[Design] Fwd: Tags, hierarchies et al.

[Mimi Yin]

Rashmi's response to my wiki post.

Begin forwarded message:

> From: Rashmi Sinha <rashmi at uzanto.com>
> Date: August 30, 2005 9:36:41 AM PDT
> To: Mimi Yin <mimi at osafoundation.org>
> Subject: Re: Tags, hierarchies et al.
>
>
> At 12:40 PM 8/29/2005, Mimi Yin wrote:

>> Thanks for the pointer to the thread. I think Chandler's
>> interpretation of this whole debate is to get rid of the distinction
>> between TAGSs in Flickr and Delicious, labels in Gmail, Categories in
>> Outlook and Entourage, Smart playlists in iTunes AND FOLDERS in most
>> email clients, sets in Flickr and explicit playlists in iTuness.
>>
>> Tags, Labels, Categories, Smart playlists, Folders, Sets and
>> Playlists are technology-centric. They focus on distinctions in
>> functionality and software behavior.
>>
>
> I think that makes sense. Here is why. I have been thinking about  
> what the cognitive difference between tags and categories is. As an  
> input mechanism, tagging is just an association (it could be a  
> descriptor, a category, or any other type of association). Its  
> whatever comes to mind, without any editing.
>
> In contrast, a category, is the right group that the item belongs.  
> Its a subset of what a tag is. Since the focus is on the "THE one  
> right category", it can be a complex task sometimes (you have to  
> reject all the other associations, and focus on one).
>
> When people talk about tagging, i notice that they emphasize that  
> its easy, because there is no one right answer. The tags themselves  
> still tend to be categories. When I look at tags on del.icio.us/ 
> flickr,  a lot of times those are possible categories. So in  
> practice, one difference between assigning tags or categories might  
> simply be whether you assign one or more.
>
> Another difference: what happens to the item when you tag /  
> categorize it. Assigning to one category generally implies I am  
> done with the item. In contrast, assigning one tag does not imply  
> that you are done with it. In del.icio.us / Flickr you can assign  
> as many tags as you want. Once you are done, the item is at the top  
> of the list - you can edit it anytime you want. But in a folder  
> system, once you categorize an item its gone from your field of  
> view, its hidden in a folder somewhere. I think this is important -  
> people seem to want one or the other. Some people want the item to  
> go away once you tag/categorize something, others want it to remain  
> in field of view. One thing that we are trying in our sorting  
> software is to always show the last three items you put in a category.
>
> There are other several other minor differences. But coming back to  
> my original point - the differences in the system are not  
> fundamental to tagging / categorization (meaning assigned tags  
> often seem to be categories). For example, I can easily imagine a  
> categorization system that allows multiple categories, and the item  
> remains available. It is a question of UI affordances.
>
>
>> I guess the important distinction in my mind's eye is not so much
>> between a folder or a tag...but more between:
>>
>> 1. Explicitly creating a group of items in a top-down manner (ie.
>> collecting them into a folder or playlist) versuss
>> 2. Implicitly creating a group of items in a bottom-up manner (ie.
>> tagging or labeling and then creating a smart playlist or smart
>> folder based on that tagging)
>>
>
> The top-down, bottom-up distinction is important. I went back and  
> read some of the cognitive anthropology work. In real life,  
> categorization or learning categories is definitely made up of  
> small, local decisions. Think of the way a child learns categories,  
> say living versus non living. Her parents make small local  
> observations about objects she encounters. Together this adds up -  
> she gets the living-non living difference. Real life categorization  
> is really small, local, independent observations (all the different  
> associations that arise when we encounter an object). The  
> categories arise out of that choas of small, local observations -  
> our minds are able to the analysis and synthesis to find categories  
> to organize those observations.
>
> Look forward to discussing more of this with you when you are in  
> the South Bay. And feel free to post this to the Design list/ wiki.
>
> -rashmi
>
>
>> This reduces the debate down to UI affordances (ie. drag and dropping
>> items into some grouping mechanism, be it a Gmail label or an Eudora
>> folder VERSUS labeling an item as belonging to some label or folder).
>>
>> Consequently we could easily imagine a world in which there was some
>> functionality that collected items into some BUCKET, and users could
>> fill that bucket in a number of different ways: bottom-up or top-  
>> down, but that they DIDN'T need to worry about what kind of bucket it
>> was (ie. tag or folder), unless it was a human-centered distinction
>> (ie. a Project bucket versus a Location bucket).
>>
>> Anyway, looking forward to talking with you more about all of this!
>> (Would you mind if I posted this exchange to our Design list?)
>>
>> Thanks, Mimi
>>
>> On Aug 29, 2005, at 10:47 AM, Rashmi Sinha wrote:
>>
>>
>>>
>>>
>>> I came across your Virtuality presentation slides: http://  
>>> wiki.osafoundation.org/bin/view/Journal/VirtualityPresentationSlides
>>>
>>> Very interesting, though I am not sure I got everything. We should
>>> catch up sometime - I am fascinated by what Chandler is doing.
>>>
>>>
>>> This thread on my blog might be of interest: http://  
>>> www.rashmisinha.com/archives/05_07/oddpost-folders.html
>>>
>>> I am really curious about Gmail tags are really being used. Are
>>> people jerry-rigging folder systems (as the last comment suggests).
>>> Wonder if there is any data out there about this.
>>>
>>> -rashmi
>>>
>>> o----------o---------o--------o-------o
>>> Rashmi Sinha, Ph.D.
>>> Uzanto Consulting
>>> www.uzanto.com
>>> www.rashmisinha.com
>>> o-------o-------o------o-----o----o--o
>>> Co Program Chair - BayCHI
>>> www.baychi.org
>>> o-------o-------o------o-----o----o--o
>>>
>>>
>>>
>>
>>
>>
>
> o----------o---------o--------o-------o
> Rashmi Sinha, Ph.D.
> Uzanto Consulting
> www.uzanto.com
> www.rashmisinha.com
> o-------o-------o------o-----o----o--o
> Co Program Chair - BayCHI
> www.baychi.org
> o-------o-------o------o-----o----o--o
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osafoundation.org/pipermail/design/attachments/20050830/92ee42b0/attachment.html