Submitted by Armando on 2009/07/08 19:21
In Smart Fields (aka smart folders) a discussion started about WikiTags.
 
So here's a thread to talk about it... :)

CamelCase, { } or comma separated...  : I personally don't mind the CamelCase. it takes less space (chars), especially if one is using the same tags all over, not just in IQ. But I think that it could be even simpler : each term put in the WikiTags field (camel case or not) should be a WikiTag, and only for those longer expression would there be a need for {} or what ever. The field already defines specific data, so why also use "{ }" ?   IMO "{ }"  should be used only to reference WikiTags from other fields.
 
Right now, IQ isn't really using a real WikiTag system anyways... since writing a word put in  { } (or [ ], or camel case... in other progs..) in any field doesn't tranform it into a wiki link as it would in a wiki application. This is why, for now, I use the WikiTags field as a simple category field (since the category field is not easily available in the add item dialog) and not use the {} system.
 
 

Comments

There is an interesting paper which I'll read in my spare time: http://www.hpl.hp.com/research/idl/papers/tags/tags.pdf .
 
At the moment, I tend to favor CamelCase as it is simpler and auto-wiki links are more easily implemented.
 
The other big question is a flat or hierarchy based tagging system (often referred to as taxonomy, while this is generally when multiple parents are not allowed)
 
I leaning towards a hierarchy based system, where the UI would be a grid where each tag is an item in the grid. This would allow flat and tree view of tags and allow multiple parents.
 
Adding a wikitag to an item which doesn't exist would create an item in this grid. Using the syntax ParentTag:NewTag, the wikitag could be created with an initial parent hierarchy.
 
Food for thoughts
 

Armando

2009/07/08 22:16

Permalink

I had a look at the paper. and found the "Kinds of Tags" part pretty interesting (it tends to "validate" and also complements some of my own observations -- not that they have anything special...). These classifications would be useful to users in general, looking for ways of organizing different types of tags.
 
>I leaning towards a hierarchy based system, where the UI would be a grid where each tag is an item in the grid. This would allow flat and tree view of tags and allow multiple parents.
 
Hierarchies are always useful for classifications, organization. etc. So I agree that this would be nice.
 
>Adding a wikitag to an item which doesn't exist would create an item in this grid. Using the syntax ParentTag:NewTag, the wikitag could be created with an initial parent hierarchy.
 
What'd be great about this is that users would be able to define tags precisely, add fields (notes, etc.). Maybe ? I currently do that in the AHK script I use for tagging.
 

gregory

2009/07/09 13:12

Permalink

Disclosure time - I'm researching personal and small-group information and knowledge management as Ph.D research while teaching in a French business school...
 
The paper Pierre references looks very good - thanks Pierre.
 
As Pierre says, "The other big question is a flat or hierarchy based tagging system (often referred to as taxonomy, while this is generally when multiple parents are not allowed)". The thing is, it seems to me, that both taxonomy (hierarchical classification) and tagging (multiple classification) have their places. A major issue in getting things done and keeping found things found is personal information management PIM, and a major issue in PIM is deciding whether to classify, to tag or both.
 
Part of the genius of InfoQube is that it (already) permits both, and in very flexible ways. It needs some development to make both easier (folder hierachies and better wikitags along the lines Pierre has just suggested).
 
Naked self-advertising (and Pierre please remove this post if you think it's inadmissible, but I have absolutely no commercial interest here, I get paid to teach and to research!) - part of my research is to get people to think about how they store and manage information (and hopefully tell me about it!). My current thinking, and a challenge to people like users of this forum and of IQ who already think a lot about personal and small-group information, is summarised in the attached PowerPoint presentation (hastily amended this morning to include a reference to the HP Labs research article Pierre highlighted then!). I hope IQ users will find it interesting; it mentions IQ more than once... Specifically, the presentation mentions taxonomic classification, tagging and introduces ontology (which is taxonomic classification on steroids!).
 
Mark Gregory, Rennes, France - GMT +1/+2; EST +6

Armando

2009/07/09 13:53

Permalink

Interesting stuff Gregory. I'm not in the same field, but I studied information management for my thesis a couple years ago -- so I'll have a close look at your presentation later on. Thanks for sharing !

Armando

2009/07/09 14:08

Permalink

Oh... And I'd like to add that it's probably not a coincidence that we're using IQ (started to use it more 1,5 years ago... and still motivated to contribute with my effort and time here). It can do so much in terms of info management ! Thanks Pierre.

KeithB

2009/07/09 14:20

Permalink

[quote=gregory]
 
...
My current thinking, and a challenge to people like users of this forum and of IQ who already think a lot about personal and small-group information, is summarised in the attached PowerPoint presentation (hastily amended this morning to include a reference to the HP Labs research article Pierre highlighted then!). I hope IQ users will find it interesting; it mentions IQ more than once... Specifically, the presentation mentions taxonomic classification, tagging and introduces ontology (which is taxonomic classification on steroids!).
 
Mark Gregory, Rennes, France - GMT +1/+2; EST +6
[/quote]
That's an excellent powerpoint--please consider posting it as a blog.