This site is rarely updated. benbrophy.com is more up-to-date. - Ben

CocoaMySQL

Carl Jones pointed me to a nice OS X front-end for MySQL called CocoaMySQL ( download | FAQ | more ). I downloaded the official MySQL OS X installer and was in business in 5 minutes. Looks great, though I don't know enough yet to do much with it.

Also I came up with an idea for my class project that will revolutionize the way we do website maintenance. More on that later.

Tags:

Comments | 2005-09-29

Sakai hacking

Man check out that crazy Sakai hacker Steve Githens, busting out with a lab notebook app for his class, popped wight into his Sakai installation. Go, Steve, go.

Tags:

Comments | 2005-09-22

The Prototype Middle Path

I saw a great article on Balancing fidelity in prototyping. It is easy to over do prototypes, leading to lots of wasted effort. I've been guilty of under-developing prototypes as well, because the ideas seem so clear in my head.

Here's how I think I'm doing on the Stellar Images prototype.

Don't get carried away in making the prototype look pretty

I haven't done any more that the minimal styling, since I assume the style of the tool will come from the course management system it's working with. Many developers would do less, many designers would do more, but I've aimed for the middle path.

Keep interactivity at a medium to high level

I think I've done that, I added in JavaScript and made all the links work. Hopefully the HTML and JavaScript and some of the CSS from the prototype will pop right into the JSP pages that present the working application.

Don't compromise on breadth

All pages covered (luckily there aren't too many). I could probably do more coverage of error messages and such though.

Compromise as much as you can on depth

Done. This is where I could really lose it though, by adding various search results pages, and long list of images. If we were doing more usability testing at this point, I'd feel a lot of pressure to go overboard on depth.

Overall my use of HTML prototypes is too much for some applications, but in this case starting with HTML helps move on development and design at the same time, and will save us QA later.

Tags:

Comments | 2005-09-22

DSpace vs. Fedora

The question was raised on the Sakai-Library discussion list: Does anyone have any opinions on dSpace vs. Fedora?

There were many insightful answers, but this was the most concise, from Hussein Suleman

dspace provides a simple installable package with a user interface, some workflow management, its own repository, elements to support preservation, etc.

fedora is all about the repository and management of items in it.

technically, one could modify dspace to use fedora rather than its own repository ... so it may not be a choice but an "and".

you would choose fedora if you wanted to integrate a repository into a larger application. you would choose dspace if you wanted a one-stop-shop digital library system.

I love a nice simple answer.

Tags:

Comments | 2005-09-21

Grampa's tales of yesteryear

I've been participating in a class at Earlham College in Indiana - my Alma Mater. The class is called Social Impact of Computer Technology. The instructor, Mark Pearson, has set up the class so that every student is required to blog on their topic, and he's invited a panel of alumni working in the field to follow along and comment on the blog. I'm interested on how blogs can be used in teaching and learning, so I signed up.

This blog entry that sounds like something I would have written as an undergrad drunk on post-modern social theory: "Whats a Banana?". Here's my response.

"I know a lot of people work in the video business. People pay them to make videos about events, or about their company or whatever. When digital video cameras became affordable, and products like Adobe Premiere and Final Cut Pro made it easy for people to edit video without a lot of equipment or training, people in the video business got nervous. Would everyone just make their own videos now?

"Nope - they have more business than ever. Thing is that they anyone can make a video, but it takes a lot of effort and skill to make a good video. Same thing for web designers. Now anyone can figure out how to create a web page - but to make a really useful website takes skill and effort.

"Just something to think about. The increased availability of tools is a great help - and will open a lot of doors for people. But it's talent and the hard work that create something valuable."

My god I'm so OLD. That's the worst thing about this project - I feel so old an grizzled. I start saying things like "it's talent and the hard work that create something valuable." I should have signed off "Grampa."

Tags:

Comments | 2005-09-21

A wee JavaScript

I added a JavaScript function so that when the 'remove' checkbox is selected the effected row changes it's visual style on the edit class images mockup. It's not a big thing, but it will make it easier for people to see what they are doing (and it won't detract from the experience of those who can't see).

I've ordered a new JavaScript book - DHTML Utopia: Modern Web Design Using JavaScript & DOM - which looks great for the kind of coding I aspire to write, with the JavaScript truly removed from the XHTML.

Tags:

Comments | 2005-09-20

Untitled Talk about Cooperative Technology

Here's the abstract for Carl Jones' and my talk at the Small Tools/Big Ideas conference.

DSpace is designed for the long-term storage and management of digital assets. Stellar is designed for managing and presenting teaching materials while a course is being taught. Stellar Images will use DSpace assets to present images as course materials, playing off of the strengths of each tool.

This approach encourages innovation, reuse, and cooperation. DSpace will be one of many repositories feeding Stellar Images, and Stellar Images will be one of many windows into DSpace content.

We will talk about our plans to connect the two concerns, and the practical requirements of the faculty and students using these tools. We will also touch on the role of OKI's Repository OSID ("Open Services Interface Definition") as an interface between these tools.

What we need now is a title.

UPDATE: "Cooperative technology: Making institutional repositories and course management systems work together"

Tags:

Comments | 2005-09-19

Listening to users

Here's a must read piece: "Listening to users considered harmful?" on Creating Passionate Users. The gist of it is that listening to users will help you make incremental improvements, but won't lead to revolutionary improvements. It's something I've been grappling with when we discuss setting priorities for Stellar development.

Other than the work we're doing right now, I believe the best major improvement we could make to Stellar is to restructure the materials, schedule and homework tools to be organized around class sessions. Instructors would come in, with slots to fill up for each time their class meets. My inspiration comes from looking at hundreds of syllabi, and looking at the way 100s of Stellar courses are organized now.

Nobody asks for this change, but I can see it buried in many suggestions and complaints, "It's hard to keep my materials organized" from an instructor or "It's hard to see what I need to do today" from a student. Yet my experience and instinct tells me this what's needed. I'm applying the general rule of listening for user goals when I hear feature requests, not listening for solutions.

By listening to users we'd add infinitely nested folders like in Sakai's resources tool, and an elaborate calendar tool like dotLRN (I love dotLRN's calendaring, btw). These could be improvements, but the problems would persist and ease of use would decline. It's not revolutionary, it just adding features to existing tools.

We're not in the position to make either of these kinds of changes to existing Stellar tools right now. Hopefully Sakai will give us that opportunity - but I'm afraid it's design goal of being generic enough to work for research teams as well as academic classes will make it difficult. If advocating for this kind of change is a hard sell at MIT, it becomes exponentially more difficult when there are so many other stake holders involved.

No solutions here, I'm just wondering how to keep a development team nimble enough to make revolutionary improvements on a project that is so big.

Tags:

Comments | 2005-09-16

Simmons Database Class

I've started a class in database design at Simmons. The class is taught by an information architect named Amy Warner. The first class was great - it will benefit my work. Predictably I plan to do a final project that's some sort of web database application, probably using PHP and MySQL.

I had been thinking of whipping together an app that would make it easy for my group to track billable hours - but Basecamp scooped me by adding their own time tracking feature today.

Tags:

Comments | 2005-09-14

Stellar use still growing

I just saw the the number of Stellar class websites has inched up above the number used last semester. There will surely be more requests trickling in of the next week or two, but as I write the number is 428, and we maxed out at 416 in the Spring. In case any one wonders, that's 2137 class websites since Fall 2001.

Stellar class websites by semester

Tags:

Comments | 2005-09-12

Image Footnotes

I brought my son to a kids event at MIT on Sunday. I talked with a grad student from BCS who I've met a couple of times. We got to talking shop, and I told him about Stellar Images. As all teaching staff ask, he wanted to know if we'll supply copyright information. In particular, he want to know how he could keep track of which of the pictures in his personal collection could be re-used. I didn't know how to do that once the picture was downloaded. He said he wished is said in little letters on the bottom of the image.

I think that's a great idea. I saw professors with collections of images on their hard drives, images they could never in use OCW without knowing their source. I added a note to each full-size image in the mockup (example) in a 15 pixel strip on the bottom. I included info about how the image may be used, along with a sample attribution. I didn't include the actual license info, because I hear instructors saying they don't want legalese. The stripe is almost black (not full black which could distract from the image). I used Jason Kottke's Silkscreen font to make type that was readable at a very small size.

Not sure if we'll be able to manipulate the images on the fly like this in the Fall proof-of-concept, but I hope this will wind up in the final product.

Tags:

Comments | 2005-09-12

Stop Wasting Time in Usability Tests

Jakob Nielsen's Alertbox entry this week, Time Budgets for Usability Sessions, is a worthwhile read. Basic idea is that we spend time in usability sessions asking focus-group or survey questions that would be better spent watching how people use the interface. This has certainly been the case in most MIT usability tests, which suffer from an exhausting list of subjective questions, answered with a rating of 1 to 7.

I do enjoy the post-test debriefings. We're testing in an academic environment, and the people testing are faculty, TAs or students. It's good for us to take the time to hear they have to say - we often get great ideas. It also teaches them more about our work, and they leave more of advocate for our service than they were before the test.

Tags:

Comments | 2005-09-12

MDID

MDID is an open source project with much in common with Stellar Images.

"Developed at James Madison University, the Madison Digital Image Database (MDID) software brings the digital image library into the teaching and learning process."

The big difference is that MDID is an image repository, ours is a tool that accesses image repositories (it could potentially access MDID repositories along with DSpace, HarvestRoad, etc.) MDID os moving towards accessing other repositories, but the general approach of Stellar Images is much lighter.

MDID is worth keeping an eye on - it has many of the feature we might strive for in future releases. One possible disadvantage for us is that it's an ASP/.NET application.

Tags:

Comments | 2005-09-08

Access Denied, with a way out

The new Stellar access denied pages are getting quite a work out in these first couple days of classes. The page is the usual access denied page, but if person trying to get into the class is logged in successfully (i.e. has an MIT certificate) they can use a form on the page to send mail to the instructors and administrators of the class asking for access. The course staff then get an email with the request, plus a link to the page where they can rant access. There have been 300 requests since the new page debuted on July 31. 155 of those requests were in the last 24 hours.

I heard from Rich Garcia in Academic Computing that the number of support requests coming in is much lower than in other semesters. All of those people getting access denied messages now have an action they can take, other than calling the help desk. This is also much easier on the instructors, who get a link directly to the access page, rather than having to puzzle through the process any time someone asks them for access.

But why are there so many people being denied access? Mainly it's that Stellar is operating off the pre-reg lists now, the since Reg day was just Tuesday, the registrar's official lists aren't out yet. Here's a sampling form the last 100 messages that have come through:

  • 40% Unknown - either they didn't leave a note or it was unclear why they wanted access.
  • 31% Not on Pre-reg list
  • 15% Auditing - either they want to audit for the semester or they are shopping for classes and wanted to look at this one more closely
  • 10% Registration problem - The student says they pre-registered, and they should have access
  • 2% Section request - The student is trying to join a section subsite
  • 1% Faculty - Faculty asking another professor to look at their course
  • 1% TA - TA assigned to the class hasn't been added as a TA yet

Looking at the first 100 messages, from before the semester began, I saw lot more requests from faculty looking to to see old versions of the class they are teaching, and from TAs trying to get into the class they've been hired to work on.

Although I'm pleased with the new page, I can see some improvements we can make.

  • Improve the reg feed. This is out of our control, but the faster we can updates from the registrar the happier our community of users will be. There are many "I turned in my registration form 5 minutes ago, and i still can't access this website" messages.
  • Different access denied page for sections. We have an good system for section signups, those emails from people trying to get into sections circumvent that system.
  • Avoid duplicates. Not a big problem yet, but I saw some potentially annoying duplicates. It seems some students were submitting the form, waiting a while, and if they hadn't been granted access yet, submitting the form again.
  • Let the class staff designate gate keepers. Currently every professor, TA and admin associated with a class gets the access request email - it would be nice if they could pick one or two people to handle these requests.
  • Make adding students even easier. I'd love to see the email to instructors have a "Add this student to the class" link which take them to a page saying "The student has been added, and they have ben sent an email letting them know." One click and the job is done.

Tags:

Comments | 2005-09-08

Why I removed bulk adds of images

In the initial wireframes for the Stellar Images showed a little check box next to each image that turned up after searching the repositories. There was a button to "add selected images" which would pull each of those images into the class collection. I recently removed that bulk add feature. Now to add an image, you click to view the image details, then click the "add to class images" button.

Why would I do that? There are more clicks involved, especially if someone wants to add 5 of the images that turn up in a search.

What I realized is that bulk adding is something some one on the development team would want to do when demoing or testing this app - get a bunch of pictures in there quickly. But it's not something I saw the professors doing when they prepared their own presentations.

The faculty, whether they were using Google Images, government archives or the slides in their own filing cabinet carefully select individual images to their presentation. For example, if some one wants a portrait of Richard Stallman for their history of open source computing class, they search for Stallman, then click the thumbnail that looks most appealing. Seeing the larger image they can see what it will look like in the presentation, and decide whether they'd really like to use it.

It also let me reduce a lot of clutter. The UIs for bulk action are always a little off putting for novice users, so by removing it, I can really improve the flow of the application. Someone somewhere is going to want to bulk add, but the people I am designing for, personified in the user personas I'm using, don't need bulk add, but do need a tool that's easy to understand.

There are 100s of little decisions like this that happen in UI design. Someday someone is going to ask me, or more likely someone one on the front lines of user support, why there is no bulk add. How do I communicate this kind of design decision?

Tags:

Comments | 2005-09-07

Mailman's HTML written by a Wildebeest

I'm cleaning up a GNU Mailman template to use with Stellar. This work involves replacing code like this:

  <TR>
<TD COLSPAN="2" WIDTH="100%" BGCOLOR="#FFF0D0">
  <B><FONT COLOR="#000000">Subscribing to <MM-List-Name></FONT></B>
</TD>
  </TR>

with code like this:

<h3>Subscribing to <MM-List-Name></h3>

Ay carumba. Mailman seems to be a lovely piece of software but HTML like that really makes a person question the entire application.

Tags:

Comments | 2005-09-06

Updated Stellar Images mockups

I spent much of the day updating the Stellar Images mock-ups and functional specs. I added a little JavaScript to make it look more like a working application.

Tags:

Comments | 2005-09-02

Small tools/Big ideas

I'll be presenting at the Small Tools/Big Ideas conference in New York on October 7.

This year's conference focuses on the ways that digital image libraries and newly developed digital tools are reshaping the practice of teaching art and art history.

I'll be there with Carl Jones, a programmer from MIT Libraries. We're working together on the image collection and presentation tool I'm designing for Stellar/Sakai. Carl works on DSpace, particularly on getting the Rotch Visual Collection online. We'll be talking about the strategies we're using to collaborate on that project.

Tags:

Comments | 2005-09-01