML option because I thought it’d be a way of getting an automated feed of registered blogs, but it appears not. If I have to add them manually I’m worried about the lag between people registering and me adding them. It’s surprising how difficult it is to crack – Tony H should’ve solved this by now!
I’m not sure VLE alone is to blame for this malaise you describe. I have long been saying that most university IT service departments are less than the sum of their parts. Lots of clever people working there but with job descriptions that are antithetical to innovation. In my experience, I have been only able to get things done when I ingratiated myself in with the right people underneath the management level. But I can speak SysAdmin so this is not an option for most academics.
I think every university should have an innovations branch of the IT services that would run essentially like a little experimental ISP hosting WP, Drupal, etc. for academics to help with their projects.
Echoing Dominik’s thoughts, I don’t think it’s the VLE alone. In a paper last year I argued that the Product (the VLE), the People (the type of issue Dominik identified + others) and the Process (teleological/strategic etc processes) were at least three interacting factors at work creating the sort of problems you’ve experienced. The argument being that there are alternatives that might serve better.
Not sure how TypePad handles links in comments so will include the paper URL below as well.
I see you’ve tidied up most of the issues in the comments. It’s probably of little use given the way FWP works but for info here’s a way to generate OPML directly using a Google Form http://mashe.hawksey.info/2013/01/registering-blog-addresses-and-generating-a-opml-file-notes-on-feedwordpress-and-mooc-in-a-box/
Dominik & David – yes, you’re right in that it’s not the VLE alone, but (& I say this as a former VLE director) they are indicative of the enterprise solution culture that pervades higher ed now. The systems, cultures & processes are all geared towards this type of product & innovative, small scale, risky things don’t fit.
@martin – thanks, plan to take a look today, but the feedwordpress thing of not updating automatically scuppers my beautiful OPML plans. I really appreciate you taking time to look at it – this must be a nut we can crack!
Really interesting conversation Martin, and glad I came across this through Martin Hawksey’s recent post in answer to it.
From a developer point of view, and to answer the question about automation that cogdogblog raises, there’s no reason why you couldn’t create a submission form on a WordPress site which uses a plugin to validate feed entries. It could ensure that the URL is correct (none of the htttp:////) and actually fire off to read it to make sure that it’s also a working feed.
I suppose it’s the moderation that’s still tricky – that would always be a human factor – making sure that it’s not some pharmacy salesmen trying to inject sales messages into your MOOC…
Really interested to hear if you make any progress with the plugin so make sure to let us know 🙂
Wading into this a couple of months late. But hey, ain’t that what asynchrony was designed for…
The reason: I want to make a couple of points.
1. I think a list of exiting plugins as well as how they might be pedagogically appropriated would be better than an off the shelf one-size-fits-all DS mega-plugin. This would allow educators to mix it up how they like, experiment with new ideas, grow the tech as it evolves and generally keep the system diverse (cf. complexity theory).
2. If you do this it also puts the responsibility of the learning design back in the hands of the educator. If there’s always gonna be someone else who builds the tools, then educators are going to continue being deskilled in this department. And that’s a crap deal for learners in the long run.
@Pepsmccrea – yes that’s a good point. In my digital scholarship book I argued that academics had let publishers take charge of academic publishing, and had lost control of it, and they were i danger of doing the same with digital scholarship. MOOCs are a good example of this – if we let Coursera dictate what MOOCs should be we’ll soon find we don’t have much say in what they are.
Hey Martin – Matt Johnston and I had to build parallel back-bedroom (wordpress) versions of picbod and phonar initially – we still do, but bury the academic backstage gubbins in Moodle and use the front of house wordpress sites as adaptive aggregators for wherever the classes and their tasks outcomes. To suit the different approaches to open T and L any out-of-the-box would have to be adaptive.
There seems to be a groundswell of us wanting this open-course co-op moment (Martin Hawksey’s words) – Brian Lamb’s Wiki space brought me here, after writing a plea myself ( http://goo.gl/ZNE0u ), but I agree, Alan Levine and Jim Groom need pulling in and genetically splicing with Tony Hirst, yourself, Martin Hawksey, Kin Lane then possibly hosting by Audrey Watters. Or if Jim’s going to be precious about his genes we could just get together and build something openly.