5 April 2004

Subway maps

I found two very interesting sites today, on subways.  The first
is a map showing the London Underground and how the map relates to the
reality of the street.  Check it out.  The next compares the size and density of various major transport systems worldwide.  It’s here.

Another site I found previously tries to fit the NYC MTA into the Tube’s design.  You’ll see what I mean when you check it out.

Posted in OnTheWeb on 5 April 2004 at 2:27 pm by Nate

Open Source conflict

This is probably below the radar for lots of you, but there’s a good
little discussion going on in the open source software world about user
interface and usability.  In regular people’s words, how easy does
the software make it for the user to do what s/he wants?  Can the
user figure out with ease how to get from A to C?

Eric Raymond, one of the high priests of open source, an expert of experts, couldn’t figure out how to use his own printerJohn Gruber, of Daring Fireball, notes

Raymond is ignoring the actual depth of the
problem. It’s easy to say, The open source community needs to do
better, we need to create software Aunt Tillie can use.
[emph. in original] But they’re so far
away from this right now that even an expert like Eric Raymond can’t
figure out how to use their software.

The “I thought I was the only one” letters that Raymond found so
interesting aren’t coming from the A.T.-set; they’re coming from Linux
geeks who read essays written by Eric Raymond. And they’re frustrated by
open source software’s terrible usability. The problem isn’t just that
dear old A.T. can’t use desktop Linux — the problem is that even Linux
geeks have trouble figuring it out.

Another perspective comes from mpt, on “why free software usability tends to suck.”

So why do I care?  Well, because the statistical software that I
use is open source (as per the trend at Harvard to push the open-source
software over commercially available alternatives with user interface
that make is easier for the novice to get up and running with his or
her data analysis), and its user interface sucks.  Yes, it’s very
powerful; yes, it’s customizable; yes, it has more flexibility than a
number of the commercial products, especially when you get to the
advanced level and need to program your own models.  But that’s
not very likely in the sort of research that most of us in political
science do.  The existing models are quite adequate, especially in
light of the fact that when one deals with social phenomena, you can
get very precise, but probably not very accurate.

But everyone here wants to see us use R (a variant of the S language),
rather than Stata or SPSS, or even SAS
So we’re all being turned
into open sourcers, whether we want to or not.  But the problem
comes when you decide, “Oh, I want to run X type of analysis.” 
You know the statistics that you want to do, but the damn interface
gets in the way of doing that without some significant programming, on
a steep learning curved.  It
takes quite a while to figure out how to get that analysis loaded, how
to run it, and how to get some results to come back to you.  Much
more
time than I remember it taking with the commercial software projects.

Posted in OnTheWeb on 5 April 2004 at 11:29 am by Nate