"Automatic Icons Organize Files"

17 Sep 2004 - 1:25am
10 years ago
11 replies
540 reads
Listera
2004

Thoughts?

Technology Review: Automatic Icons Organize Files
<http://www.technologyreview.com/articles/04/09/rnb_091704.asp?trk=top>

Everyone knows it is more difficult to memorize a set of facts or words than
a set of pictures. Everyone also knows it is way too easy to misplace files
on a computer.

Researchers from the University of Southern California, MIT, and ESC
Entertainment are aiming to improve the lost-in-cyberspace problem with a
tool designed to tap people's facility with pictures.

The system, dubbed VisualID, automatically generates detailed icons for
specific files. It assigns similar icons to related files by mutating the
original icon in a series. The degree of mutation depends on the degree of
similarity of the file names, which gives the user an approximate visual
sense of saliency, according to the researchers.

Ziya
Nullius in Verba

Comments

17 Sep 2004 - 9:00am
Schlatzer, Kurt
2004

Listera wrote:
> Thoughts?
> Technology Review: Automatic Icons Organize Files
>
<http://www.technologyreview.com/articles/04/09/rnb_091704.asp?trk=top
>
> already) http://subscribe-announce.ixdg.org/

I particularly enjoyed the following statement in the article:

"The sticker versions of the icons could be applied to real-world
objects like nearly identical tools on a submarine or the space
shuttle."

This statement illustrates the disconnect between academia and the
real world as suggested by Ziya in a previous thread. As a former
submariner, I can assure you submariners must be able to identify and
distinguish their tools by touch, not by sight, in the event of a loss
of power or in thick smoke. Glowing stickers would not be permitted as
the material is radioactive and not permitted on board. Strange as
that might sound, the stickers minute radiation could affect the
personal radiation measurement devices worn by crew members. Tools on
submarines are also kept in small, tight compartments where a visual
reference system would be ineffective. I suspect astronauts have very
similar requirements.

The other evidence of disconnect between academia and the real world
is the following line:
"It would take three to five years to develop the system fully,
according to the researchers."

;-)

Kurt Schlatzer
E-Commerce Project Management Analyst
Oxford University Press, USA
www.oup.com/us

17 Sep 2004 - 11:19am
Alain D. M. G. ...
2003

--- "Schlatzer, Kurt" <kurt.schlatzer at oup.com> a écrit :

> The other evidence of disconnect between academia and the real world
> is the following line:
> "It would take three to five years to develop the system fully,
> according to the researchers."
>

On the other hand, I think 5 years is a reasonable amount of time
because there is very little chance that we are going to have (before
that period) the huge, cheap high resolution wall sized (or let us say
cubicle sized, but spread out everywhere) OLED screens that are needed
to show so many significant icons at the same time.

Alain Vaillancourt

__________________________________________________________
Lèche-vitrine ou lèche-écran ?
magasinage.yahoo.ca

17 Sep 2004 - 11:21am
Alain D. M. G. ...
2003

--- Listera <listera at rcn.com> a écrit :
> Thoughts?

> Technology Review: Automatic Icons Organize Files
>
<http://www.technologyreview.com/articles/04/09/rnb_091704.asp?trk=top>

Great topic, mild to fair article and horrible bibliographical
reporting of where to get to the original research and who did it.
The original SIGGRAPH paper is titled VisualIds: Automatic Distinctive
Icons for Desktop Interfaces. There is a PDF of it available here:

graphics.usc.edu/cgit/pdf/papers/visualids.pdf

The first author is J.P. Lewis, but the person who really has a solid
past in this is the second author, Ruth Rosenholtz. She co-authored an
equally interesting paper on distinctive thumbnails for representing
Web pages in search results in 2001: Allison Woodruff, Andrew Faulring,
Ruth Rosenholtz, Julie Morrsion, Peter Pirolli: Using thumbnails to
search the Web. CHI 2001: 198-205. There is a PDF available here:

www2.parc.com/istl/projects/uir/pubs/
items/UIR-2001-01-Woodruff-CHI2001-EnhancedThumbnails.pdf

One of the most interesting things about the VisualIds research is that
unlike the few other serious and advanced attempts (by persons doing
research or practical R and D in things like Zooming User Interfaces
for instance) to give a distinctive graphic identity to digital
documents on the desktop, these scientists are focusing on existing
desktop infrastructures. Instead of reinventing the wheel (of the
direct manipulation interface) or doing away with it (as some
intelligent agent advocates would have it) they are proposing to look
at ways to change it gradually, but in fairly significant ways.

I think there should always be both: Research on totally new vectorial
systems that go beyond Windows (or Gnome or Linux or the 3D Looking
Glass prototype at SUN for that matter) and the down to earth focus on
things like generated 64 by 64 pixel images that are formed by a
mathematical and abstract graphical transformation of the terms used in
the file names.

I wonder, though, why they are using data glyphs but do not mention the
work that has been done, by Flury and Riedwyl, De Soete and others, in
building upon Chernoff faces, finding out the salient features of human
faces to represent multi variate data.

The thing to bear in mind the most, I think, is that (given the
limitations of scholarly research and publishing) they offer only one
way of generating distinct images for the desktop. There are many
more, and in my opinion they are far from exclusive.

Alain Vaillancourt

__________________________________________________________
Lèche-vitrine ou lèche-écran ?
magasinage.yahoo.ca

17 Sep 2004 - 1:17pm
Jef Raskin
2004

> Glowing stickers would not be permitted as
> the material is radioactive

Not all are; most are not. But your main point that experienced
operators don't need labels is the important one (and one I wish more
designers would realize). Still, glowing indicators (of the
non-radioactive kind) would be useful for training; it is usually
better to use words rather than icons, however.

17 Sep 2004 - 2:33pm
Schlatzer, Kurt
2004

Listera wrote:
> Thoughts?
> Technology Review: Automatic Icons Organize Files
>
<http://www.technologyreview.com/articles/04/09/rnb_091704.asp?trk=top
>
>
> Researchers from the University of Southern California, MIT, and ESC
> Entertainment are aiming to improve the lost-in-cyberspace problem
> with a tool designed to tap people's facility with pictures.

What about accessibility or cultural issues? I guess those researchers
will need another 5 years to create AudibleID, then 5 years for
TactileID, then...

Kurt

17 Sep 2004 - 4:16pm
Jef Raskin
2004

But are icons a good idea in such a context?
On Sep 17, 2004, at 9:19 AM, Alain D. M. G. Vaillancourt wrote:

> [Please voluntarily trim replies to include only relevant quoted
> material.]
>
> --- "Schlatzer, Kurt" <kurt.schlatzer at oup.com> a écrit :
>
>> The other evidence of disconnect between academia and the real world
>> is the following line:
>> "It would take three to five years to develop the system fully,
>> according to the researchers."
>>
>
> On the other hand, I think 5 years is a reasonable amount of time
> because there is very little chance that we are going to have (before
> that period) the huge, cheap high resolution wall sized (or let us say
> cubicle sized, but spread out everywhere) OLED screens that are needed
> to show so many significant icons at the same time.
>
> Alain Vaillancourt
>
> __________________________________________________________
> Lèche-vitrine ou lèche-écran ?
> magasinage.yahoo.ca
> _______________________________________________
> Interaction Design Discussion List
> discuss at ixdg.org
> --
> to change your options (unsubscribe or set digest):
> http://discuss.ixdg.org/
> --
> Questions: lists at ixdg.org
> --
> Announcement Online List (discussion list members get announcements
> already)
> http://subscribe-announce.ixdg.org/
> --
> http://ixdg.org/
>

18 Sep 2004 - 12:49am
Jef Raskin
2004

>>
>
>> Technology Review: Automatic Icons Organize Files
>>
> <http://www.technologyreview.com/articles/04/09/rnb_091704.asp?trk=top>
>
> Great topic, mild to fair article and horrible bibliographical
> reporting of where to get to the original research and who did it.
>

From the article: Everyone knows it is more difficult to memorize a set
of facts or words than a set of pictures.

I wonder about that. I notice that tooltips, with words, are often used
to explain icons, but rarely does an icon pop up to explain a word.

18 Sep 2004 - 3:42pm
Michael R. Bernstein
2003

On Fri, 2004-09-17 at 22:49, Jef Raskin wrote:
>
> From the article: Everyone knows it is more difficult to memorize a set
> of facts or words than a set of pictures.
>
> I wonder about that. I notice that tooltips, with words, are often used
> to explain icons, but rarely does an icon pop up to explain a word.

I think different (but related) tasks are being conflated here; simple
memorization of arbitrary lists of words or images (which is what I
assume the quotation refers to) is not the same thing as using words to
disambiguate (and thus help learn) the meaning of an icon, where the
icon is subsequently a more easily recognized/recalled signifier.

The idea behind the technology described in the article seems to be that
subtly altered images can be used to disambiguate between otherwise
similarly named and placed digital files, as well as between otherwise
similar physical objects. This is disambiguating identity, not meaning,
but the image is then apparently also more easily recalled.

I'd be interested to know whether the image change pattern can be
learned, so users can *predict* what they should be looking for (as
naming schemes can be learned), or if the pattern of changes is
arbitrary enough that it only serves as an aid to recall.

--
Michael R. Bernstein <webmaven at cox.net>

20 Sep 2004 - 8:14am
Alain D. M. G. ...
2003

--- "Michael R. Bernstein" <webmaven at cox.net> a écrit :

> I'd be interested to know whether the image change pattern can be
> learned, so users can *predict* what they should be looking for (as
> naming schemes can be learned), or if the pattern of changes is
> arbitrary enough that it only serves as an aid to recall.

The patterns are abstract and their original assignment is arbitrary
but the algorithm has the effect of clustering files with similar
names, and only then giving a common shape for each cluster, which is
then "mutated" for each individual file. In addition, the authors say
(in their paper ViusalIDs, which I urge you to read, but not in the
general article) that they intend to leave to each user the choice of
adjusting the parameters. So, at several levels some form of prediction
should be possible, after experimenting.

Alain Vaillancourt

__________________________________________________________
Lèche-vitrine ou lèche-écran ?
magasinage.yahoo.ca

20 Sep 2004 - 8:28am
Alain D. M. G. ...
2003

--- Jef Raskin <jef at jefraskin.com> a écrit :

> From the article: Everyone knows it is more difficult to memorize a
> set
> of facts or words than a set of pictures.
>
> I wonder about that. I notice that tooltips, with words, are often
> used
> to explain icons, but rarely does an icon pop up to explain a word.
>

This is partly an intellectual property issue. The images (still and
moving) are out there to explain anything and everything, but you have
to get the rights to use them first. The words are free. And then there
is the issue of indexing. It takes a controlled vocabulary and/or a
thousand words to really describe an image so getting the right image
at the right place is tricky. Finally, there is the sad fact that much
of computer science and its practitioners inhabit a conceptual world
where alphabets reign supreme and visual elements are few. You can
sometimes lure them out with colors though.

Alain Vaillancourt

__________________________________________________________
Lèche-vitrine ou lèche-écran ?
magasinage.yahoo.ca

20 Sep 2004 - 8:38am
Alain D. M. G. ...
2003

Traditional generic icons, no! Significant generated icons, which are
in fact a desktop outgrowth of the data glyphs used in Information
Visualization, and are in a way conceptual thumbnails, yes!

Of course, there is also the possibility that ininitely scalable
vectorial representations of the documents themselves (such as proposed
in the interface formerly known as THE) might also be a good idea in
this context. But for this to be so they would need to have richer
representations than just showing the first page of a text (or
spreadsheet or whatever) and would thus require a variety of accretion
and annotation mechanisms such as those that exist in the real "paper"
world to create visual diversity in documents.

--- Jef Raskin <jef at jefraskin.com> a écrit :
> But are icons a good idea in such a context?
> On Sep 17, 2004, at 9:19 AM, Alain D. M. G. Vaillancourt wrote:
> > On the other hand, I think 5 years is a reasonable amount of time
> > because there is very little chance that we are going to have
> (before
> > that period) the huge, cheap high resolution wall sized (or let us
> say
> > cubicle sized, but spread out everywhere) OLED screens that are
> needed
> > to show so many significant icons at the same time.

Alain Vaillancourt

__________________________________________________________
Lèche-vitrine ou lèche-écran ?
magasinage.yahoo.ca

Syndicate content Get the feed