CHI-WEB Archives

ACM SIGCHI WWW Human Factors (Open Discussion)


Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
"ACM SIGCHI WWW Human Factors (Open Discussion)" <[log in to unmask]>
Hal Shubin <[log in to unmask]>
Mon, 10 Nov 2003 16:31:53 -0500
text/plain; charset="us-ascii"; format=flowed
Hal Shubin <[log in to unmask]>
text/plain (163 lines)
I asked this question:

>... We're looking at how tall to make the list -- it can either hold an
>integral number of rows of data (fitting n rows exactly) or the last row
>could partly fit and be cut off at the bottom of the list.
>Visually, I prefer to have an integral number of rows. Functionally, I
>wonder if having the last row cut off will help users realize that there
>is more information they can scroll to see.  ...

Y'all wrote these messages back. Seems to be a consensus that the last line
should be clipped to provide an indication that there is more information.

Two people pointed out that it may be out of my control anyway.

Thanks for the information. My original design was to ask to not have any
rows clipped. Glad to have people's support for changing my mind. We'll see
if they can do anything to make sure it happens.

                                         -- hs

From: "Scott Berkun" <[log in to unmask]>

There is definitely an unresearched rule that I've always used here - yes
you do want to clip that last line to help indicate that things scroll.
I've never seen any research anywhere on this (though I do recall looking
for it at least once through and other places, and not finding
any), but it does roughly follow the logic of what newspapers and magazines
do: whenever an essay spans multiple pages or jumps to the back of the
magazine, it's *always* broken in the middle of the sentence. It seems
natural that this helps convey to the reader that there is more to read,
and in the same way, I think that broken line has a similar, though
probably weaker, influence on user expectations.

At various times on Windows and IE we looked into improving ways to
indicate possible scrolling to users that were, to be blunt, less ugly and
more refined than clipping a line of text, but we never got around to
building them. We were motivated to look into this at all based on the
frequent observations we made of few people recognizing the smartness of
the scroll bar (how it indicated by it's size whether there was more to
read) Initially some folks believed that this was enough feedback, but our
observations didn't bear this out. Anyway, this was a few years ago, so
perhaps more people notice it - but I bet many still won't.

More important than my above ramblings however is probably this: What is
the user task? If it involves finding a specific entry, *and* the user is
1) pretty sure they're on the right page, 2) pretty sure they're looking at
the right list, and the thing they want is not immediately visible, I bet
most people will try the scroll bar most of the time, regardless of how
smart it is. So my bet is this specific design choice is relatively moot,
compared to the importance of designing to satisfy the conditions I just
listed (#1 & #2).

[and after a reply from me...]

Cool - good to know my assumptions were based in some way on reality.

One more idea: I know precisely the problem you're dealing with, and I've
done some hacks here and there to get around it. The one that comes to mind
is to stuff a placeholder into the last whole entry in the list, that
automatically loads the next data in the list. Something like:


It helps trigger the reaction that there is something special going on, and
it provides a way to click and automatically refresh the list without
having to mouse over to the scroll bar (arguably avoiding the costs of
Fitts' law by avoiding that big mouse move). I've seen this work well a few
times, though it does eat that last entry in the list, and it can be ugly.
Also, this usually requires some hacking to the list control that your
developers might not be bribed into doing.

From: "Karuna Venter" <[log in to unmask]>

I have always found that with web pages that scroll, it's better to have
what looks like an incomplete page above the "fold" to cue users to scroll.
If the page when it loads looks complete, they are less likely to scroll,
because the scrollbars on web browsers are pretty much ignored, and with a
complete-looking page those scrollbars are the only visual cue to scroll
down.  No data to share on that, just the experience of watching users.

From: Jeff Johnson <[log in to unmask]>

I favor letting the list cut items off in mid-line.  Otherwise, you need an
explicit indication of whether there are more items below those
shown.  Cutting items off mid-line is a subtle but clear way of indicating
the same thing.

From: Leo Frishberg <[log in to unmask]>

Without a doubt a partial row at the bottom is absolutely, hands down, the
only way to go.

Can I make my opinion more forceful? <smile>

Actually ran a few tests on this one in the past, and while I don't have
the data in front of me, my hypothesis was that we'd see far fewer errors
(mistaken perception of complete data set) with a partial row than with a
"visually" neat complete row.

This played out pretty cleanly.

I've even begun to consider a whole new theory of screen design as a
result: messy artifacts are key indicators of "see more data".  In other
words, we like things neat and tidy and when they aren't, it's cause for
investigation.  The notion of the "loose thread".

I'm sure there's literature to support all of this as well.

From: Nancy Shea <[log in to unmask]>

The experience I have is with a financial application, with a core grid. We
were confronted with the same issue and decided to allow for partial row
display at the bottom for the following reasons:

- (the one you list) it exposes the fact that there is additional data.

- you avoid the awkward system adjustment of the screen to fit the precise
measurements you define when the user does scroll. this can be jarring.

- this is what excel does, so our users are accustomed to this behavior

I agree with you that on a static screen the integral number of rows is
more appealing, but I think interacting with such a system can be awkward.

Looking forward to hearing what others think.

From: "Lew Clayman" <[log in to unmask]>

If you refer to the visible portion of the list, it may be out of your
control - that is to say, it will likely render differently on computers
with different browsers, display settings, etc.  On the web, you have be a
lot be very fluid about layout details.

From: "Sahala Swenson" <[log in to unmask]>

I'm a little confused, in almost any current implementation of a scrollable
list, isn't it pretty universal to partially show rows that are not within
the current scroll view?

Hal Shubin
Interaction Design -- Design & strategy for the Web, 617 489 6595

    Tip of the Day: Use the archives to research common questions
     CHI-WEB: POSTINGS: mailto:[log in to unmask]
              MODERATORS: mailto:[log in to unmask]