Pagination vs Sortation

15 Nov 2005 - 5:05pm
9 years ago
13 replies
833 reads
Jeff Howard
2004

I'm working out the relationship between pagination and sortation and wanted to get some
feedback. Pagination has traditionally been a web paradigm, while soration has been more of
a desktop convention. As the distinction between the functionality of web apps and desktop
apps breaks down, it's possible to see both in one place. That creates some friction.

Say you're on the 2nd page of a paginated set of tabular results. Say, Books by Author. You
see Daniel Boorstin's books and decide to sort that page by publication date to see his works
in order. You don't want to lose the books you were looking at, so you appreciate that the
page sorts in place without requiring a page refresh. But then... you decide to see page three.
What constitutes the third page? Is it the original third page of results, say, authors with last
name starting with "C" or is it now the 3rd page as it would have appeared had the list been
sorted by publication date from the beginning? Or, from the date of the last book on the 2nd
page?

Apps like Hotmail address this issue by requiring a page refresh when sorting that sends the
user back to the first page of the newly sorted result set. Apps like Gmail address this issue
by not sorting.

Is there some middle ground where pagination and sortation can co-exist?

Comments

15 Nov 2005 - 5:39pm
Baldo
2005

> Is there some middle ground where pagination and sortation can co-exist?

quite ot:
there is another method that works fine with pagination and sortation
and this is filtering:

take a look at the "Find By" links
http://computing.kelkoo.co.uk/b/a/c_114401_monitors.html

15 Nov 2005 - 6:08pm
Daniel Harvey
2004

I can't think of an implementation that works the way you describe in the second paragraph. If I'm taking an action to sort then I don't want to keep the prior list. Thus, in your scenario from the second paragraph I would expect it to be by publication as opposed to alpha or whatever the prior sort was. Amazon and most others behave this way.

Most have the sort function trump pagination (as it should since the default pagination is still determined by some defailt sort even if it's not visible to the user although it should be so they can return to it) as you cite via the hotmail example. Conceptually, I'm okay with this since the user is initiating the change and may even need to see a refresh to get feedback that their action has been processed. Technically, this is because the db is processing the query and displaying another form.

Where this could change is with AJAX models and even then I suspect that the content would noticably refresh if not the page. Flickr, for example with large photsets, refreshes the content and flashes the screen (pagination is now determined by the new sort, fyi) even tho the page itself does not refresh.

-----Original Message-----
From: discuss-bounces at lists.interactiondesigners.com on behalf of Jeff Howard
Sent: Tue 11/15/2005 6:05 PM
To: discuss at lists.interactiondesigners.com
Subject: [IxDA Discuss] Pagination vs Sortation

[Please voluntarily trim replies to include only relevant quoted material.]

I'm working out the relationship between pagination and sortation and wanted to get some
feedback. Pagination has traditionally been a web paradigm, while soration has been more of
a desktop convention. As the distinction between the functionality of web apps and desktop
apps breaks down, it's possible to see both in one place. That creates some friction.

Say you're on the 2nd page of a paginated set of tabular results. Say, Books by Author. You
see Daniel Boorstin's books and decide to sort that page by publication date to see his works
in order. You don't want to lose the books you were looking at, so you appreciate that the
page sorts in place without requiring a page refresh. But then... you decide to see page three.
What constitutes the third page? Is it the original third page of results, say, authors with last
name starting with "C" or is it now the 3rd page as it would have appeared had the list been
sorted by publication date from the beginning? Or, from the date of the last book on the 2nd
page?

Apps like Hotmail address this issue by requiring a page refresh when sorting that sends the
user back to the first page of the newly sorted result set. Apps like Gmail address this issue
by not sorting.

Is there some middle ground where pagination and sortation can co-exist?
________________________________________________________________
Welcome to the Interaction Design Association (IxDA)!
To post to this list ....... discuss at ixda.org
(Un)Subscription Options ... http://subscription-options.ixda.org/
Announcements List ......... http://subscribe-announce.ixda.org/
Questions .................. lists at ixda.org
Home ....................... http://ixda.org/
Resource Library ........... http://resources.ixda.org

This message is the property of R/GA and contains information which may be privileged or confidential. It is meant only for the intended recipients and/or their authorized agents. If you believe you have received this message in error, please notify us immediately by return e-mail or by forwarding this message to postmaster at rga.com, and destroy any printed or electronic copies of the message. Any unauthorized use, dissemination, disclosure, or copying of this message or the information contained in it, is strictly prohibited and may be unlawful. Thank you.

15 Nov 2005 - 6:24pm
Larry Marine
2005

> there is another method that works fine with pagination and sortation
> and this is filtering:

True, but filtering, in my experience, has been particularly difficult for
the average user. Much of the difficulty stems from the fact that filters
often behave like a modal interface. And we are all familiar with the
usability failures with modal interfaces. Users often forget that they have
filters on and cannot understand why they can no longer see something that
they think should exist. Worse yet, they forget that they have the filters
on then get no results and think that something doesn't exist at all.

The original question was:
> Say you're on the 2nd page of a paginated set of tabular results. Say,
> Books by Author. You
> see Daniel Boorstin's books and decide to sort that page by publication
> date to see his works
> in order. You don't want to lose the books you were looking at, so you
> appreciate that the
> page sorts in place without requiring a page refresh. But then... you
> decide to see page three.
> What constitutes the third page? Is it the original third page of results,
> say, authors with last
> name starting with "C" or is it now the 3rd page as it would have appeared
> had the list been
> sorted by publication date from the beginning? Or, from the date of the
> last book on the 2nd
> page?

The users, having changed the sort, are likely to be confused by any
automatic regression to a previous sort if they choose to go to the next
page. What would happen if the Boorstin books were cut off at the bottom of
the page and the user meant to go to the next page to see the rest of the
books?

The interface should 1) let the user know what sort they have selected, 2)
provide a means for them to change/revert to the original sort. You are
correct in assuming that the user would want the list to rescale so that the
selected object is still selected, independent of where it ends up in the
new sort. Think of typical email clients, like Outlook, where the selected
email remains selected regardless of how you sort, such as by date or
subject.

BTW, users don't like nor perform well with pagination. They do much better
with scrolling sortable tables, especially if there is an obvious
progressive search mechanism that jumps them to the first instance of
Boorstin as the user types in BOORST. There are many problems with
navigation a paginated table, not the least of which are the moronic
implementations that merely give you a page number. How in heck is anyone
supposed to know what is on page 48 or 49? But I digress. Even alphabetized
page links don't do well, especially in long lists such as when there are 18
pages of B's. How does the user get to page 14 where Boorstin is located?

Mi dos centavos,
Larry

15 Nov 2005 - 6:58pm
Jeff Howard
2004

> rescale so that the
> selected object is still selected, independent of where it ends up in the
> new sort.

I guess the crux of my problem is that by doing a sort without a page refresh the user hasn't
left page 2, but Boorstin's books wouldn't really be on page 2 anymore (or even all on the
same page) if the list had been sorted by publication date from the beginning. Plus, unless
the pagination links update dynamically (this page is now 64 instead of 2... bad) then there's
no way to navigate to all the books that were published before Boorstin.

> What would happen if the Boorstin books were cut off at the bottom of
> the page and the user meant to go to the next page to see the rest of the
> books?

Right. If a user sorts in place (that is, limits their sort to the 20 items already onscreen), then
until they leave the page their results are ultimately sorted by the intersection of author
_and_ publication date. Once they leave the page, they lose the vestigal author sort... don't
they?

15 Nov 2005 - 7:11pm
Andy Kirkwood, ...
2005

Hi Anders,

We've implemented something similar to what you describe on a video
store website: < http://www.arovideo.co.nz >

When a user searches there are a number of methods for both sorting
(year of production, title) and filtering (Rental/Sales DVD/VHS), you
can also customise display options. Results are paginated but also
have a 'See all' option.

There are still issues when transitioning from sort to filter. The
model we've employed is to return the user to the beginning of the
search result (assuming a new task). A solution that does not
assume/rely on an earlier action/state is often preferable with web
interfaces. This is possibly a result of the limited interaction
opportunities previously available and the user's expectation that a
website is not an application. So in the example you've outlined,
pagination should update to reflect the new record set.

Ideally we'd enable the user to also select an item from the search
result as an anchor point. The new pagination or sort being centered
to that item. This would be similar to the desktop interface
convention for the Classic Mac: you can select an item in a folder
then resort but ensure the item selected remains visible in the
window. (More of an issue when there are significant numbers of files
in a folder).

The BOORST issue is less of a concern if it is accepted that the user
had changed task: from viewing all publications to only viewing works
by BOORST.

As a counter to Larry's note regarding user preference for long lists
over pagination, we've opted to provide controls to transition from
one to the other. This way user's can decide. By default, results on
the Arovideo website are paginated to improve the browsing experience
and minimise hosting fees. Pagination may be preferred for some users
as it provides stronger spatial orientation cues. Finally, at a
cultural level, a list of 100s of film titles can be perceived to
devalue each of the individual film records which is counter to the
intended positioning of the store.

Best regards,

--
Andy Kirkwood | Creative Director

Motive | web.design.integrity
http://www.motive.co.nz
ph: (04) 3 800 800 fx: (04) 970 9693
mob: 021 369 693
93 Rintoul St, Newtown
PO Box 7150, Wellington South, New Zealand

15 Nov 2005 - 7:25pm
Andy Kirkwood, ...
2005

Sorry Jeff (not Adrers),

Read the wrong details from my inbox ;)

å

15 Nov 2005 - 8:29pm
jbellis
2005

Looks like several potential cases-o'-use:
sort by publication date:
A. I want to see the earliest n publications
B. I want to see the latest n publications
sort author X's works by pub date:
C. I want to see (all of) Boorstin's works in pub order (irrespective of
current items-per-page setting)
D. I want to see (all of) some-author-that-I-haven't-clicked-on's works
in pub order (irrespective of current items-per-page setting)

When one clicks the column heading, do we know the user's intention from the
four choices above? Are there other choices I've omitted?

Regarding A vs B, above, when I sort by Price on eBay after searching on
"Shackleton," I want to see most expensive first (because those are the only
legitimate items) but eBay lists least expensive first. With those painful
expeditions to the server, it's like using cups-and-strings connected to my
2,000,000,000 cycles/second processor. So the nuclear option
(Ajax/Flash/Laszlo) might be the ultimate solution but probably not the
answer most of us seek today.

Perhaps when someone clicks a column heading in simple-minded systems
(Hotmail/eBay) we should make the dumbed-down choice for them [because
everyone reminds me all day long that users are dumb and don't want to make
decisions, right?], restarting the list at Pub Date=00/00/00 (all authors),
but in power systems, we should be prompting the user to choose from A, B,
or C, above?

Or if we persist in paginating results long after the main reason for doing
so (28.8 modems) has disappeared, maybe the prompt technique should be a
user preference? (Yes, typical Google searches return too many results to
eliminate pagination, but that wasn't Jeff's context.) I feel that user
control is the only absolute amid all the tough choices.
www.jackbellis.com, www.usabilityInstitute.com

----- Original Message -----
From: "Jeff Howard" <id at howardesign.com>
I'm working out the relationship between pagination and sortation
> Is there some middle ground where pagination and sortation can co-exist?

16 Nov 2005 - 3:21am
Ben Hunt
2004

I wonder if anyone's used AJAX with search results simply to get rid of
pagination, which does feel like an archaic interaction device.

I'd like to do a Google search and start getting results on the page, which
just keep a-comin'. Then I could use my mouse wheel or keys to keep
scrolling down. Fitt's Law comes into play. The pain of locating the 'Next>'
link over and over is reduced to zero because of the direct interface to the
mouse wheel/kb.

Perhaps it could always fetch 3 pages' worth below where I'm looking, to
prevent flooding the system?

- Ben

16 Nov 2005 - 3:22am
stephan at wiss...
2005

Jeff,

this is a very interesting question. My take would be to send the page
back for sorting (you could consider Ajax, so you don't send back the
whole thing) and then open not the first page, but the page the currently
visible/selected/focused record will be visible on. So if the title of D.
Boorstin's book is "Zulu programming" you probably would open page 756.
The challenge is to establish a clear clue what record has the focus when
doing so.
What do u think?

:-) stw

discuss-bounces at lists.interactiondesigners.com wrote on 16/11/2005
07:05:48:

> [Please voluntarily trim replies to include only relevant quoted
material.]
> Say you're on the 2nd page of a paginated set of tabular results.
> Say, Books by Author. You
> see Daniel Boorstin's books and decide to sort that page by
> publication date to see his works
> in order. You don't want to lose the books you were looking at, so
> you appreciate that the
> page sorts in place without requiring a page refresh. But then...
> you decide to see page three.
> What constitutes the third page? Is it the original third page of
> results, say, authors with last

16 Nov 2005 - 3:55am
Navneet Nair
2004

The Ajax based data grid implementation have a minor flaw in that there a
lag in the time that they actually send the request and fetch the data, at
this point the user is usually not sure how far they have scrolled in the
un-paginated list and so this may not really be really useful to them except
that they have not refreshed the page.
One way to handle this would be to diplay the first and the last data
element on the page in the pagination itself, something like this:
http://www.onclipevent.com/pagination.gif
Thoughts?
Navneet

----------------------------------------------------
Navneet Nair
Interaction Architect
onClipEvent: form follows function();
----------------------------------------------------
Website: http://www.onclipevent.com
Blog: http://www.onclipevent.com/enterframe/

16 Nov 2005 - 7:21am
Todd Warfel
2003

And then have a "wishlist" like function. When I'm surfing through
search results, I tend to open the one's I might be interested in in
a tab. That way I don't have to go back and forth. I open a series of
results, but continue to scan the search results list. Then when I'm
done, I go through the tabs, look at the results I've selected for
further reading and then close the ones I'm either done with, or no
longer interested in.

Basically, the search results gives me a pre-filter. Then I filter
again (tabs). Then filter a third time (either read and close, read
and keep open, read and print, read and bookmark).

I do the same thing with my RSS feeds.

On Nov 16, 2005, at 4:22 AM, stephan at wissel.net wrote:

> (you could consider Ajax, so you don't send back the whole thing)
> and then open not the first page, but the page the currently
> visible/selected/focused record will be visible on.

Cheers!

Todd R. Warfel
Partner, Design & Usability Specialist
Messagefirst | making products & services easier to use
--------------------------------------
Contact Info
Voice: (607) 339-9640
Email: todd at messagefirst.com
AIM: twarfel at mac.com
Blog: http://toddwarfel.com
--------------------------------------
In theory, theory and practice are the same.
In practice, they are not.

16 Nov 2005 - 8:27am
jbellis
2005

The types of valid objections that Ben raises are already greatly surmounted
in various solutions. Here are some:

1) Copernic.com provides a sort-of-free Windows app that gives users full
RAM power for web searches (completely eliminating pagination), not to
mention providing sorting. How often do you really need a true web (zero
footprint, anywhere-anytime) solution?
2) An advanced option on Google lets you set your results list to 100 (after
which even a search for references to Mother Theresa probably degrades into
scumware results). M. Fitts would probably then use the biggest key on the
keyboard, the spacebar, to scroll.
3) Vivisimo.com: clustered results, making pagination irrelevant. Not great
but promising. Try changing your home page to this one for a while and
always "open in a new window."

Regards, www.jackbellis.com, www.usabilityInstitute.com

PS, the more fascinating question is why we persist in using Google, myself
included. A friend tells me it's because they use RAM all over the world to
cache their data, rather than hard drive retrieval. Speed kills, but in this
case, competitors. And there's the GUI factor: they stay out of the web
design business and we all somehow blame the pathetically unusable search
results on everyone but Google.

----- Original Message -----
From: "Ben Hunt" <ben at scratchmedia.co.uk>

> I wonder if anyone's used AJAX with search results simply to get rid of
> pagination, which does feel like an archaic interaction device.
>
> Perhaps it could always fetch 3 pages' worth below where I'm looking, to
> prevent flooding the system?

16 Nov 2005 - 8:52am
Jared M. Spool
2003

At 06:39 PM 11/15/2005, Baldo wrote:
>there is another method that works fine with pagination and sortation
>and this is filtering:
>
>take a look at the "Find By" links
>http://computing.kelkoo.co.uk/b/a/c_114401_monitors.html

Also commonly known as faceted browse.

Jared

Jared M. Spool, Founding Principal, User Interface Engineering
4 Lookout Lane, Unit 4d, Middleton, MA 01949
978 777-9123 jspool at uie.com http://www.uie.com
Blog: http://www.uie.com/brainsparks

Syndicate content Get the feed