|
 |
 |
|
| |
 |
Thursday, November 28, 2002 |
Dave Winer est loin d'être le dernier venu... Il vient pourtant mettre un bémol à un concept que je crois pourtant essentiel, le Web sémantique. Encore Seb, le moteur de recherche, qui ramène cette perle :
The truth about RDF. "I am a disbeliever of anything that requires as much documentation, head-scratching, hand-waving, and eyes-glazing-over as RDF does in 2002. Forget the problems with the formats, that can be dealt with later, after you figure out how to explain it to someone who knows a lot about computers, networks, users, XML, HTTP, etc. If you can't explain it to me so that I understand what you're doing -- you've got a big problem.
It's a cute, and all-too-common tactic to say that people who don't get it are dumb. I'm not dumb, but RDF makes me feel that way. After all these years, I've concluded that if I can't understand it, it doesn't have much of a chance in the market. All the powerfully successful technologies of the past have had simple explanations anyone could understand. If RDF is one of those, I strongly believe it must too. Therefore I conclude that it isn't. |
11:33:58 PM 
|
|
d'influencer notre manière d'amasser, organiser et partager la connaissance. Seb rapporte que :
Jim McGee has a discussion of four kinds of knowledge we might want to share: useful answers, useful questions, practices, and generative patterns of knowledge. (via AI)
Un article dense qui mérite d'être analysé.
11:23:57 PM
|
|
Gilles réfère à cet excellent document de fond sur le Web sémantique. J'en ai extrait des trésors d'information dont voici quelques extraits. Les concepts de base d'abord :
At its most basic level, the Semantic Web is a collection of information about information, also known as metadata (...)
The mechanics of adding semantic content is accomplished through a standard framework for describing information called Resource Description Framework (RDF). This framework allows developers to link the "words" in your website, documents, and databases to decentralized vocabularies which define the meaning of those words. These decentralized vocabularies are the heart of the Semantic Web and are frequently referred to as ontologies. (...)
(...) anyone can extend or write their own ontology for whatever words they wish to describe. By writing your own ontologies, you help the computer understand the information on the web more closely to the way you do. Ontologies are commonly written in RDF schema (RDFS) but a new ontology language (OWL) is in the process of being standardized.
Le passage suivant a d'ailleurs suscité mon hilarité : It has been suggested that there exists a classic chicken and egg situation with the Semantic Web; i.e. without applications using semantics no one will build semantic content and conversely without content no one will build applications. Ça correspond bien à un débat qui, sans pour autant faire rage, en occupe tout de même certains d'entre nous... Plus loin, l'article de semaview suggère une progression du Web sémantique, progression qui, à son deuxième stade, me rappelle étrangement les réflexions de Mitchel Resnick sur l'ordre émergeant de « systèmes massifs parallèles » :
Stage 1: Semantic Islands
The Semantic Web is currently in Stage 1, in which islands of semantic content and applications are being developed within corporate intranets. (...)
Stage 2: The Islands Connect
(...) Autonomous software programs, also known as Agents, will be deployed across the Internet to make use of the vast collection of semantic content. Agents will seek out specific information or web services on behalf of users without any interaction with humans.
Stage 3: Artificial Intelligence
(...) Stage 3 is many years away and no one is certain how it will unfold. (...) AI researchers may be able to use the data to provide some true artificial inference abilities.
Parmi les bénéfices suggérés pour illustrer la puissance des données annotées et associées à des ontologies décentralisées, l'exemple suivant rejoint justement mes intérêts :
Intranet Content Management. (...) Typical content management systems attempt to place all data within a few master taxonomies. Problems arise as each employee brings their own unique view and a general taxonomy isn't often equivalent to this personal perspective. A semantic intranet would have all documents, web pages and data annotated, and linked to company and individual specific ontologies to define a personal view. Relationships between items on the intranet could be visualized to give employee's views of data not previously possible.
Et pour terminer, tout l'aspect de l'échange de données entre applications et entités constitue la puissance véritable du Web sémantique, en termes économiques il s'entend :
The second major set of applications that can use semantics are concerned with the storage and exchange of data. Today's method of storing data in tables or file formats presents a problem when one wishes to exchange data with other software systems.(...)
If a proprietary solution is used to store data, one-to-one conversion software must be built so that other applications can "understand" the data being exchanged. As the number of different devices wishing to communicate increases, the number of conversion programs grows exponentially. XML data storage is, in effect, a proprietary solution because each programmer is able to define their own tags. (...)
The Semantic Web offers a solution by allowing developers to define data storage formats by linking them to decentralized vocabularies so that data can be queried globally. As a result, applications are able to quickly and easily lookup meanings and exchange data. (...)
Rather than use XML or a proprietary data format to store data within a digital device, a company may choose to semantically markup their data format. The semantic data format would allow other vendors to use the standard and even modify and extend it. Digital devices would be able to quickly learn to read new data formats they never encountered before.
Je me rends compte, après avoir digéré tout cela, que le copaing pense exactement comme les auteurs de cet article en ce qui a trait aux solutions ouvertes. Bien vu! D'excellentes références sur la pensée de Tim Berners-Lee, le concepteur du premier fureteur Web et serveur Web, accompagnent cet article de semaview. Seb nous en suggère également d'autres.
10:14:00 PM
|
|
[Gilles en vrac...]
Bien pratiques en effet que ces stratégies. Ce site, Boxes and Arrows est d'ailleurs fort bien fait : simple, clair, agréable. Pas surprenant, la spécialité du groupe d'experts alimentant ce carrefour : l'architecture et le design de l'information!
D'accord avec ce commentaire. Toutefois, pour avoir à exécuter cette tâche d'implantation d'un intranet dans un avenir rapproché, je suis toujours à la recherche d'exemples de contextes moins « corporatifs » (petite entreprise, travailleurs du savoir) et plus progressifs en termes d'incorporer le journal web et ses fonctions parallèles dans l'interface. Trop de ces stratégies d'implantation prennent en effet pour acquis la présence d'un collègue préposé à la mise à niveau de l'intranet et/ou d'un logiciel spécialisé. Qu'en est-il de l'intranet « sur mesure » mis à niveau par les utilisateurs eux-mêmes? Je demeure à l'affût d'un environnement d'intranet qui « respire » avec ses utilisateurs.
4:50:25 PM
|
|
© Copyright 2002 Robert Gregoire.
|
|
|
|
|
|
