TED Talk: Jennings v Watson

Deep Blue versus Kasparov was an pivotal moment for me. At the time I was playing lots of chess on the computer. (And usually losing.) So the prospect the best player in the world cannot beat a computer was a depressing prospect.  Maybe it is a sign of how much I had matured (or immatured) to think a computer could beat someone better than me at trivia. Maybe it has something to do with routinely using computers to compile things for me in minutes that would take me days.

If you have not seen anything about IBM’s Watson, then start with NOVA’s Smartest Machine on Earth. Even better, here is Day 1 of Jeopardy.

If the below video does not play, then try Ken Jennings: Watson, Jeopardy and me, the obsolete know-it-all.

TED Talk: We Are All Cyborgs

Technology is evolving us, says Amber Case, as we become a screen-staring, button-clicking new version of homo sapiens. We now rely on “external brains” (cell phones and computers) to communicate, remember, even live out secondary lives. But will these machines ultimately connect or conquer us? Case offers surprising insight into our cyborg selves.

Superhero Economy

Watched a number of episodes of Justice League Unlimited today. Buildings, roads, and machines get pulverized by the violent actions. Someone rebuilds all the destroyed stuff because in the next episode, everything is pristine to get pulverized again. So much rebuilding must suck for insurance premiums. However, it does ensure lots of construction employment, material sales, structural engineering services, and designers.

🙂

… And you thought they were just weekend morning cartoons!

Digital Legacy

A book on time management in talking about long-term goal planning suggests we define the legacy we wish to leave. Coming from academia, I typically think of a legacy as a name on a building, an applicant with an alum for a parent, or a scholarship. However, the artifacts left behind by previous cultures are also a legacy.

Our digital footprints both could be part of this legacy or easily lost. I lean toward all this data we spew about the Internet will be lost eventually. I have seen floppy disks and hard drives die, taking with them the only copy of critical data. I have seen companies report their hard drives stolen from their machines in co-location as why customers lost their data. I have seen companies close web sites because they ran out of money. Let’s not forget natural disasters like earthquakes and floods.

So we keep backups.

Who will preserve these backups once we are gone? Are you able to read the data from computers 40 years ago? Maybe we’ll be better about being able to read the data from past when we reach 40 years into the future?

Not likely.

Anthropomorphism

Do you talk about computers, software, or web sites as manifesting human-like behaviors? Personally, I have.

At work, we manage several machines who collectively provide the GeorgiaVIEW service. When a machine is completely unresponsive, then we refer to it as having died. When a machine stops a performance issue, we refer to it as happy. I call my car cranky when it fails to run well.

Besides, electronics and vehicles, are there other examples?

RE 2007: Administering Sakai

.

  • Timeline:
    • Oct 2006: Faculty committee selected Sakai over Blackboard Vista
    • Jan 2007: Developed a roll-out plan.
    • Jun 2007: Pilot
    • Aug 2007: Production
    • Still: Some classes still running on CE4.1, being phased out of use.
  • Needs – no more than 5% of code custom written by GA Tech or professional services.
    • Integration with Banner.
    • Grade book
    • Assignments – no resubmit. Professor had built a section aggregation tool which combined with assignments started killing the servers. Ate all the database connections. Dead in the water Monday afternoon to noon on Tuesday until a code change was implemented.
    • Assessments
    • Course lists would not show unpublished sections. Students were concerned registration failed to take place. Wrote a “More” link to show students the course exists but not yet available.
  • t-square
    • GA Tech – 30,000 users and 10,000 sections. 16,000 users login.
  • Sakai out of the box not very good.
    • Java Server on Tomcat.
    • User administration sux. No way to see what classs a student is taking without an outside application. Created an administrative console, available to machine room admins. Monitors services, processes. Admin console Perl connecting to APIs.
    • Admin role can access every course. Built in admin console to link directly to courses to go help troubleshoot.
  • Unicon – professional services
    • Built t-square implementation.
    • Created some tools.
  • Staffing
    • Systems – OS and hardware – 1 person
    • Application support – 1 person
    • Code developers, Quality Assurance – 3 people
    • Database Administrator – 0.5 people (spends part of time on other projects)
    • Instructional technologists – 2 people
  • Why Sakai? Faculty hate WebCT. Possibilities of integration (aren’t there possibilities for this with Blackboard?).
  • Costs – refused to say. Rumor is 2x our costs for making available for 200,000 students for Blackboard.

Kinda weird. Suggested it was nigh impossible to create accounts except through Banner in WebCT CE. Except the same APIs which create Banner accounts can create guest accounts?

On the Fourth through Sixth Loops of Ready 2 Wear

I really have to stop listening to the same song played over and over. It may affect my thinking….

We had another node crash due to the Sun JVM issue. Our start script failed to make a file in /var so the node did not become fully operational as expected. While waiting for those with permission to delete some stuff to free up space, I went looking for what I could delete myself. Naturally /var/tmp seemed a likely place. I found 1,171 files named Axis#####axis. (Replace the #s with well… numbers.) They used up only 42MB. Most were small. Looking across all our machines there are thousands of these dating back to February of this year.

I love the Unix file command. It will tell you what kind of files are there. So I used file | sort -k 2 to sort by the type. Almost all of the files were either plain text or JPEG or GIFs. One file, called a “c program file” turned out to be a JavaScript (based on the C syntax). I downloaded a JPEG file locally, renamed it to have the .jpg extension, and opened it in an image viewer. It opened correctly. Seems its a graphic of a table.

It would seem our Blackboard Vista 3 has been collecting these files for months. They do not take up very much space. There are not nearly enough files to represent a download of content by all users. Our /var would fill up hourly in that case.

Axis is an Apache SOAP project. Vista’s exposed APIs use Axis, I believe. So, the running hypothesis is several of our campuses are using a product which is contacting the APIs to upload content. Its spread out enough that all four clusters are affected. Its something that started about February.

Suspect #1 Respondus – Chosen because we know it hits the APIs to upload content. Discounted because the content is lecture materials. Respondus works with assessments (aka quizzes, tests, exams).

Suspect #2 Impatica – Chosen because the JavaScript file references PPT. Impatica compacts PowerPoint (aka PPT) files and allows them to play without needing a PPT player. Their support pages teach users how to use the Campus Edition 4 user interface to upload content into a course. O-kay….

Suspects #n Softchalk, Diploma, Microsoft .Learn, etc. – I haven’t really investigated any of these. They are just names to me at the moment.


UPDATE: So… There is a bug in Axis which dumps these files into the file system. The files can be deleted as long as they are not current.

links for 2007-08-17