Paul Holbrook's Radio Weblog : Worth $40 a year? You decide ..
Updated: 4/8/2003; 8:56:21 PM.

 

Subscribe to "Paul Holbrook's Radio Weblog" in Radio UserLand.

Click to see the XML version of this web page.

Click here to send an email to the editor of this weblog.

 
 

Friday, July 12, 2002
Georgia Tech's portal project, and how extreme programming and interaction design might help

A bit of introduction: the major project I'm working on at Georgia Tech is a campus portal. I'll try to blog more about the subject of portals at another time, but for now, portal software aims to give each member of the campus a personalized view of the resources of the campus: a student might see their class schedule and grade information, a faculty member might be able to update a class calendar, staff members might have a view of their own, alumni another .. you get the idea. But to the point: I've been thinking about two schools of thought that seem to be in conflict, but both of which might have applicability in Tech's campus portal project: Extreme Programming, and Alan Cooper's Interaction Design.

I've come into this portal project after a long evaluation process that involved both the central IT staff and people from what we've been referring to as the "functional side" of the campus. Tech went through a selection process, and eventually decided to work with one of the vendors who's big in this space. (I won't say who right now; I'm not sure the campus has officially made the announcement.)

I'm concerned that we run a big risk of not delivering something that the customers - those "functional" people - will find useful. I've heard it said that when the functional folks saw the demo from the portal vendor, they said "That's what we want." That's not surprising; most of us think that way. "I don't know what I want, but I'll know it when I see it." But it's a long way from a idealized demo by a vendor to being able to pull that off in a real-world environment.

The trick is to give customers the ability to give feedback along the way. A model that starts with talking to the customer to get requirements, and then goes off for a year to build what they said they want is headed for a train wreck.

Both Extreme Programming and Alan Cooper's Interaction Design get at this problem in different ways.

Extreme Programming (universally referred to as XP) is a methodology of programming which involves short cycle times working closely with the customer. In XP, the development team works in two week cycles, and works on implementing customer stories: stories being small enough pieces of functionality that can be implementing in a short space of time. The theme underlying XP is "embrace change:" since we know user requirements will change, we make a virtue of it. XP uses the metaphor of driving to illustrate this: you don't drive from Los Angeles to New York by pointing the car in the direction of New York and then closing your eyes for the next 40 hours: you have to make small corrections along the way.

Interaction Design relies on a tight loop between customer and interaction designer before coding begins. The interaction designer concentrates not on what the software will look like, but on what it will do. Interaction design concentrates on using this process to drive the requirements. This makes sense: you can't know the requirements until you've got a good idea of what you're trying to build.

Cooper's methodology also relies on personas: archetypal users that are used as stand-ins to help define the requirements. In the campus portal project, we might have a persona for a faculty member, another for a student, one for a research faculty, and so on. Interaction Design goes through the goals and needs of each persona to make sure the resulting design serves everyone.

Some resources follow-up on these topics:

Extreme Programming

The canonical book is Extreme Programming: Embrace Change

A Google search on "extreme programming" will yield boatloads of results, but here are a couple of good places to start:

Interaction Design

The reference to have is Alan Cooper's The Inmates Are Running the Asylum : Why High Tech Products Drive Us Crazy and How To Restore The Sanity. This is one of my favorite CS books out there. The first chapter of the book is available at Cooper's web site. This first chapter hooked me right away with this riddle/koan for the computer age:

Q: What do you get when you cross a computer with a camera?
A: a computer!

Fitting the two together

Finally, there's a couple of other resources that look at the interaction between XP and Cooper's thinking. Fawcette Technical Publications has a splendid joint interview between Kent Beck, the author of the first XP book, and Alan Cooper.. Jon Udell also has a very nice essay on the same topic.


11:36:35 PM      comment []
Using Cytrix to support machines in lecture halls

The Education Technology group here at Tech has to support over a hundred machines in classrooms, and keep them in sync with common software. Using a Citrix terminal server might make it easier to support one-off applications rather than having to install them on every machine.

I attended a "Talk with with the CTO" lunch a few weeks back with other Tech IT staff. There was some conversation about what they have to do to support the interactive classrooms tech has been building out over the past few years.

The comment was made that Professors sometimes want to install software on the shared computers in the classrooms. That's a problem, because the classroom computers are all supposed to be set up the same way, so a professor can count on finding a common setup even if they have to move classrooms. The professor has two choices: bring in a laptop and plug it in, or convince the central IT people that the program should be installed on all the classroom machines.

I have another idea, based on what CNN.com did for rolling out their common content management tools: have a Citrix terminal server installed centrally, and have the ability to log into that server from each classroom machine. The Citrix server doesn't replace installing common local software on all machines, but rather gives the professor an alternative: if they have a program they think would be useful, the IT folks could install it first on the Citrix machine, where it would then be available from any classroom. If the program proved useful, it could be installed more widely, or perhaps just left on the Citrix server until the Prof didn't need it. That would allow the IT people to be more responsive to their clients with less overhead.

Citrix isn't the answer for everything, but with high speed local networks and a reasonable load, I'll be it would be pretty reasonable for more applications.


11:01:07 PM      comment []
Items coming in from my Tech weblog

For a variety of reason, I haven't pointed to my Tech weblog from here; it's really meant for a Georgia Tech audience. But there's some good stuff over there, so tonight I'm pulling over items from that blog and publishing them here. One of the advantages of blogging for a smaller audience that isn't part of the larger blog community is that you can bring up stuff that's been blogged a hundred times in the larger community, and you get to sound like the expert, even though you're only the end node in a long chain. Hey, weblogs can make all of us look like experts!


10:55:50 PM      comment []
Universities work towards Superarchives

The Chronicle of Higher Education describes the rise of superarchives:

Several colleges are now looking to share more of that work by building "institutional repositories" online and inviting their professors to upload copies of their research papers, data sets, and other work. The idea is to gather as much of the intellectual output of an institution as possible in an easy-to-search online collection. One college has called its proposed repository a "super digital archive."

I'm skeptical. People do what they are motivated to do. Researchers motivation is much the same as everyone else: money. You do research, you publish, and if you do well you attract research dollars and build a kingdom. I don't see a superarchive fitting into that model. Oh, I'm sure some researchers will play, but I can't see them motivating everyone to contribute.


10:46:41 PM      comment []
Federal Grants and computer security

One of the security people at Georgia Tech mentioned that the Federal Government was considering writing some words about computer security into the rules governing Federal grants. I've now found some concrete information about that.. Here's a powerpoint on the topic by Bob Mahoney of MIT from 1/2002; he also points to two documents with the actual text of the proposed changes: the Safe Computing Environment Requirements, and an appendix to those requirements.

Here's a summary of the high points taken from Bob's presentation:

  • Grantees are required to certify that they will provide a "safe computing environment".
  • SCE refers to both initial config and ongoing maintenance of covered systems.
  • Changes to SC environment must be reported.
  • Requires statements of appropriate use to be given to all employees.

  • Requires prompt notifications of "significant events".
  • Specifies actions to be taken for security events

10:39:09 PM      comment []
In search of a cool wireless port on a Friday afternoon

One of the advantages of working at Georgia Tech is that they have wireless access at various places on campus. It's Friday afternoon, and for once it's actually bearable outside. (Atlanta has three good seasons, but this ain't one of them.) My building is a few blocks off campus, so for a change of pace I pack up my laptop and walk over towards the student center. There's a little coffee place called CyberBuzz there with that has a wireless hub, and Macs for those without a machine of their own. That was the plan, anyway.

Problem #1: CyberBuzz is closed. I guess students don't need coffee at 3:45pm. (At least not in the summer.) That's ok, because CyberBuzz has a little patio outside with tables and chairs, and I've been thinking of sitting there all along.

Problem #2: The chairs are locked up. The tables are still out, but without a chair, they don't do me much good. There are a couple of benches near by, but using a laptop isn't a whole lot of fun without reasonable back support. I spy a metal folding chair sitting next to a table, so I decide to sit down.

Problem #3: the chair has a plastic cover over a foam cushion. It's been raining. The back of my pants are now soaked.

I walk into one of the student services buildings. There are armchairs, and they don't look wet. I sit down in one.

The wireless connection is strong. My pants are drying. And it's Friday afternoon. I'm ok.


4:04:39 PM      comment []
CMSes and blogging: it's Xerox all over again

James Robertson of Column 2 pointed to the John Hiler article on why blogs are a disruptive influence.  I've already blogged the Hiler article, but Robertson's comment that Hiler's article will "make CMS vendors have second thoughts about their business models."  Don't be so sure ...

Robert's comment reminds me of when I worked at Xerox back in the 80's.  I worked on the Xerox Star, which was one of Xerox's attempts to make money on the brilliant work of Xerox PARC.

When the Apple Macintosh came out in '84, we bought one to see whether Apple had "got it."  That Mac sat outside the office of our software development manager.  That first mac didn't have a hard drive. Our reaction was that it wais a toy; we had nothing to worry about.

That's ironic, because around that time, Xerox was getting it's lunch eaten by Japanese copier companies taking over the low end of the copier market.  And that's exactly what the Mac did.  Xerox sold the Star for $15,000 (or more like $100,000 when you added in all the a couple of network servers).  Apple went with what they could get out the door for $2500.  It wasn't much, but what they were able to do for that $2500 kept getting better and better.  Apple defined the market for graphical computers.  Xerox sold off Xerox PARC this year.


12:03:25 AM      comment []

© Copyright 2003 Paul Holbrook.



Click here to visit the Radio UserLand website.

 


July 2002
Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30 31      
Jun   Aug


 4/7/03
 1/29/03
 1/26/03
 1/23/03
 1/23/03
 1/23/03
 1/21/03
 1/21/03
 1/16/03
 1/16/03
 1/15/03
 1/14/03
 1/14/03
 1/14/03
 1/5/03
 1/5/03
 1/3/03
 1/2/03
 1/2/03
 1/1/03
 12/30/02
 12/28/02
 12/26/02
 12/24/02
 10/8/02
 10/8/02

Home Page