Adrien’s old blog
=> moved permanently to blog.adrienjoly.com

Just a thought I wanted to share: have you ever been frustrated of not being able to bookmark a file of your computer on delicious? …to find the URL of a file (e.g. an image) you downloaded, in order to share this URL instead of sending the file? … or event to be able to work seamlessly on any document from any computer, with or without installed software? and I won’t even talk about synchronization of files between people, computers and devices… (is that the actual last version of the document?)

If feel so, you’re just like me! Wouldn’t it be nice that we only work with URLs instead of OS-specific and computer-specific local file paths? That way we could leverage thousands of cutting-edge tools available on the internet, in order to better manage, annotate, visualize, share and thus better browse through information!

Of course putting everything on the internet is scary because we’re not always online, and because some stuff sometimes disappear from the web (e.g. older versions of software which are not free anymore)… But what if we “cache” our web-stored information on our computers instead, and assume that the web-hosted version is the reference of the resources (with an Unique Resource Identifier… URI!). This URI could lead to several hosting spaces, including private storage (e.g. computers, devices, portable hard drives), that would keep synchronized.

I promise that, if such a system exists (and works well of course), I would never have to keep downloaded research papers on my computer any longer!

My homepage needs some freshing up but I have so many ideas (sometimes conflictual) that I don’t know where to start!

Firstly, I must define what I understand by “homepage”: it must be the web page that I will spread around to represent my identity on forum/email signatures, on wikis, etc… So, it must provide some personal introduction and links to several facets of my identity (e.g. the researcher, the musician, the technology enthusiast…) on the web.

I considered using one of my existing public social profiles (netvibes, friendfeed, my lab profile, my wiki, or even this blog) as homepage. But:

• those public profiles are either too specific (representing only one of the facets of my identity) or too messy when too general (e.g. netvibes);
• my homepage is very well referenced on Google, so I’d better leverage that;
• …and I would like to manipulate some code!

So, let’s share some thoughts on what I’d like to make:

• The homepage must be nice looking on a computer and adapted for easy navigation on a mobile device => should I make a WAP version or use some generic HTML instead?
• I would like my page to be machine-understandable, by adding some semantics (RDFa, FOAF…)
• I want to be able to edit quickly and precisely (no WYSIWYG please!) the content of my homepage directly on the web.
• The content must be backed up on every change, so that I can revert a change, either accidental or malicious.
• I want the page to be lightweight and respectful of W3C standards.

Err.. this sounds like a neat semantic wiki, doesn’t it?

And now, some more precise ideas I’ve had:

• In order to have a self-contained homepage for all the facets of my identify, I’m thinking of a unique HTML file with fragment anchors for each facet (e.g. http://joly.adrien.free.fr/#music would only display the “music” part of my homepage). It would be nice to display a menu/tabs to switch from a facet to another, without having to load another page. The nice thing about this solution is that legacy web browsers (including MS Pocket Internet Explorer on Windows Mobile “smartphones”) that do not support modern javascript will still be able to render the full page and leverage anchors.
• I want to include some lifestreaming on the page. For that I could deploy Noserub to federate my feeds, or simply embed a friendfeed gadget. The possibility for visitors to comment lifestream entries directly on my homepage (incl. through Facebook/Google friend connect) is a plus.
• EditArea seems like a great way to make my page editable online.
• Concerning the semantics, I see two possibilities: the simple one is to embed some RDFa and/or microformats directly in the HTML code and to have a separate FOAF file that I would have to maintain separately; the geeky solution is to store the content of my homepage (mostly links, anyway) in a FOAF file (in RDF) and to render it as HTML pages using stylesheet-based transformations (e.g. XSLT). As a geek I obviously prefer the second option! However it is also a quite heavy solution, and I’m not sure that its complexity is worth the result, and it might not comply with all the requirements I expressed above…

Do you have any thoughts to share on this?

This question excites business people, researchers and web enthusiasts… What will Web3.0 be? Gurus say it’s gonna be a set of technologies allowing to get rid of the browser, to work semi-connectedly (sometimes off-line) or it’s gonna be the so-called Semantic Web that W3C have been dreaming about for years, or maybe it’s about moving from the keyboard/mouse/screen paradigm to new human-machine interfaces… I think everyone is being too specific, in my opinion.

Here is my vision of the Web evolution:

• Web1.0: sites that broadcast information to people
• Web2.0: services that allow people to contribute and exchange information
• Web3.0: platform that brings relevant and contextual information to every individual

If you read between the lines, you see different dimensions:

• the information flow: broadcasting -> exchange -> relevant delivery
• the platform: web sites -> services -> platform
• the social aspect: people look for information -> people exchange (blogs, social networks, tagging…) -> every person is informed (personalized and contextual aggregation)
• the relevance: company or service-based sites (brands, forums, chats…) -> convergence to topic-based sites that allow manipulation of information (widget, mashups, RSS feeds…) -> feed of personalized information based on the context of the user
• the technologies: HTML (server-based pages) -> AJAX (rich client interfaces) and RSS feeds (extracting information from websites) -> ??? (A.I., semantic technologies…)
• the interfaces: keyboard/mouse/screen -> support for more interfaces (webcams for video conferencing…) -> multimodal access (adapting to the context)
• the community spaces: forums/irc/chat/im -> social networks, metaverses (second life…) -> augmented reality?

Sorry for posting such a draft, that’s just a bunch of random thoughts I just felt I had to write down!

As you may know, I believe that the future of computing is through the use of Semantic Technologies in order to link knowledge (note the contrast with the words “data” or even “content”) in a way that is more natural for humans but also understandable for machines. Although many enthousiasts (including the W3C) try to push semantic standards to the public to follow the vision of Tim Berners-Lee (well know for having written the foundings of the Semantic Web in Weaving the Web), many detractors think that this vision is naive/idealistic and that it will thus never work. I believe that we are ready to Semanticize the Web, but it will take time and we should not expect the masses to do the hard work.

Concerning the Semantic Web, Mor dared to declare publicly that the “The Semantic Web is Dead“. On his article The Emerging-Semantics Web (”The Semantic Web is Dead”), he claims that, even with Web2.0 in which users annotate data and the use of microformats, we cannot expect the masses to generate Semantic data (or should I say knowledge). As I commented on his article, it’s true that nobody but geeks can write pure Semantic data (using RDF for example, and it would not be perfect anyway). But wouldn’t it be the responsability of the platforms (i.e. websites/webservices that allow users to feed their databases with user data) to enrich the user data with semantics? I explain this idea on Mor’s article, check it out.

On his article “Moving Towards the Semantic Web: Grassroots vs. Ivory Towers“, James Simmons shares a similar belief that Semantizing the Web must be done by developers on their publishing platforms, not by the masses using these platforms. In his explaination, he smartly introduces two categories of Semantic Web believers: “Grassroots” and “Ivory Towers”. The Ivory Towers (we’re mainly thinking of the W3C) make the Semantic Web by building its foundations in a top-down approach. They define standards to store, link, query and represent semantic data on the web but their vision of how the web should be is too idealistic for now, as they are on Tim Berners-Lee’s side. Whereas the Grassroots make the Web more Semantic, following a bottom-up approach. For example, they enrich state-of-art websites using smart tricks like microformats to embed metadata within XHTML pages.

Like James, I believe that the Semantic Web is starting to grow as developers build bridges between the current Web and semantic standards. According to the fact that most content of the web is stored using platforms like WordPress (for blogs), Joomla (for websites) or even Facebook (for social networking), this evolution could become exponential soon if the developers of these popular platforms integrate some semantics in them. It’s a long way to Tim Berners-Lee’s (and other Ivory Towers’) vision for sure, but we’ll get there progressively!

It’s now become a common practice for blogs and web sites in general to provide RSS feeds that notify subscribers when new information is published (or any other update). That way, visitors don’t have to visit their favorites websites regularly (”PULL” fashion), instead they are notified (”PUSH” fashion) by their RSS feed aggregator as long as they subscribed to the corresponding feeds. That way, the aggregator watch for updates of visitor’s favorite websites, and the visitor only has to read the updates using the aggregator.

The problem is that some websites don’t have RSS feeds. Fortunately, there are a few services that will create that missing RSS feed for you by watching a web page regularly, thus notify you when updates occured.

On this page, I propose a quick review of the services I tried.

Don’t you realize that people don’t socialize much in the train anymore? Most people have their own personal entertainment means: they watch movies on their laptop, listen to their ipod, read books… But in the same time, many of them have a profile on myspace or any other social networking website in order to meet new people. Isn’t that a paradox?

In France, I’ve found two initiatives that extend the idea of social networking to train journeys. Idtgv&co is proposed by SNCF (the French railroad company), and Train d’Union is a free and independant alternative. In both of them, passengers are expected to create a personal and/or a professional profile, register the references of the trains they will be travelling on and meet other people sharing the same journeys online before they can eventually meet on the train.

While some people may think that it’s ridiculous to use a website for meeting people whereas you can talk sponteanously with the person next to you, I think the idea is great because you can make new interresting contacts instead of doing stuff on your own during long journeys.

Now, the next step is to be able to browse people’s profiles and exchange messages on line, during the train journey, like you would do on myspace, in order to meet them right away and make your journey much enjoyable.

Hi,

Today I decided to keep track of the programs that I use. In order to make this list versionnable and shared, I’ve made that list as a publicly-visible Wiki page.

Feel free to consult it there:

adrienjoly.schtuff.com - FavoriteSoftware

Facts: With the increasing popularity of the "web 2.0", internet users are invited to get involved with the content proposed on web sites by annotating, commenting, tagging… This way, information is reviewed and linked to other resources, helping people to learn and navigate in a more intuitious way - because these annotations are human.

BUT such annotations are found on specific websites, e.g. YouTube for videos, del.icio.us for bookmarks, Last.fm for music, and so on…

Challenge: What about bringing human annotations to another level by generalizing it to any kind of information and pushing it to the user when relevant, without expecting him to look for it on specialized websites?

Uses cases:

1. Reviewing resources and sharing opinions
2. Warning about resources: "content is not up-to-date" etc…
3. Linking to other human-relevant resources
4. Creating communities: meeting other visitors

• Yakalike, Chatsum and many other services intergrate your favorite browser to chat with other people visiting the same page as you. This is an interresting solution for all of the 4 proposed use cases, but there are some drawbacks:

• the lack of popularity (too few users installed the software) prevents the idea from taking off
• the identification of a "page" is based on its URL, which can be irrelevant (especially with web 2.0 apps?)
• this chat program, as a browser extension, may be too intrusive for permanent use: too much space is wasted, sometimes for nothing

• Blogger Web Comments is another browser extension which notifies the user (by a small icon) when the visited page is commented in blogs, allowing her/him to consult the relevant posts. This approach is less intrusive and relying on blogs, which are commonly used nowadays, but also has its drawbacks:

• this add-on only relies on blogs hosted by Blogger.com, limitation that is not acceptable since it would chunk the information for each blog host
• commenting on a page consists in posting in your blog, and this blog must be hosted by Blogger.com. is a small note worth a post on your blog?

• Del.icio.us and other bookmark sharing communities are services that invite users to submit their bookmarks on their account by tagging them. By sharing them with the community, we can evaluate the popularity of a website, discover new websites related with certain tags and read user reviews. We can also consult the people who tagged a bookmark to interact with them. This approach is very popular and interresting but:

• you have to go on the del.icio.us website in order to browse this information
• tags are not necessarly efficient: their are sometimes subjective (synonyms and interpretations) and culture-dependant (translations), this may result in improving ranking bookmarks which are trivial to tag and forgetting content which is not easily categorizable but valuable.
• like a few people, I like to keep my bookmarks stored locally too, and I don’t know about any possible synchronization between del.icio.us and Firefox. It would be hard anyway since browsers use hierarchical categorization while del.icio.us use tags.

I like the idea of chatting with people concerned with certain websites, when they could represent an interresting community, but I think that this concept should not be integrated in all cases. In the blog approach, I like the fact that the user can be notified if some annotations were identified about the website which is currently visited, this is less intrusive. Unfortunately, it also too specific because most interresting annotations are not worth a post on your personal blog. Nevertheless, del.icio.us, as a bookmark sharing service, is the perfect place to store these annotations. But the current service is sticking to its website too much, forcing users to look for information instead of the information being brought to her/him.

A perfect approach?

I think that annotations should be proposed to the user when browsing websites or consulting results of a search engine. In the case of the search engine, these annotations should have an influence on the ranking of the results, by the way. In the browsing side, an extension should connect to a service like del.icio.us to propose sticky notes, a link to an irc channel (chat), human-entered links to related resources and other ranked annotations about this page. The sticky notes and the proposed chat channel must be elected democratically by user vote, favorizing commonly agreed and concise information. The sticky notes would consist in objective information, like the current status of the the content: telling if it’s not up-to-date for example. Any other annotation would remain in a list in which users can rank them (like in Digg). Adding a bookmark would propose to the user to rank the annotation which best describes the content. Ideally, we’ll find a more efficient way than tags or hierarchies to classify bookmarks both locally in your browser and remotely on the bookmark sharing service on the internet.

An interresting post about two evolvements of taggings: hierarchichal and clustered tagging.

Hello World: Hierarchical Tags