root navigation content

Delavy for the Web

Bilingual thinking to improve the Web

Version FrançaiseEnglish Version

Bookmarks Classification

samedi 3 janvier 2009 [21:27:38]

I propose a new bookmarks classification system, to exit the impractical arborescent storage. I ask "why we bookmark ?" and make a sort system with 3 category 01:

  • Data
  • Tools
  • Websites

Cet article existe aussi en version française.

I take as a prerequisite that whit the awesome bar, we bookmark a website to remember it, but also for faster access. I always remember the url and the name of my university website, but with the awesome bar (firefox address bar who allow to access on bookmarks), I ask for "fac" or "univ" and go ! So we bookmark website on which we regularly going to have easy access.


We wand to keep a picture, a video that we like. I called it favs but Wei Zhou ideas, data seems more appropriate :

  • Theme : we recall the theme because we search it : a "humoristic" picture or video.
  • Type : we recall the theme because we search it : a "video" of skiing
  • Name : we don't always remember the name of the film/game/...
  • Url: we never remember the url : it's too vague (a video on dailymotion) or too complicated (too long adress)

=> Classification : Classification by type : picture / video / text / flash. And associate it tags concerning the contain, with a maximum of tags

Actually 02, a sub-folder by type remains effective. Remember to tag with the type too, to make search like "humor video".


We often bookmark a page/website that we use, on which we do an action/search on content : search engine, translators, dictionaries, search engines, directories, encyclopedias, validators, ....

  • Theme : we recall exactly the theme because he was the tool
  • Type : there is no type
  • Name : we don't always remember the name
  • Url : we don't always remember the url

=> Classification : It must tagging the website with its function. Then, I type "translator" and just have to chose.

In the actual system, it was difficult to make folders : some groupings are possible (language tools, irl , ...) but it quickly becomes very chaotic. Notice that we can use the keyword system to quickly make request.

Websites 03

It is the remaining of the bookmarks, content generating site (most of the time), where we go to read content : mozilla labs, cnn, ... There may also be simple pages, but i don't have ideas for the name so I call it "websites". This sites are note classifiable by frequency, and hardly by theme : I think the folders classification to limiter : as we evolve, we change everything the classify, and we make folders by themes and many theme go with a same site (and many sub-folder as different folders).

  • Theme : the theme evolve with user, and many theme are for a same site.
  • Type : type often in combination (mozlabs : pictures, texts and videos)
  • Name : we always remember the name
  • Url : we often remember the url

To better understand, I compare bookmarks with music. The music tracks are easily sortable even if they are a millions, because attached data where fixed : this track make always 2'03" and was always Reggae and the artiste was always Bob Marley. We don't choose "music tags", so we remember it, and they can't change ! But imagine a bookmarks on the artist previously cited : When we search it, we don't know if it was in the folder music ? artist ? singer ? reggae ? The solution is in the tags : we can tag it "musician" and "artist" and "reggae", .... A research of one of this term found the page.

Tagshifter example

Menu display thank to Tagsifter, allowing a research by tag on the bookmarks

I propose a tags navigations like with tagsifter (see picture below). If I search Songbird for example, I don't know if I should search in Music/Software or in Software/Music. In a folders browsing, it make a big difference ! In a tag browsing, no : if my website is tag "music" and "software", I ask one of the two and so Firefox show me the associated tags. Always in take the music tracks comparison : at everyone, Marley where in the gender Reggae. But for a website, everyone use at as he want, and nothing is universal. That's why I think that the user must spend time in tagging Bookmarks.

=> Classification : "folders" of tags like tagsifter

In the actual system, tagsifter is very efficient. If the bookmarks are correctly tags, with a little request form address bar, with 2 or 3 keyword we find they. So in 2 or 3 clicks on tagsifter we find what we seek, with several different "ways" (so we make fewer mistakes and find faster).

The goal is to have a maximum of fixed data, i.e. data who didn't depend of the user, which facilitates research : you know on which category the page you are looking for, there is (almost) no error possible.

Notes and references:

Texts, pictures, and design : Copyright © 2008 - 2009 Delaby Pierre.
Copyleft : This is a free work, you can copy, distribute, and modify it under the terms of the Free Art License. You find a copy of this license on the site and other sites.
Icons by Mark James in Creative Commons | All Credits and Attributions.

Logz Powered : | Host by | Copyleft Licence Art Libre | Vrs ?, Vxl ? | site map