technology
The "Build a Computer" class has begun to shift gears with the new semester. We finished building the computer last semester and are now trying to actually use it. In the first class meeting in January, we tried to use installers I had made on the mac in my office -- they didn't work. In the intervening week, I built new installers using a linux box and found that there was a step that the Mac directions don't include, where you install something in the bookblocks of the device. That was consistent with the failure mode we were seeing, where the device was visible in the BIOS but did not appear bootable. Unfortunately, there was a holiday and then the next class period was on the first day of the new semester and I felt like I needed to skip North Star to be in my office for people with first-day problems. But I sent the installers with Daniel so that he and the others could try them out. They called me a couple of times with questions, but succeeded in finishing the installation and named their new server "polaris".
This week, we built accounts for everyone on polaris. We updated the groups file and then tested to make sure that everyone could log in and could become root. Then we tried to set up networking.
My general plan has been that eventually the server could become the router and firewall, plugged directly into the cable modem. We're continuing with that possibility in mind. For the meantime, however, the server will reside behind the wireless basestation and so we wanted to set up wireless networking. It took some reading of man pages and a bit of googling, but eventually we got everything in /etc/network/interfaces, stopped and started networking, and we were able to ping out. We ran apt-get update and then were able to install tinymux.
My plan is to have them set up tinymux as their first service: something fun. And then move to other services. I'm looking forward to next week.
Eight years ago, I set up a text-based adventure game with my older son called "Muppyville". He was in fourth grade. We showed it to all the kids in his class for one class session and then he and a handful of his friends played it for months.
There was a little boy who was fascinated by the potential of Muppyville to exert power over the other children. He created places in the game that he could lead unwary people into where they would be trapped and he could taunt them and then leave them stuck there. He eventually figured out that he could change his "name" to "1imako", which looked a lot like the username name I used "limako" (the first character is a a one instead of an "L"). And then he would pretend to be me to other kids but then insult them and use obscenities -- among other things. I gave him a couple of chances to straighten up, but eventually I had to block his access and tell him he would have to have his parents contact me if he wanted access again. He was understandably reluctant to do that.
Fast forward to 2011: According to my son, this boy got an administrative password to "Powerschool" -- the database that the school uses to maintain records -- and subtly adjusted a few of his own test grades to change his grade from a B+ to an A. And he got caught! Reportedly, he was suspended and all of the colleges and universities he had applied to were notified of his academic dishonesty.
Some people just never learn...
Last spring, I spent a lot of time thinking about what kind of contribution I might make to the teaching at North Star. I suggested several possibilities and they chose to invite me to offer a course on building a computer. We met for an hour a week starting in September.
The first class attracted a lot of initial interest: more than 20 students came. I led a discussion where we started with what kinds of computers are there and it was great! Students started listing brands of computers (Apple, Dell, HP, etc) and, then we moved to form factors (laptops, desktops, towers, etc), and then purposes (workstations, servers, thin-clients, etc), and eventually came to architecture (RISC, CISC, and different processor families). We established our mission as trying to define the components, purchase, and assemble the components to build a server computer.
The next class, fewer students came, and even fewer came after that . In the end, there were only three who came every week. It may be that the others were imagining I would come in with boxes of parts and have everyone build their own. It was more work to try to figure out everything and buy it first. But we persevered.
We discussed the kinds of services we thought we might want. We agreed that it should be a file server and probably have a webserver. There we a lot of interest in having something that might be able to improve wireless performance in the building. And I suggested that we could also run a Minecraft server. And Daniel very much would like to run a MUX or MOO server as well. We talked about the other ubiquitous services like DHCP and DNS, that make the internet work.
We looked briefly at operating systems. We looked at and rejected Windows (which, if you try to buy the server version, is fantastically expensive). We considered Open Indiana and the various flavors of BSD, but settled on Ubuntu 10.04 LTS as probably the way to go.
By now, the fall was disappearing quickly. We identified some sources for components and began looking at the cheaper components, looking for something that would be most appropriate for a server (making sure there were drivers for linux) without breaking my wallet. We selected a processor/motherboard combo and picked out a DIMM. The first case we wanted was too expensive and didn't include a power supply, so we went with a cheaper case. When we first looked at hard-drives, they were quite cheap, but due to the flooding in Thailand, prices tripled, so we ended up getting a smaller hard-drive than we'd initially hoped. It took an extra day to find a wireless card that looked like it would support the HostAP stuff that (hopefully) will make it easy to build a captive portal.
Today, in class, we started building. George had teased me good-naturedly regarding whether I'd be able to let the students actually do the work, but I think my photographs provide testimony that I let them actually do almost everything. We got the motherboard installed and installed the CPU, CPU fan, and DIMM. Next week, we'll hopefully install the hard-drive and wireless card and hook up all the internal cables. Then we should be good to try to boot it up and install gnu/linux.
Next semester, we'll continue with sysadmin training and actually setting up all the services.
I got an email from one of technical staff in another department wanting to consult with me about setting up a streaming server. When we met, he explained that the faculty in his department were doing "class capture" and wanted to set up a server to provide access to the video files. We mostly talked about the technical issues involved, although I couldn't resist at least mentioning the pedagogical underpinnings.
So-called "Class Capture" is stupid. If you're teaching in such a way that a video recording of the screen is useful, then you're doing it wrong. The time the students and faculty are together is incredibly special and can be used much more effectively than as a memory dump by one person. In particular, you can have students discuss problems in small groups and report out to the larger group: class capture doesn't work for that. You can have students work on group projects. You can have students actually do things and not just sit there. If you do anything interesting with the class time, class capture doesn't work: it would be pointless. To do class capture in an environment like that, you'd need a team of camera-persons and sound persons. And an editor to provide a comprehensible stream of footage. Class capture only makes "sense" if you have someone lecturing. Sigh...
I pointed out that I understood that the poor tech couldn't make faculty choose different pedagogy and that he needed to just make something that would work for them. I showed him an approach that I thought would work. I pointed out that you really only need a "streaming server" if you want to provide live feeds and that for posting files, you could probably get away with just posting video files (probably .flv, transcoded using ffmpeg) and a player like flowplayer. I pointed out that you could put a pretty front-end on it with Drupal, but he said that people just needed to embed their videos in their course websites, so that probably wouldn't be necessary.
We also talked a bit about hardware and OS: he suggested getting a tiny Dell computer and using Windows. I pointed out that you probably wanted more reliable hardware than that, but he said that his department was notoriously cheap. It turned out the only reason he was talking to me was that they'd looked at purchasing some kind of video streaming solution that cost $5000. If money hadn't been an issue, they'd have just bought it, I guess. Sigh...
When I heard about ContactCon, I signed up almost immediately. The issues being raised have been of interest to me since I started using the Internet: how to make sure the net can be used for empowerment rather than oppression. The net is clearly useful for both, but the trend has been shifting in the wrong direction for years.
A corporation would have never made something like the Internet in the first place. When I was a kid, we still had The old AT&T and Bell Telephone network. You weren't allowed to own a telephone: you were required to rent one from the phone company. And everything was monetized. Now, I suspect that the most expensive thing about current cell-phone operations is the overhead necessary for administration, metering, and billing. And that's the direction we've been going: give users a dumbed-down box that only enables what the monetizers want you to be able to do.
There were a lot of interesting people at ContactCon most of whom I'd never met before. The demographic was mostly white, largely male, and somewhat younger than me. There were some folks my age or older, but we were the exception. Many were young entrepreneurs and freelancers looking to network to support their project. It reminded me of the luxury of my current circumstances: I have a steady job and don't need to spend half my time trying to market myself or bill people. I don't have to work on spec or limit what I do to what people are willing to pay for. I get to spend most of my time actually just working and being creative. I lament for this generation that is so circumscribed and limited in their choices -- and will probably end up permanently stunted by the economic conditions that have been imposed on us by the 1%. Or, if you prefer, that through my generation's lack of engagement, we have allowed ourselves to be disempowered.
I wore my "Official Red Hat" red hat and took my ubuntu netbook to demonstrate my free software street cred. I actually met the guy who'd ordered the stock of red hats when he worked at Red Hat in that time period. I had completely borked my install of Ubuntu a few days ago (or maybe the update from Easy Peasy never really worked right). In the event, I completely wiped the netbook the night before and re-installed everything. I've started using Dropbox to maintain the rough drafts of my writing, so it was easy to get my data back. I could have just taken my macbook, but it wouldn't have been as fun. In point of fact, I hardly used it, but it was nice to know it was there.
I met dozens of people, learned about many new projects, and also touched base with projects I've known about but haven't had time to explore. I've been interested in the Freedom Box since I first heard about it: it's consistent with my vision for people having their own server. And it's also the only way to have any assurance of privacy: you can't trust third parties not to reveal all of your private information to the government or corporations.
I organized a discussion about education and unschooling. It was a very receptive audience to the ideas and there were a number of people working on interesting things. The most interesting was probably Be You, but there were many, many others. ContactCon reminded me of what John Jungck used to say about the goals of BioQUEST: to begin conversations worth continuing. I suspect I will continue to interact with some of these people going forward.
A system for badges for life-long learning, based on previously-established Department Learning Goals, driven by student applications, and implemented in Drupal, could encourage students to align their own learning with the Goals. In addition, it could provide a much-needed impetus for building improved assessment of learning.
The Biology Department at the University of Massachusetts at Amherst has fostered revolutionary pedagogy and pioneered team-based learning approaches in lab and in lecture. With funding from Howard Hughes, we built innovative introductory and upper-level biology laboratory courses. With support from the Pew-funded Program in Course Redesign, we developed on-line resources to support in-class small-group problem solving. With funding from NSF, we've developed cutting-edge model-based problem solving activities for use with clickers.
About ten years ago, the Biology Department adopted a forward-looking set of Learning Goals that has been influential in the redesign of existing classes and the design of new ones. Acknowledging that, while the curriculum defines the depth and breadth of disciplinary knowledge, the Learning Goals can establish a vision for the skills and perspectives that every Biology course should foster and develop in students. The Learning Goals in many ways encapsulate scientific literacy and an appreciation for life-long learning from a biological perspective.
While the Learning Goals have been influential, both in terms of curriculum and course design, the Department has not established assessments to monitor and evaluate student progress toward the goals in any centrally-reportable way. That's not to say that the assessments don't happen at all: they're simply embedded in courses with no mechanism for sharing them or their results. Finding ways to share the outcomes of assessments would help the department substantially, both by providing information about the Department's mission to help guide further efforts, but also to enable the department to communicate our mission to outside constituencies. It has been difficult, however, to build the necessary consensus among the faculty to accomplish the work. I believe a system based on badges could solve this dilemma.
I propose a slightly subversive mechanism to turn the problem on its head by enabling students to request badges through submission of an artifact from a course or experience (a project, paper, photograph, video, examination, etc) and writing a brief statement that explains why it merits receipt of the badge. By creating a student-driven system, the incentives for faculty could be inverted: the student requests will drive the system and provide the information about where the learning goals are already being assessed.
Badges will be aligned with the major learning goals and perspectives with four levels that students could potentially achieve for each badge (one for each year of study). Ideally, the instructor of the course and another faculty member would be required to certify that the work submitted represented progress above and beyond the level the student had previously achieved. In very large courses (some Department courses have more than a thousand students), this could be unrealistic, however, and course TAs might be needed to manage the requests.
By encouraging students to achieve the Leaning Goals directly, they will develop increased familiarity with the Goals and begin looking at their assignments and activities with an eye for how they could potentially meet Goal and earn badges. This will undoubtedly give students greater appreciation for how the Learning Goals articulate with class goals and with the Goals the faculty have for students in the major.
There are several places where the student badges could be used to confer privileges and opportunities to students. The website itself will offer a feed where new student badges are posted (at the discretion of the student). Another place where the badges could be leveraged is the Biology Undergraduate Research Apprenticeships application program where students already apply online and their badges could appear when their application is viewed. Changes to the Honors College will also necessitate the Department managing an admissions process for Honors students and the badges could appear when this determination is made.
Minimally, this system would provide a wealth of information to the Department regarding where the Learning Goals are being practiced and implemented effectively. One potential outcome of this system would be that faculty might begin redesigning their courses and assessments to facilitate student badge applications: faculty might feel some pressure to redesign their assessments if their classes yield few or no applications for badges.
I believe this can all be built with tools already available and in use in the Biology Department: We have extensive experience using Drupal and we've already started using the user_badges Drupal module (for Science Scouts badges). This module does not currently support the Mozilla Open Badges initiative, but I have already begun trying to organize a group to put in a subsequent proposal to extend the user_badges module to support it and it seems like a good platform to start with. Students would submit the material on-line and, via the workflow module, the faculty members would be notified to evaluate the submission, and assign the badge.
If other departments or entities on campus were interested in pursuing similar badges, it would be easy to replicate or centralize the infrastructure for the badge system. The General Education group and the new Integrative Experience program have similar kinds of learning goals that badges could be developed for.
For nearly ten years, I've been trying to convince the Biology Department to embark on a process to assess the Learning Goals we developed and adopted without success. If a system of badges could get students to start identifying where we're already doing these assessments and provide the mechanism for sharing the information centrally, it would be a transformative step in moving the department forward. But the biggest winners would be students taking greater control of aligning their education with the Department Goals.
Submitted as a proposal to the Digital Media Learning Badges for Life-long Learning Competition with Tom Hoogendyk & Coherent Bytes listed as a collaborator.
It's fascinating to watch the varied reactions regarding the revolutions sweeping through the middle east. The people are excited. The real-politik folks are worried. The doomers are eschatological -- But then they always are.
A lot of people have been commenting on the role social media have played on the revolutions. Bin Laden began Al Queda with the goal of toppling the dictatorships that had been supported by the West and his goal is now realized -- but for the perfectly opposite reason. Part of it is purely economic, but a big part has been that the new media have enabled the people to see what liberal democracy looks like and they want it. They won't want a fully Western society, but I think they do want the trappings of a modern civilization: political and social freedom, an open media, and the ability to be part of the modern world. But what they want and what they'll get may be something different.
Nevertheless, I'm hopeful. The real-world potential for technology to let people communicate and organize may yet carry the day. And we might have new allies to help America throw off the yoke of neoliberal imperialism...
I saw a funny report this morning in about students worrying about becoming one of those iPhone people. I don't worry about that, but my iPod Touch has been helping me understand it. I probably spend too much time looking at computer screens already but, with the iPhone, you can be looking at a computer screen without being at a computer. It's terribly easy to keep up with mail, look stuff up, and keep current with social media -- you don't have to be at a computer anymore.
Speaking of social media, while I reading Inside Higher Ed, I saw this cautionary tale about faculty being surprised when their posts were more public than they had expected. I've been trying to tell everyone who would listen for years that you shouldn't expect privacy if you post stuff electronically. I forwarded the link to the Biology faculty with my admonition: I think the take-home lesson should "you should act as though everything you post online can be read by anybody". I think the wrong message is to tell people to carefully configure their privacy settings to make sure their posts remain private.
One other bright spot recently was reading that Diane Ravitch, one of the architects of the failed so-called "standards based educational reform effort", has woken up to how much damage we're doing to kids with those policies. Better late than never. I've tried to tell people: if you set policy based only on what's easy to measure, you leave out most of what's really important in education.
Yesterday, I participated in another two sessions related to developing a master plan for the library. The first was about computing spaces. On their first slide, they had labeled it "OIT Spaces", however, which drew a fairly strong response from the campus CIO and Director of OIT. "I thought this was about the library -- Nobody told me that this was about OIT space." The consultants were abashed and apologized for not using the correct term, but I thought it was a good example of how the upper administration builds processes that are blind to what actually happens on the campus. The announcements of these sessions made it sound like they were "teaching tip" workshops, rather than having anything to do with developing a master plan for campus space. But I digress.
I thought the session had too much "reading of the slides" by the consultants. We can read the slides -- really. You really don't need to read that long list of buzz words.
I said several things that probably made some people angry. In particular, I talked about the fact that the networking business model left 2/3 of the network in the ISB dark. The NSM networking group had tried to get control of the wired network so we could build out gigabit ethernet and make all the jacks live, but we lost that battle and got only 300 jacks turned on. This has resulted in a nightmare where most of the network jacks don't work and the spaces are not reconfigurable. You want to move the printer over there? Sorry -- that's the only place with a network jack. The jacks there, there and there? They're turned off. Their solution is that everyone should use wireless, but that doesn't work for the huge datasets that we have students building. There are jacks in all of the student study areas, but they're all turned off. And even in the labs, half of the jacks are turned off, so it's always a guessing game: does this one work? No. Does this one work? No. And, when the issue was raised the OIT response was that the building was "overwired".
I also gave my speech about what the campus would look like if it were built like "Spark" -- the webct/blackboard CMS. If the campus were like Spark, it would only have one door and you'd have to swipe to get in. When you got inside, you would only be able to see the doors that you're allowed to walk through. Everything else happening at the University would be invisible. MIT has their "open resource" policy: all of the courses at MIT post all of their resources openly. Why does a private university do that and our public university locks everything behind a door. And the campus itself doesn't do that. Did anybody check IDs of people at that meeting? Of course not -- even though I found out that at least one person had crashed the meeting without RSVPing.
One guy who was there was a professor of Computer Science. I wonder if I rubbed him as wrong as he rubbed me. He said that Computer Science didn't maintain its own computer labs anymore because students all just used their own laptops. I pointed out how we monitored our students and were aware that not enough of them had laptops to enable us to do that, but that we also wanted to be sure we could present a consistent and functional environment because the students' computers all had different versions of software where the software was also often broken or misconfigured. "Yeah," he said. "I've spent lab periods where I spent all my time trying to help students who had different versions of things who couldn't find this menu or open a kind of file." I pointed out that one solution was to use open-source software because then every student could install the same version. His eyes turned into big red X's: "Well, I'm a Computer Scientist and I can tell you that you should never trust that Open Source Software -- it's full of bugs and is always broken." I think my mouth must have fallen open. What planet is this guy from?
I did meet the new OIT guy who support's Apple systems. We had a good conversation and then walked over to the ISB so I could show him around and talk about our support structure and issues. He said he'd invite me over to see their enterprise next week.
I also met the new technologist at the library and said I'd give her a tour of Morrill so I could show her some of the best and worst spaces on campus. I also told the DEGW people to visit the Intro Labs in Morrill. We built those a dozen years ago and they've been incredibly successful.
It did show me how insular and lucky we are in the Biology Department: we've had the freedom to build and support great stuff. But it means that we don't always see the culture in other departments, where people take issue with the very idea of "group work". We need a lot more transparency on campus to shine light into the dark corners that are full of dust and cobwebs.
On Thursday and Friday, the content management system selection committee saw two presentations, one from Hannon Hill about their Cascade Server product and the other from Lucidus, a small web development company in New Hampshire about Drupal. I felt that there was greater enthusiasm for Drupal and am cautiously optimistic that drupal will be ultimately chosen.
Cascade Server is a proprietary software product that is a "push server". Essentially, it is a content management system that pushes out static webpages for deployment. Content developers and managers connect to the system and see a file-system like view (in a web browser) with a number of tools to edit pages and workflows for content approval and vetting. The pages that are actually deployed don't have any directly interactive components. Things like RSS feeds and lists of headlines are all pre-generated on the system and posted (e.g. by cron) on the actual site as static pages. This means that if any real interactivity is needed -- ie, any kind of response to user input -- it needs to be accomplished by other kinds of one-off packages or php scripts, rather than being an integral part of the system itself. This seems like a serious shortcoming to me.
Given that we've been working with Drupal, I don't need to describe what Drupal is or how it works. The presenter offered a presentation about Drupal and then showed how it could be used to address the three scenarios that he had been tasked with. There was intense interest on the part of the selection committee and many questions -- many more than when the cascade server was presented.
The presenter for Drupal made a persuasive case that the system selected needs to be able to grow and adapt to the changing needs of the campus. He argued that commercial companies need to focus on the key features that people know they need now, whereas an open source system, by encouraging the participation of the user community, has a "long tail" of additional add-ons and modules that only a small fraction of users want now. Not all of these might be ready or needed immediately, but some of them undoubtedly represented features that the University was going to want in the coming years -- we just don't know it yet. In the end, that was the key difference between the two approaches.
Using the Cascade Server, would be like the Red Queen -- running as fast as we can to stay in the same place. It would allow us to continue building websites as static pages, but would not provide transformational change in the kinds of services nor prepare the campus for the future. Drupal was hands-down the winner if the goal is to revolutionize the kinds of services available to the campus community.
I believe the process still has a couple of steps, including opportunities for the selection committee to actually use the products. The Drupal presenter encouraged the committee to contact a webmaster who had overseen the implementation of the Cascade Server at Plymouth State University (I think), but who was now someplace else in New Hampshire. One got the impression that she would not provide a favorable perspective of the Cascade Server.
The first week of the semester has come to a full stop in the middle with a snow day today. It's fun to watch the snow coming down when you don't have to go out into it. I have many memories of having to go out into cold and snow. It's more fun to just watch through the window. If it changes over to freezing rain and we lose power later, that might put a damper on the fun. But the snow is pretty, even though I don't look forward to the shoveling to come later.
It's the first semester in years when I haven't had to teach. I enjoy teaching and I will look forward to being back in the classroom next semester, but for a change its great to have time to focus on technology. I've been trying to learn the ins and outs of ldap for the BCRC and department computers we support. I spent a lot of the day yesterday trying to get ldap over ssl working and, by the end of the day, I think I have a working set up. Now I have to figure out how to make the MacOS part talk over ssl to the server. There are a lot of things I really like about MacOS, but -- from a unix standpoint -- its really frustrating when you find perfectly functional and transparent parts of the system replaced with an opaque black box. My ldap connection works fine in plain text and doesn't work over ssl -- even though the command-line ldap tools work fine on the same computer. Argh.
The Governor has released his budget projections. The University has taken a $27 million dollar cut from this year's budget -- this year, that's already half-over. Next year isn't going to be any picnic. Tomorrow there is a General Faculty Meeting where the chancellor seems likely to lay out his plan. I am hopeful that Obama's economic stimulus package will help the state and the University. The University has been cut so much over the past ten years that there really isn't much spare to cut.
I'm installing Windows for a teaching lab that wants to have Windows software available. It's been an interesting experience. Using Windows is a really unpleasant experience: it constantly pesters you and gets in the way of trying to get work done. Ugh.
Using VMWare for virtualization seems pretty cool. I was more inclined to go with Virtualbox, but others wanted VMWare and the ability to run VMWare Fusion, with MacOS and Windows windows interspersed does seem cool. Both have parts of their virtual machine open sourced and other parts closed, so it's hard to make a determination on that level. We were able to get an academic license to use either for free, so the price difference didn't matter.
One important feature of VMWare for interacting with radmind is that you can have sparse filesystems that are split up into files that are not larger than 2GB. This suggests that we won't have to copy more than a few GB of data per machine each night to restore the image of Windows available in the lab. I was worried it would be a disk image the size of the full drive.
At startup, VMWare would ask whether the image had been copied or moved -- I found a post that suggested putting this line: uuid.action = "keep" into the vmx file would prevent it from asking the question. We'll see. I hate not being able to make dialog boxes go away -- its one of the things I hate about Windows.
One important trick I found for running in Unity mode: turn off the desktop background. By selecting "none" and the color black, when you move windows or close windows, the screen updates much faster and without jarring redraws of the a grassy field.
The first time I tried a virtual machine copied over using radmind, it looked like most stuff worked fine. I did notice an error associated with tpconnect (which appears to be something related to printing, but I'm not sure). I haven't tracked that down.
It's been time consuming to get this far, but I'm hoping we'll have a working Windows installation in the lab when students use it tomorrow morning. We can work on refining the details over time, although it stinks to not be able to do file-by-file improvements -- 2GB a pop is better than nothing, but it still sucks when compared to the file-by-file control we get with radmind.
Recent comments
15 weeks 5 days ago
16 weeks 5 days ago
30 weeks 12 hours ago
37 weeks 3 days ago
41 weeks 3 hours ago
41 weeks 3 hours ago
46 weeks 4 days ago
47 weeks 5 days ago
48 weeks 2 days ago
48 weeks 3 days ago