The latest release delivers better control over tagging, custom names for tags, defining new tags, Save API and OpenID support.

We are happy to announce the addition of several new features. The purpose of the new features is mainly to facilitate the use of common tags from Wikipedia, as well as to overcome Wikipedia’s limitations as a controlled vocabulary for semantic tags.

Tagging emerged as an extremely popular way to integrate and organize data, due to its simplicity and flexibility. However, free-word tags do not have defined meanings, so it isn’t always clear what a particular tag represents. Does the tag “jaguar” represent the animal, the car company, or the operating system?

On the other hand, common, “semantic” tags are unique, well-defined concepts that allow people to state what a web page is exactly about. Semantic tags come at a price, though. They reintroduce structure, the absence of which was the main reason why tagging has become so popular.

The question is: Is it possible to make semantic tags as flexible as classic ones? Can humans accept and love the format intended for machines? Today’s release is Faviki’s attempt to answer this challenge.

Features in this release include:

Enhanced tagging interface

Universal Wikipedia tags are often too long and too hard to enter and the exact name of a tag has to be known beforehand. Furthermore, tags are personal items – a private association to some concept. They are often based on emotions, for instance: the nickname “Pippo” instead of the full name of the soccer player “Filippo Inzaghi”.

The new release makes it possible to use custom names for tags. Tags are added in free form, resembling classic tagging. If Faviki doesn’t understand a tag provided by a user, it will ask her to disambiguate it. It will then remember her choice and, next time, it will know what she means.

Faviki “learns” about user’s name of the tag

Faviki “learns” about user’s name of the tag

This is possible by connecting the idea of tagging with the idea of searching. Tags are used as keywords for a Google search that is restricted to Wikipedia’s domain. After all, tags and keywords are subjective associations to unique concepts and Google search is a great way to find URLs that represent these concepts. This way, users can use keywords as custom tags for Wikipedia URLs.

In addition, custom names for tags can also be modified explicitly on the Tag page.

Defining new tags

Wikipedia is the world’s largest encyclopedia, but it still covers only a small portion of the real world. There is a large number of concepts that are either too specialized or do not possess sufficient “notability” to be included in a common encyclopedia.

We already take for granted that every company or organization has a URL and that most people we know have some kind of web page, a blog, a social network profile or a company page that represents them online.

Faviki exploited this fact in one of its new features – defining new tags. New tags are added the same way as Wikipedia tags. The difference is that, this time, Google search is not restricted to Wikipedia’s domain, but only a few of the top results are allowed to be selected. Google returns web pages from the whole Web and users collaboratively create new tags and decide which URLs are the best candidates for new concepts.

Users collaboratively decide the best URLs for a concept

Users collaboratively decide the best URLs for a concept

Save/Edit API

The Faviki Save/Edit API is a simple API that provides a way to save and edit bookmarks from other applications.

OpenID support

Faviki finally supports OpenID. It uses RPX, a service which integrates various OpenID implementations from Google, Yahoo, AOL, Microsoft, along with plain OpenID.

Other features/improvements

  • Smarter autocomplete list
    The autocomplete list is an alternative way of finding and adding tags. It is now powered by DBpedia lookup – a powerful search API for Wikipedia concepts.
  • Converting tags
    This feature allows users to convert any of their tags to another tag across all of their bookmarks.
  • Spam control
    Bookmarks of no value for users can easily be marked as spam. Bookmarks that were marked as spam by a certain number of users are hidden.
  • Export/backup bookmarks
    Bookmarks can be exported along with semantic tags in the standard HTML bookmarks format.
  • Tag description tooltip
    A short abstract with an image, if there is any, shown when a mouse is held over a tag name, helps users choose the right tag. The data is fetched from DBpedia in real time.

Thanks to all of our users who have given us the feedback regarding the new features on Faviki. Stay tuned, further information will be released on the blog soon!

Reblog this post [with Zemanta]

We’re excited to announce that Faviki has officialy started to use Zemanta API to suggest possible Wikipedia concepts.

Zemanta is a platform for accelerating on-line content production, by recognizing contextual content and instantly serving relevant images, smart links, keywords and text to the user. Recently they launched an early release of the API and allowed web developers to use Zemanta engine in their applications. The API allows developers to use simple RESTful interface to get suggested images, articles, tags and links in a structured format (XML, JSON, ..) for a given piece of text.

The fact that it suggests Wikipedia links, among others, was particularly interesting to me, so I tested it right away and figured out that it works very well. Now Faviki users don’t have to start from the stretch, because tag suggestions are given through an analysis of the web pages’ content. If a suggestion is OK user just needs to click on the ‘+’ button and the tag is added.

However, do not expect it to do the entire job for you – you are the one who makes the final decision!

We also started using Zemanta on our blog (it can be deployed on all major content publishing platforms). It’s easy to use and it saves time, so we highly recommend it. Great job, Zemanta!

Many thanks to Andraž Tori for the support.

Reblog this post [with Zemanta]

What is it?

You probably noticed the ‘G’ button on the right hand side of the field for adding new tags, and of the ‘tags’ field in the search. That is the Google search button that we wrote about on our Help page here. However, we thought that this feature deserves its own post on the blog, because it helped us with finding tags many, many times.

How does it work?

With Google search button, you can search for tags as you would search Wikipedia pages on Google. For instance, if you type in ‘apple’, and click on the Google button, the system will automatically add ‘wikipedia’, so your query will actually be ‘apple wikipedia’, and search result will be retrieved from the domain en.wikipedia.org only.

Faviki google search api button

Experience showed us that this way of finding tags can be quite helpful and time saving. Sometimes it is hard to find the most appropriate tag with autocomplete list, and Google is pretty clever when it comes to finding the most popular/representative tag for an acronym or ambiguous term, for instance. So, it is often the case that the tag that you are looking for is at the top of the list. To add it just click on the ‘copy’ link.

Cases in which it beats the autocomplete list

  • Acronyms and their disambiguation:
    • EU = European Union
    • RHCP = Red Hot Chili Peppers
    • CSS = Cascading Style Sheets, Content Scramble System, Cansei de Ser Sexy
    • LCD = Liquid crystal display, Lacida, Lowest common denominator
    • SEO = Search engine optimization, Seasoned equity offering
    • RDF = Resource Description Framework, Robotech Defense Force, Radical Dance Faction
    • REST = Representational State Transfer
  • Ambiguous terms:
    • apple (fruit, digital technology corporation, Fiona Apple, bank…)
    • keyboard (computers, music, magazine…)
    • office (software, place where you work, series, film…)
    • flash (software, superhero, photography, song…)
  • Searching for the right term for the concept:
    • programming = Computer programming;
    • baby = Infant;
    • tiredness = Fatigue (medical);
    • moonlight sonata = Piano Sonata No. 14 (Beethoven);
    • rachmaninov = Sergei Rachmaninoff. (Note that in this case the term is not even spelled correctly)
  • When you know what you think of, but you don’t know/can’t remember how to name it:
    • belarus capital = Minsk
    • eu lead body = European Council
    • kaiser chiefs singer = Ricky Wilson (British musician)
  • If you wish to search for related tags or tags concerning a broad topic:
    • online social (Social network, Social software, Online identity, OpenSocial, Virtual community, Social bookmarking, Social computing)
    • vegetarian (Vegetarianism, Vegetarian cuisine, Vegetarian Society, World Vegetarian Day, Veganism)
    • olympic games (Olympic Games, Summer Olympic Games, Winter Olympic Games, Ancient Olympic Games, Youth Olympic Games)
  • If the tag contains non-English characters, and you don’t want to deal with them:
    • roisin murphy = Róisín Murphy
    • motorhead = Motörhead

Drawbacks

  • It is slightly different than autocomplete list, e.g. you have to click on the ‘copy’ link instead of on the tag name (which is a link to a Wikipedia page)
  • Search results list will also contain some Wikipedia pages which are not tags, like pages whose names start with ‘Special:’, ‘Template:’, ‘User:’, ‘Wikipedia:’, ‘Help:’, ‘User talk:’, ‘Wikipedia talk:’, ‘Category:’. These are special Wikipedia pages and obviously cannot be used for tags, so you cannot add them.

We hope we’ll be able to fix these issues soon.

Summary

Inserting correct tags is essential for Faviki in order to use its potentials to the maximum. But finding the right tag is sometimes a bit tricky. We hope that Google search API can make your tagging easier and more accurate.

Follow

Get every new post delivered to your Inbox.