Back to dissertation / research

 

2.2. Social web and new mass media: Web 2.0 features

 

What is a vertical portal?

The role of vertical portals is to gather people and information belonging to the same community.

 

But there are already communities on blogs and on social networking sites?!

Blogs ease information production and SNSs (social networking services) maintains links between people but a community still need his home, his own channel with specific services and tools.

 

A wine community and a basket ball community are sharing common tools on internet but not all of them. The goal of our reflexion here is to identify what are these common needs.

 

2.2.1. Web is social: many-to-many

 

2.2.1.1. The blog phenomenon

 

Web is different from old medias by the nature of the information flow. Old medias have channels containing shows which gives information to people. The new media has always channels but now everyone is free to produce in it and to deliver the information. Nowadays some people have their own information channel, most common are blogs (text) but they are sometimes vlogs (video). These private channels are closely connected to their community, exchanging thoughts through comments and sometimes publishing the best comment as an article. Blog is becoming the first tool for producing content on the web because it's simple and thus used by the crowd (Sifry 2007).

 

In the same way, any member of a community should have a producing tool in order to deliver information with the rest of his community (functional concept n°1).

Blogs are popular because anything can be posted simply whenever you want, maybe your post will be popular, maybe it won't.  And if your post makes some buzz, it means that you have just produced something interesting for your community. A lot of people often said that you had to produce one topic per day on your blog, otherwise people won't come on it anymore, I have some concerns with this idea. If your blog is specific to a community (e.g. Blog about your hobby: wind surfing) and aggregated by the community's portal (e.g. French wind surfers), then your article will be promoted by portal mechanisms (using tags and quality markers) and maybe it will become popular, known by peers and recognised (Wikipedia Virtual Community) (functional concept n°2). People won't come on your blog every day to check if a new interesting article has been posted because the portal will do it for them, portal is used here as an aggregator. All information is now able to move from the end of the tail to the head (Anderson 2006:119).

 

By the word blog, you don't have to think about a place where someone talks about his day-to-day life. Here we are focusing on the format; you have to imagine community's blog as a unique producing tool where all the information produced by someone is referenced. When implementing this functionnality into a vertical portal it might be a good idea to find a new name for it. In fact the word "Blog" is now too much linked to a personal journal. If we keep going on the french wind surfing portal, we could name it something like "Waves logbook". We will keep user behaviours, using a blog presentation but we will integrate attributes which are specific to this community. Indeed the information produced should use a specialized semantic. This will help the portal engine to classify the information produced. Shared folksonomies (tags) could be a good solution because it is widely democratized in the blogosphere (Sifry 2007) and these tags could be enhanced with more semantic (Spivack 2006:17).

If you want to recover more semantic from a message you can also create templates for specific information types. Wikipedia already uses templates for a lot of different topics (Wikipedia Infobox). In our wind surfing case, the "Waves logbook" functionality could add a specific message type called "I found a new beach". If the user decides to use this template, then the portal will suggest a number of information to integrate in his message like the location, type of waves, wind, etc... Semantic would be added to each peace of information using RDFa and a beach ontology (Spivack 2006:16) (technical concept n°1).

The portal engine is now able to understand this semantic and we can maintain a list of all the beaches referenced by our users.

 

2.2.1.2. Comments, because everyone has an opinion on everything

 

We were treating about information spontaneously produced by community's members but this information can also be a reaction to another article, video or any other piece of information.

Conversations are starting around pretty anything on the web, so community's members have to be able to react everywhere. These reaction have to be ranked as well and using a proper format.  Discussions have to be in live, using instant messaging and recording conversations for future users. People participating to a discussion could know when a new response has been added (technical concept n°2). (e.g. a new discussion service has just opened up and responds perfectly to that problematic www.tangler.com).

Because comments are often as important as the original message commented, the portal should automatically add an entry on the user producing tool (expressed upper). Comments are too often downgraded compared to the initial information whereas people are proud about their comments. I've already seen a lot of time someone making a comment to a blog article directly on his own blog and then simply post a link like: "Hey check out my comment on my blog: http://ww....". Because he knows that his information is more valuable as a message than as a comment.

 

2.2.1.3. When one-to-'a few' is replacing one-to-many

 

Vertical portals have to provide tools in order to let the community living and growing by itself but they can also provide one-to-many information flow.

We will keep our French wind surfing portal example. If that portal have been created by the institutional French federation of wind surfing then this institution may would like to inform all the users about a new regulation or a new law concerning wind surfing. This example shows exactly the edge of our reflexion on common functionalities for vertical portal; all these one-to-many functional will have to be developed beside the library.

Where the library can still be useful is how to implement one-to-many and many-to-many information into the portal design. We will develop this thought later in this dissertation.

 

2.2.4. Web is social: we create profiles

 

Profiles are used on every web site. They allow authentification and then identification. User's profile is linked to one unique ID in the website database; it makes possible several functionalities around the user identity: functionalities customization, link between information produced and authors, relationship between profiles, etc...

The main problem nowadays is that profile is already created on every web site with more or less details. Even relationships between profiles have to be recreated.

The issue here is to share user profiles in order to ease authentification and first steps of the registering processes. Nowadays some solutions are growing up but aren't widely used yet.

The two well known systems are OpenID for user authentification and FOAF files for handling profiles. OpenID provides a way to be authentificated only one time on all the web sites you will visit. You will just have to give your OpenID URL which will be used to authenticate you on the current web site.

 

Concerning your basic information, like age, gender or friend list, semantic web can provide an easy way to share a common configuration file for all the sites. This profile configuration file could be FOAF. In this RDF file you will find your basic description and a list of your friends.

The creation of a unique identifier on internet is critical; it could avoid identity violations (e.g. fake posts) and it will increase the efficiency of researches with more semantic. Nowadays it is very difficult to perform a research on articles written by a specific author because of the lack of semantic in the page.

OpenID+FOAF+RDFa provide a quite good solution to that problematic and there are already systems managing both authentification and remote profile (e.g. videntity.org storing remote profiles and the PeopleAggregator framework for managing them).

Every portal should implement OpenID authentification and import/export FOAF files (technical concept n°3).

 

(We will discuss later how the semantic web can be generally implemented in vertical portals).

 

2.2.5. Web is social: profiles are linked

 

A vertical portal should of course implement social networking functionalities. Objective here is not to recreate a full SNS but to create an architecture allowing people to maintain their social links. This should include at least:

 

2.2.5.1. Private and public messages/conversations

 

Public and private messages can become live conversations because they use instant notification thanks to Ajax (see 2.2.1.2.). Public messages are displayed on the user's profile, this system already exists on MySpace (comments list) and Facebook (The Wall) but these messages can't turn in a live conversation (Functional concept n°3).

 

2.2.5.2. Friend of a friend links

 

It's important to let the user builds his own network thanks to invitation mechanisms.

Thanks to FOAF, not only profiles could be shared between different systems but also his list of friends and even groups.

 

2.2.5.3. Groups and sub-communities

 

Danah Boyd writes:

 

In 1980s and 1990s researchers argued that the Internet would make race, class, gender, etc. extinct. There was a huge assumption that geography and language would no longer matter, that social organization would be based on some higher function. Guess what? When the masses adopted social media, they replicated the same social structures present in the offline world. Hell, take a look at how people from India are organizing themselves by caste on Orkut. Nothing gets erased because it's all connected to the offline bodies that are heavily regulated on a daily basis.

(Boyd 2006)

 

A vertical portal is created for one specific community and the portal has to let the community manages its sub-communities (Functional concept n°4) .Specific functionalities could be allowed to sub-communities like a private board and tools for sharing various contents (videos, photos, widgets), etc ...

 

2.2.6. Mass media means multi-media

 

Because internet is replacing traditional newspapers, radio and TV, vertical portals should attempt to broadcast all these different kind of media, text, audio and video (functional concept n°5).

The first type of content is textual. Text is everywhere on internet, so it's quite easy to understand why newspapers felt the web's wave first (Fadner 2005).

 

Vertical portals should also broadcast audio, maybe creating a user driven radio linked to the vertical portal? (e.g. Last.fm)

Anyway, many softwares allow to generate an audio version of an article. These audio versions are really useful for people doing something else at the same time (e.g. driving) or people with disabilities. It can be even possible to imagine reading articles on the portal radio.

 

Of course every vertical portal should include video. The main problem around videos on internet is accessibility for people with disabilities. Special versions for deaf are already available on TV, why not on internet? Videos have also a lack of semantic. It's very difficult to know what a video exactly contains, despite folksonomies which describe the global subject, we don't know exactly when is the information you are looking for in the video. Indeed it is a technical problem; most of the video on the web are running under a proprietary technology: Flash. Flash hasn't been created for broadcasting videos but for creating vector and raster graphics (Malone 2007). SMIL, a Synchronized Multimedia Integration Language, seams to be the best alternative (W3C 2007a) but unfortunately a lot of people are not concerned about accessibility and non-proprietary issues  (Axistive 2007) (e.g. Firefox won't implement last versions of SVG and SMIL before the end of 2008, they are currently focused on security issues). Thus we will have to wait 2008 for a full implementation of SMIL 3.0.

Considering this major issue, nowadays it is very difficult to state on a specific technology for broadcasting video+audio+text. I still consider SMIL implementation as a priority and so video handling in PortalLib will be built on two different layers: one uploading and storing videos without any format modification and one for the broadcasting (technical concept n°5).

 

What about a richer media? People are more and more talking about richer interfaces, with technologies like Silverlight (Microsoft), Flex (Adobe), full 3D interfaces, etc... But what is exactly the benefit except decreasing accessibility to zero, spreading proprietary technologies on the web and finally shackling the semantic web? War continues, first victims are users (see 2.1.2.).

 

2.2.7. Mass media means multi-devices

 

We just said that vertical portals use multiple forms of content at the same time. You can find on the same page, text, embedded videos and audios and this rich content might be difficult to render on very different devices. We will focus here on mobile devices because we are entering in a new era of mobile devices. Indeed the next generation of mobile devices are merging all the existing mobile services. Nowadays you can find mp3 player, photo, video recorder, video player, radio, internet access and all powered with more and more storage capacity, processing power, larger bandwith and wider screens. So I don't think we should create specific web technologies for mobile devices; we don't need anymore old mobile technologies like Wap or Imode, we just have to work on the user experience, surfing on a vertical portal using his mobile device.

W3C standards hopefully let us separate content and form. Content is stored with a markup language like XHTML, RDF or any XML based technology while the form is designed thanks to stylesheets (CSS).

 

So do we just have to create a new CSS stylesheet for mobile devices? I don't think so.

In fact mobile devices won't replace our 17 inches screen, our large keypad and our lovely mouse. User needs are different because he is on the move, input device is worse than his home keypad, his bandwidth isn't such good and his screen even with a good resolution has not been made for reading content on web sites.

I think that vertical portals and even all web services should provide a specific version of their systems for mobile devices (e.g. The excellent: mobile.yahoo.com).

A vertical portal should summarize its services into a user-friendly way designed specifically for mobile devices and maybe creating specific functionalities (technical concept n°4).

It appears relevant to focus on how we can ease information access while information production will be more reserved for the 17 inches interface.

 

2.2.8. Vertical portals are open

 

In his well known paper 'What Is Web 2.0', Tim O'Reilly wrote:

 

Much as the rise of proprietary software led to the Free Sowftware movement, we expect the rise of proprietary databases to result in a Free Data movement within the next decade. One can see early signs of this countervailing trend in open data projects such as Wikipedia, the Creative Commons, and in software projects like Greasemonkey, which allow users to take control of how data is displayed on their computer.

(O'Reilly 2005:3)

 

In fact current social network sites are not social because all the information is bound to their technical edge. MySpace and Facebook aren't really pushing RSS. MySpace limits his site to specific HTML based widgets while Facebook creates a proprietary widget technology. None of them of course share his network through the web semantic but Facebook has at least an API (but only for embedded applications...).

 

The team who has written a framework for web science says 'On the Web, the relevant information is likely to be highly distributed and dynamic, personalisation is expected to be one of the big gains of the Semantic Web, which is pre-eminently a structure that allows reasoning over multiple and distributed data sources' (A framework for web science 2006:47).

In order to be distributed and dynamic a vertical portal should be open, that is why a vertical portal has to implement RSS (RDF Site Summary) everywhere an information is likely to be aggregated (technical concept n°6). Web services can also provide a good way to share processed information. Creating widgets for your vertical portal is also a critical action because widgets are a piece of your site included in another portal (functional concept n°6). Widgets are now everywhere: on client side (e.g. windows vista), on aggregators (e.g. netvibes), on mobile devices (e.g. iphone) and even on other portals (e.g. MySpace). Widgets have to be developed as a mini-site, so you should identify services that could be embedded externally. Our wind surfing portal could create various customized widgets like "last wind surfing photos", "wind on the beach", etc...

Vertical portal is a community's home, when you are moving in a new home, you are coming with your stuff and then you personalise the place. So vertical portals have to be highly customizable, enabling widget inclusions, multiple skins and more generally a flexible user interface.

 

Semantic web is also "an attempt to bring together data across the Web so as to create a vast database transcending its components, which makes possible applications that infer across heterogeneous data" (A framework for web science 2006:19).

The first rule when you create a database is to avoid duplicated information. Because the PortalLib project is about common tools for vertical portals, these tools could have a common database thanks to a web more semantic (technical concept n°7).

• Profiles are authenticated outside thanks to OpenID.
• Profiles and networks can be imported/exported thanks to FOAF
• Videos could be embedded in another site (we are waiting for SMIL).
• Photos can be imported/exported (e.g. FlickR)
• RDF and ontologies should be created and shared for verticals information.

 

2.2.9. User interface

 

The user interface has always been a tricky issue. How to increase usability? How to include ads? What level of user customization?

PortalLib is a library which doesn't provide any graphical template, so webmasters are free to implement components as they wish.

In all cases user-friendliness and ads depend of the nature of your vertical portal and your business model, so we won't argue here.

Nevertheless customization is a critical point (A framework for web science 2006, Anderson 2006:169) and there are many different ways to enhance customization.

The first thing is to allow users to choose a graphical themes (e.g. MySpace, Netvibes), because a portal should be your place, you can decorate it (CSS eases that process) (Functional concept n°7)(Technical concept n°8).

Moreover these services could also be removed and added; the best way to do that is to provide customizable areas in the portal (e.g. Facebook).

These customizable areas can contain widgets; it means pretty anything, and is of course based on the W3C work (W3C 2007b) (Functional concept n°8)(Technical concept n°9).

Like we said before vertical portals have to create widgets for people using portal services on others devices (e.g. iphone) or others portals (e.g. netvibes). Widgets allow to broadcast outside the portal edges but also to let users customize them inside. It means that our vertical portal is using his own widgets which they are draggable, customizable or even removable.

Maybe it's too obscure for now; we will try to clarify it by an example.

Our wind surfing portal is composed of a network layer, different kinds of production tools for text, video and photos and maybe an online newspaper written by the community.

Different people are using the portal for multiple reasons and by multiple ways, someone will be attracted by only reading and sometime commenting so he will move up all widgets containing pure content: "Last articles of the e-newspaper", "Best videos", "Best photos", "Best user articles" and he will move down or remove the following widgets "Your wind surf community", "Most popular surfers", "Beach weather forecast", etc ... This user could also add external widgets from various sites (e.g. emails, agenda, RSS, etc...) and so create an highly customized portal around his favourite activity: "wind surfing".

Of course each webmaster would be free to decide how far he would allow customization of his interface; we can guess ads won't be removable ;).

 

2.2.10. Accessibility

 

Because I've realised that a lot of people do not know what accessibility is, the following definition may be useful: 'Accessibility is a general term used to describe the degree to which a system is usable by as many people as possible. In other words, it is the degree of ease with which it is possible to reach a certain location from other locations. It is not to be confused with usability which is used to describe how easily an entity (e.g., device, service, environment) can be used by any type of user... focuses on people with disabilities and their right of access to entities, often through use of assistive devices such as screen-reading web browsers or wheelchairs' (Wikipedia Accessbility).

Any web developer should remember that 20% of people in Europe/America have disabilities. Population world-wide is aging, 20% over 60 now, 30% by 2020 and 70% of people over 60 have a disability. It is true, most people over 60 don't even know internet, but because they have not lived with it, it is not the same case for youngest generations.

Moreover accessibility isn't only a problem for disabled but also for motorists, mobile workers, people in noisy environments and all people who needs hands free.

Remember that 20% of Europeans have disabilities (90M) and only 40% are employed, telework on web apps could be a huge benefit for these people.

We have already said that a vertical portal represents a specific community composed of several smaller communities, disabled are one of them (technical concept n°10). We can guess that in our wind surfing portal more than 10% are permanent or temporary disabled, we can't simply forget them.

 

Every component of the PortalLib project will be specifically analysed in order to evaluate his accessibility level. We will follow the WCAG 2.0 (Web Content Accessibility Guidelines) and we will use the conformance level defined by the W3C for each component.

 

2.2.11. Trust

 

Trust is involved at different levels.

The simple question "is this information reliable?" involves multiple subconscious questions like "what are my past experiences on this site?", "do I know the author?", "who recommend me this information?". Vertical portals are mostly considered as niches so you are supposed to find a good specialized content. This is true when you are coming from an upper layer. Indeed the web can be divided by different layers of niches where each niche belong to an upper niche, the long tail phenomenon also traduce this concept explaining that each tail is founded by multiple smaller tails (Anderson 2006:139).

Let's make an example around trust... Imagine you are looking for a new wind surfing board. So you make a research on your favourite search engine and you finally land on a personal blog. You read an article comparing the two last boards of a 14 years old kid and you notice that other messages on the blog are dealing with mangas.

We can assume that in a vertical portal dealing with wind surfing, more trusted information concerning boards quality and prices should be found.

 

So you can imagine that trust is a very good lever for buying online and that's why niches are very interesting for advertisers, it is a matter of trust. Online ads services are displaying random ads on your site based on a poor semantic analyse of the content. Creating a vertical portal means you know well your community and his needs, the community should be able to find what kind of products or services are relevant to suggest. You don't need to create a new e-business site because you can customize your own shop (e.g. Zlio.com).

What do you think about an online shop specialized in wind surf items, with thousand of comments and people talking about boards all the day?

 

2.2.12. Business model

 

In the last IAB report, Randall Rothenberg, President and CEO, said:

 

Interactive advertising revenues continue to show solid growth as advertisers and agencies recognize that it is a medium that can uniquely impact consumer behavior from product awareness, to purchase intent, to actual purchase and then brand loyalty. We have every confidence that this growth trend will continue as marketers allocate more of their total marketing dollars to interactive and the industry delivers effective and innovative platforms for connecting with consumers.

(IAB 2007)

 

We will keep two thoughts. The first one is that the web "can uniquely impact consumer behaviour from product awareness". Indeed Ads revenues are increasing by 33% each year and it would represent more than twenty billions dollars in 2007 (IAB 2007).

The second interesting point is that advertisers are looking for "innovative platforms for connecting with consumers" and these innovative platforms are vertical portals. The Long Tail also reaches advertisers world, you can find ads for anything.

So how can vertical portals attract advertisers? The definition we've actually given of vertical portals provides different tools for advertising. We have included more meaning thanks to RDF and OWL, so the machine is able to understand the content of the page and so include ads highly relevant. We have also advice to create vertical shops for sharing experience and comments around vertical products.

With this 33% growth rate, we can expect that more and more small niches will be soon attractive for entrepreneurs. Especially when advertisers realize they are targeting the wrong community in a too wide horizontal portal (BBC News 2007).

The advertising around wind surfing is mature? I'm not sure, but if it isn't, it won't be the case for a long time.

 

2.2.13. Web 2.0 technologies

 

Here we will talk about the famous web 2.0 technologies. How can they be used in our vertical portal concept?

 

2.2.13.1. Ajax

 

Ajax the well known, like we mentioned earlier, isn't such new, technologically speaking. It's a very interesting mix of different technologies and has allowed the creation of very famous and useful tools like googlemap, netvibes, etc...

But Ajax must not be used everywhere, you should implement it carefully or it will vanish your site's accessibility. Ajax mustn't be critical in the user path, any task can be achieved without Ajax.

 

2.2.13.2. Semantic Web

 

Actually the Semantic Web is more know by his folksonomies (tags) than the real W3C Semantic Web (RDF+OWL).

We believe strongly that the semantic web will be the next big step in the web history and vertical portals have their role to play.

Indeed who knows better about a subject than its community itself?

On this purpose Nova Spivack says:

 

Where might we see this content initially arising? In my opinion it will most likely be within vertical communities of interest, communities of practice, and communities of purpose. Within such communities there is a need to create a common body of knowledge and to make that knowledge more accessible, connected and useful.

(Spivack 2006:18)

 

So, Semantic Web should be implemented as far as possible in vertical portals, creating vertical ontologies and spreading his content through the web.

Shifting from the old web to the semantic web won't be easy and fast. Anyway, some tools might help this transition. Promising technologies like RDFa or GRDDL should be widely used in the next years.

But what is the real benefit for vertical portals?

In fact a simple index of subject areas may not provide the community with sufficient ability to search for the content that its members require. To allow more intelligent syndication, vertical portals can define an ontology for the community.

 

2.2.13.3. RSS

 

RSS is a good transitional technology for sharing content. It's not obvious that RDF Site Summary will be replaced by an arising semantic web. I think people will always need summaries.

Nevertheless RSS will continue to be transformed, containing more and more semantic.

 

2.2.13.4. Widget

 

As written above, widgets have also a promising future. It's a good way to embed portal's functionality in another device or another portal.

 

2.2.13.5. Web services

 

Web services should also find his way in a web more and more opened. Web services can provide an higher level of security and they are still useful to resolve complicated requests.

 

2.2.13.6. RIA

 

I do not see RIAs as the future of user interfaces but they could be used for games or for highly interactive applications (I think about 3D visiting).

However RIAs still have a serious problem concerning accessibility and they shouldn't be used for diffusing information.

 

2.2.14. Other information relative to vertical portals

 

2.2.14.1. Security note

 

Working on security is of course a critical issue for any vertical portal but PortalLib won't support the whole security of your site because it's a library which is already based on another layer: Prado.

Webmasters have to check as often as possible last versions of Prado and PortalLib in order to ensure the higher level of security possible.

A system is never 100% secure but PortalLib will do his best.

 

2.2.14.2. Environment

 

Because the hearth of the planet is critically linked to our technological run, we advise developers to have an ecological policy (Spivack 2006:24-26).

Vertical portals help building communities, and these communities will grow as long as the web grows. So your servers will need extensible capacities.

If you don't want to waste energy you should take a look to utility computing services (e.g. Amazone EC2, Flexiscale).