Forum Posts Feed
Organizing vs. searching
< Next Topic | Back to topic list | Previous Topic >
Posted by Paulo Diniz
Nov 26, 2006 at 05:33 PM
Hi Daly,
Wouldn’t it be easier to make food a master category, and put all those kinds of food as slave categories??
If i tag something “vegetable”, the system would know it is also a “food”, because it is a slave category for food.
But i was doing a quick test with master/slave categories on IH right now, and it seems that when you assign a master category for an item, it also assigns all the slave categories altogether, and not the opposite.
I once talked about this with a member of the evernote forum (also a tag/category-supportive PIM) on our thoughts as how a tagged PIM should work, and he mentioned a feature that is similar to the master/slave categories on IH, but that worked a little bit different. I will quote him:
“I see your point about the categories vs. tagging. The problem with plain tagging, i think, is that it is very flat. It’s all on one level. There’s no multi-dimensionality. So, while i don’t need to have *data* hierarchical organized, I *do* want to have their *metadata* organized.
For example, say I have an article about Blaise Pascal, the French mathematician, physicist and religious philosopher. And let’s say I want to tag it. How would I tag this article? Simple: “Blaise Pascal”. Then what? Then I create tag relationships (hierarchies):”
“-Science
—-Physics
——-Physicists
———-Blaise Pascal
—-Mathematics
——-Mathematicians
———-Blaise Pascal
-Social Sciences
—-Philosophy
——-Philosophers
———-Blaise Pascal
-Spirituality
—-Religion
——-Religious philosophers
———-Blaise Pascal”
“As you can see, in this way, the tags are organized in logical groups, and, unlike the tree hierarchies of, say, the windows file system, a tag can belong in many different places in the hierarchy (the same tag, not copies of it). This also eliminates the need to assign many tags to each item so that it can be found later (one of the major problems of del.ic.ious. In delicious, I assign 6-7 tags to each url, and as you can understand, I always forget one tag or another, so my tagging system is not consistent at all. For example, i may have tagged one article about Aristotle with 6 or 7 tags, and another article again about Aristotle with 4-5 tags—it’s hard to remember your own tagging methodology, the mind can only remember so much.) In the above example, I only have to tag as “Blaise Pascal” and then the system “knows” all the different places it needs to put it based on the tag relationships I have previously defined (in this example, I have already said to the system (via the tag hierarchy/relationship) that Blaise Pascal is a philosopher, mathematician and physicist, so why should I have to say it again and again and again everytime I tag something about him?? It should be done *once*, not every time). “
The way he proposes the system seems better than the one on IH. My only fear is that by implementing a full-blown hierarchical system of all meta-data, where every tag resides visually in one (or more) places of a tag structure created by the user, this would risk making the user feel the urge to manage a huge structure of metadata relations before deciding which *single* tag will be most appropriate for use for the data, thus making the switch back from what *would be* an tagging system to an actual meta-data hierarchical system.
I don’t know if i’m making myself clear. I just know that i haven’t found my ideal tagging PIM yet. I hope that programs like Evernote, InfoHandler and Personal Knowbase evolve or inspire other developers to make new tagging PIMs, with new solutions. I’d better stop now, because after all, this is a forum for discussing *outliner* software and not for discussing how the ideal tagging PIM should be, but the openness of the first post of this thread invited for a bit of “philosophical” questioning…
But, if someone is developing any tagging PIM, i have great interest on being a tester for it, please contact me at orcsbr(at)gmail(dot)com
Cheers Daly,
-Paulo
>I think IH gives you a way around that problem. If you want
>food to be a category, you can have it as a category. And if you want anything you
>categorize that is a food to show up under food as well as categories you may have for
>vegetables, meats, fruits, home grown, easy-to-cook, etc. you just make the food
>category a slave to each of those categories. The when you categories chicken a meats
>it will automatically show up as a food also.
>
>The slave/master and parent/child
>category system Manfred provides for IH allows some really neat tricks with
>categorization.
>
>Daly
Posted by Kenneth Rhee
Nov 26, 2006 at 05:54 PM
Paulo Diniz wrote:
>Hi Daly,
>
>Wouldn’t it be easier to make food a master category, and put all those kinds
>of food as slave categories??
>If i tag something “vegetable”, the system would know
>it is also a “food”, because it is a slave category for food.
>But i was doing a quick test
>with master/slave categories on IH right now, and it seems that when you assign a
>master category for an item, it also assigns all the slave categories altogether, and
>not the opposite.
>
I believe for what you need to do, you need to make vegetable your master and food your slave. When you think about it, it makes sense since vegetable always belong to food, but there are many categories under food. So, make vegetable your master and food your slave. If you use this format for your example for Pascal, you will see that even if the hierarchy goes down multiple levels, it should work as well. The beauty is that you only need to do this once, you don’t have to do it again. For your example, if I select Pascal, all the upper relationship will be tagged according to the system I created.
Hope this helps.