November 2007

Knowledge management and Social networking and The Web23 Nov 2007 at 11:04 by Jean-Marc Liotier

I set up a link blog and a collaborative bookmarking site for our tiny geek community. My friends have initially been slightly confused by the conceptual similarities. So here are a few general guidelines to provide a clearer distinction of use cases.

Both tools are relevant for posting links with no significant value added by the poster. If there is value added by the poster in the way of analysis, context, story telling or anything else, a traditional blog entry is a better choice.

A social bookmarking tool must focus on resources that the user might want to come back to in the future, or that he thinks that his friends might be interested in one day. The accent is on easy recall through various means of discovery such as search, feed reading and folksonomic exploration.

By contrast, a link blog focuses on immediate sharing. It is the place to show off the spectacular, the anecdotic, the exceptional – novelty items that you want to share with your friends but whose future recall value for practical use might be low.

The motive for link blogging is not just altruistic : posting in a link blog is also a way to elicit reactions to the content you discovered. And that is why the community gathered around your link blog is important : you want to gather contributions from the people that matter to you. And if you have enough feedback, then there might just be enough new material to warrant more synthetic capitalization in a proper blog article.

As you can see, although the niches of social bookmarking and link blogging in knowledge management do overlap a little, they are definitely distinct and educating the users in extracting the highest value from them is worth the effort.

Email21 Nov 2007 at 14:13 by Jean-Marc Liotier

Yes, we do hate stupid email disclaimers that much : they are a completely meaningless waste of electrons, they waste display real estate and they make proper quoting more awkward. And most of all they are unilateral clauses whose acceptation is most dubious.

Lerouge dug out an article by Russel Coker about striking back at worthless email disclaimers by configuring Postfix’s smtpd_banner in /etc/postfix/ with the appropriate legalese of its own. Lerouge suggested that we do likewise and I jumped on the case and implemented it immediately…

Looking for something really outrageous and fit to point out how ridiculous those things are, I found inspiration at and modified it for my purposes…

So there you go :

13:52 jim@kivu ~% telnet localhost 25
Connected to localhost.
Escape character is ‘^]’.
220 ESMTP (M Sexchange) – READ CAREFULLY. By transmitting email to this server you agree personally and on behalf of your employer or organization, to release the recipients from all obligations and waivers arising from any and all NON-NEGOTIATED agreements, licenses, terms-of-service, shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure, non-compete and acceptable use policies (hereafter BOGUS AGREEMENTS) that the recipients have entered into with your employer, its partners, licensors, agents and assigns, in perpetuity, without prejudice to the recipients ongoing rights and privileges. You further represent that you have the authority to release the recipients from any BOGUS AGREEMENTS on behalf of your employer or organization.

So in case you wondered, here is proof : the humongous disclaimer from hell does fit into smtpd_banner’s limits !

Design and Identity management and Knowledge management and Social networking and The Web20 Nov 2007 at 6:47 by Jean-Marc Liotier

Open is everything – the rest is details. That is why we must take the best use cases of the closed social networking world and port them in the open. This is a lofty goal in all meaning of the adjective, but a surprisingly large number of potential basic components are available to cut the way short.

Friend of a Friend (FOAF) enables the creation of a machine-readable ontology describing persons, their activities and their relations to other people and objects. This concept is a child of the semantic web school of thought that has its origins about as far ago as the Web itself. In a narrower but deeper way, XFN (XHTML Friends Network) enables web authors to indicate their relationships to people simply by adding attributes to hyperlinks.

Microformats such as hCard, xfn, rel-tag, hCalendar, hReview, xFolk, hResume, hListing, citation, media-info and others provide a foundation for normalizing the information sharing. Some major operators are starting to get it – for example my LinkedIn profile contains hCard and hResume data. If you like hresume, take a look at DOAC while you are at it !

Some code is already available to process that available information. For example, identity-matcher is a Rails plugin to match identities and import social network graphs across any site supporting the appropriate Microformats. This code extracted from the codebase of and this is probably how Dopplr now supports import from other social networks like Twitter.

But part of the appeal of a social networking platform is how it empowers the user with control of what information he makes available, how it makes it available and to whom. So microformats are not sufficient : a permission management and access control system is necessary, and that requires an authentication mechanism. That naturally takes us to OpenID.

OpenID is a decentralized single sign-on system. Using OpenID-enabled sites, web users do not need to remember traditional authentication tokens such as username and password. Instead, they only need to be previously registered on a website with an “identity provider”. OpenID solves the authentication problem without relying on any centralized website to confirm digital identity.

The OpenID project is going even further than just authentication – authentication is just the surface. What OpenID really is about is digital identity management. OpenID Attribute Exchange is an OpenID service extension for exchanging identity information between endpoints. Although the list of attributes included in the OpenID Attribute Exchange schema does not match a nice collection of microformats, a process is defined to submit new attributes. And anyway, such a standard looks like a great fit to cover the need for keeping the user in control of his own content.

Finally, the social graph is the support for applications that must interact with the user’s information wherever it is hosted. That is why Google’s OpenSocial specification proposes a common set of API for social applications across multiple websites.

So a few technologies for social networking do exist, and they seem able to provide building blocks for an open distributed social networking. The concept of open distributed social networking itself has been in people’s mind for a long time. But until now only large proprietary platforms have succeeded in seducing a critical mass of users. Thanks to them, there is now a large body of information about the best practices and use-cases. What is now necessary is to think about how those use-cases can be ported into a decentralized open environment.

Porting a closed single provider system into an open distributed environment while equaling or surpassing the quality of the user experience is a huge challenge. But social networking and digital identity management are such critical activities in people’s life that the momentum behind opening them may soon be as large as the one that led Internet pioneers to break down the walls between networks.

Brain dump and Email and Jabber and Social networking and The Web19 Nov 2007 at 10:43 by Jean-Marc Liotier

Most social tools come and go. The ones that stay share a common feature : openness. For example, email is open : chose any technology, chose any provider or be your own provider, chose any client, any platform – any way you do it you are still connected to the whole world.

If you have the slightest understanding of your own interest, then there is no way you should even consider using a closed platform as your primary mean of communication. Why would you willingly chose to put your most critical asset outside of your control ?

Many users will object that they gladly surrender control to closed social networking platforms because plain email does not meet their sophisticated communications needs and they are not willing to invest in developing the skills currently required to participated efficiently in the blogging sphere. That is a tragedy because the social graph is quickly becoming the glue of the connected services.

And even if the functionality was sufficient, we would still have a huge mindshare gap to bridge. XMPP provides nice basic instant messenging and presence management in an open environment, but most users still prefer proprietary centralized networks and happily trade freedom for webcam compatibility.

But similar battles have been fought and won in the past : Compuserve, AOL, The Source, Prodigy and their ilk have all dissolved in the Internet. The forces of openness now have a new crusade to embark upon : we must take the best use cases of the closed social networking world and port them in the open !

Open is everything – the rest is details. That is what drew me to the Internet fifteen years ago.

Design and The Web10 Nov 2007 at 16:48 by Jean-Marc Liotier

Because I live and breath computing and the Internet, I often forget what the casual users experience. So sometimes I watch one at random over his shoulder just to keep abreast of what the Myspace generation and the typewriter generation are doing. And every time my battle-hardened sysadmin heart still shudders at the sights.

These days, my favorite casual user habit is searching for an obvious URL using a search engine. For example I have seen “yahoo mail” typed as an argument in the Google search form – no once, but several times and by different users ! I would have thought that or or or whatever other variations that have been setup by Yahoo provide enough obvious ways to reach the service. But apparently they don’t.

If even the most obvious of all are not typed, we can infer that today’s casual user does not memorize any URL anymore. Maybe the clean URL that we strive to produce are intended for the sole consumption of search engines and power users. And that’s one more reason why portals are so important.