[Design] Fwd: Feature request for Chandler or other OSAF app - Tag-based organizer implementation

[Mitchell Kapor]

Begin forwarded message:

> From: Paulo Diniz <orcsbr at gmail.com>
> Date: November 15, 2005 6:58:33 PM PST
> Cc: mitch at kapor.com
> Subject: Feature request for Chandler or other OSAF app - Tag-based  
> organizer implementation
> Reply-To: orcsbr at gmail.com
>
>
> Hi from Brazil, Mr. Kapor,
>
> I'm pretty sure the idea has been evaluated by the OSAF staff, for  
> what i've seen on a brief reading of  Mimi Yin's reports on
> (http://wiki.osafoundation.org/bin/view/Journal/MimiYin)
>
> But even though, i'd like to make an "special" request for this  
> feature: make personal organizer software that is fully  
> incorporated with data "TAGGING" , just like del.icio.us. I mean,  
> extend the flexible tagging concept from del.icio.us to any kind of  
> data, not just bookmarks. And with this concept, make it possible  
> for the user to TOTALLY abandon rigid taxonomies (outlining), which  
> is only good for a limited ammount of data, but turns out into a  
> problem when you have to organize HUGE quantities of data.
>
> I'm no app developer, just a long-time PC user, but i think the  
> following idea would be compatible with the semantic web effort.  
> (maybe?)
>
> My concept of the ideal PIM software is the one that can deal with  
> the massive ammount of info we have to deal with everyday, and make  
> that availiable over an very big ammount of time (a lifetime  
> span?), and is efficient for filtering the exact info you desire  
> from that lifetime database. I've been thinking a bit about how  
> that application would have to be, and i've gotten to the following  
> conclusions (the following is copy of a draft i wrote, so sorry for  
> any repetition of ideas):
>
> ----
> 1) hierarchical is a waste of time, because it works just like real  
> world. one data only can be at one place, unless you are willing to  
> copy and paste like crazy. Rigid taxonomies are good for science,  
> but no good for inputing info on a fast way, and then finding it,  
> after you’ve grown a massive database.
>
> 2) hyperlinking is a waste of time, because it envolves the same  
> work as copy and pasting. you’d have to go beserk hyperlinking all  
> around after you build a large database, to make it work.
>
> 3) the human brain works just like tags on delicious. when you  
> think of a chicken, you dont think about kingdom, phylum, class,  
> orden genre, specie. You basicly think “ANIMAL, BIRDS, CHICKEN”. We  
> tag stuff free of a rigid structure.
>
> 4) massive search (e.g google) is good, but you dont have any  
> control on the results other than the search keywords. If you tag,  
> you do, because you reviewed that info, and added the tag metadata  
> to it. You say what it is.
>
> 5) if you want to build a big database out of tagging, you need to  
> have a good tag searching mechanism, because sometimes it takes 4  
> or 5 combined tags to filter the things we dont want. Just like  
> del.icio.us: you can combine tags and see which bookmarks have all  
> the tags you selected. You need to be able to merge or split chunks  
> of info, and you also need a very good tag manager: you need to be  
> able to mass tag stuff, mass un-tag, etc. You have to be able to  
> fastly select TAGS for an information you will be inputting. The  
> organizer cant get in the way.
>
> 6) Its known that the the more you input information, the better  
> your database is. If your database uses tags, theres another factor  
> on the equation. The smaller are your tagged bits of information,  
> the better your database is, upon retrieval.
>
> Example:
> I go to my history class, with laptop, about american history.
> the teacher starts to talk about the american independence, lets  
> say im a history illiterate and dont know the very basics, so i  
> start typing most of the information told to me by the teacher:
>
> “The american independence was in 1776. The first president was  
> George Washington. He is a founding father of the american  
> republic. The expression founding fathers is related to…”
> And tag the class information as: history, world, USA, independence.
>
> When i got home, i would review the class for a exam on the next  
> day. I could take the chance to create new tagged bits of  
> information, more specific, with what i consider the most relevant  
> information. This helps further searches, as well as helps in the  
> learning and memorizing process. Example:
>
> “The american independence was in 1776.”
> TAGS: history, world, america, USA, independence, year
>
> “The first president was George Washington”
> TAGS: history, world, america, USA, country, president, first, leader
>
> “The expression founding fathers is related to…”
> TAGS: history, world, term, expression, america, USA, republic
>
> and then 6 months later, i go into brazillian history. Off all of  
> the information i captured, as “history, local, brazil” i can make  
> an example of this specific bit i could tag:
>
> “The brazillian independence was in 1822.”
> TAG: history, local, america, brazil, independence, year.
>
> So, 50 years later, when helping my grandson on a school paper  
> about independence of american conutries, i could filter the tagged  
> bits with only the following tags:
>
> “history, america, independence, year”
>
> and would get a list of tagged bits containing all the independence  
> years of countries i have studied on america. Cool huh? I just dont  
> know if the ammount of data generated by this could be handled over  
> an extensive ammount of time. Ask the people that wrote Google  
> Desktop Search.
>
> 7) its good if you could be able to input data from the web and  
> other documents, in a snap, with a key combination. Lets say you  
> are in a web page and you find that interesting fact you just cant  
> miss. you select the text, use a key combo (lets say ALT+Z), and  
> the organizer imports that bit of info, giving you the chance to  
> tag it on the fly.
>
> -----
>
> Concluding, my ideal PIM has the following  UI elements:
>
> 1)  no outline pane, instead of it, on the left side of a screen, a  
> "thin" panel act as an index of tags, from where you can select  
> tags (one or multiple) to filter the chunks of data.
> 2) Below the menubar and toolbar of the app, a "tagbar", from where  
> you can see what tags are active (being used to filter data on a  
> given moment). Instead of using the left pane to select the active  
> tags, you can type directly the desired tag on the tagbar, and the  
> application would suggest the nearest tag to what you are typing.  
> As you do this, the above-mentioned tag-index left-panel scrolls  
> automatically to the suggested tag enabling you to see other  
> similar tags.
> 3) On the rightmost portion of the "tagbar", a chronological  
> filter. This would enable you to filter only chunks of information  
> that were recently edited.
> 4) Taking about 75%-80% of the screen, the data visualizer. If no  
> filter is on, and no tags are enforced, you see all the chunks of  
> information (of course that would not fit on the screen, but thats  
> what scrollbars are for) you have inputed, sorted chronologically,  
> the most recent on top. A thin line separates each and every piece,  
> and on the top of each one, you can see the tags attributed to that  
> piece.
>
> Well, thats it, i 'm just selling this idea because i think  it  
> would be great to have such a software.  I'd be glad if that   
> input  would be of any use on chandler or by OSAF. Keep up the  
> great  work on  opensource!!!
>
> - Paulo Diniz
>