vanderwal.net

Feed On This

The "My" portal hype died for all but a few central "MyX" portals, like my.yahoo. Two to three years ago "My" was hot and everybody and their brother spent a ton of money building a personal portal to their site. Many newspapers had their own news portals, such as the my.washingtonpost.com and others. Building this personalization was expensive and there were very few takers. Companies fell down this same rabbit hole offering a personalized view to their sites and so some degree this made sense and to a for a few companies this works well for their paying customers. Many large organizations have moved in this direction with their corporate intranets, which does work rather well.

Where Do Personalization Portals Work Well

The places where personalization works points where information aggregation makes sense. The my.yahoo's work because it is the one place for a person to do their one-stop information aggregation. People that use personalized portals often have one for work and one for Personal life. People using personalized portals are used because they provide one place to look for information they need.

The corporate Intranet one place having one centralized portal works well. These interfaces to a centralized resource that has information each of the people wants according to their needs and desires can be found to be very helpful. Having more than one portal often leads to quick failure as their is no centralized point that is easy to work from to get to what is desired. The user uses these tools as part of their Personal InfoCloud, which has information aggregated as they need it and it is categorized and labeled in a manner that is easiest for them to understand (some organizations use portals as a means of enculturation the users to the common vocabulary that is desired for use in the organization - this top-down approach can work over time, but also leads to users not finding what they need). People in organizations often want information about the organization's changes, employee information, calendars, discussion areas, etc. to be easily found.

Think of personalized portals as very large umbrellas. If you can think of logical umbrellas above your organization then you probably are in the wrong place to build a personalized portal and your time and effort will be far better spent providing information in a format that can be easily used in a portal or information aggregator. Sites like the Washington Post's personalized portal did not last because of the cost's to keep the software running and the relatively small group of users that wanted or used that service. Was the Post wrong to move in this direction? No, not at the time, but now that there is an abundance of lesson's learned in this area it would be extremely foolish to move in this direction.

You ask about Amazon? Amazon does an incredible job at providing personalization, but like your local stores that is part of their customer service. In San Francisco I used to frequent a video store near my house on Arguello. I loved that neighborhood video store because the owner knew me and my preferences and off the top of his head he remembered what I had rented and what would be a great suggestion for me. The store was still set up for me to use just like it was for those that were not regulars, but he provided a wonderful service for me, which kept me from going to the large chains that recorded everything about me, but offered no service that helped me enjoy their offerings. Amazon does a similar thing and it does it behind the scenes as part of what it does.

How does Amazon differ from a personalized portal? Aggregation of the information. A personalized portal aggregates what you want and that is its main purpose. Amazon allows its information to be aggregated using its API. Amazon's goal is to help you buy from them. A personalized portal has as its goal to provide one-stop information access. Yes, my.yahoo does have advertising, but its goal is to aggregate information in an interface helps the users find out the information they want easily.

Should government agencies provide personalized portals? It makes the most sense to provide this at the government-wide level. Similar to First.gov a portal that allows tracking of government info would be very helpful. Why not the agency level? Cost and effort! If you believe in government running efficiently it makes sense to centralize a service such as a personalized portal. The U.S. Federal Government has very strong restriction on privacy, which greatly limits the login for a personalized service. The U.S. Government's e-gov initiatives could be other places to provide these services as their is information aggregation at these points also. The downside is having many login names and password to remember to get to the various aggregation points, which is one of the large downfalls of the MyX players of the past few years.

What Should We Provide

The best solution for many is to provide information that can be aggregated. The centralized personalized portals have been moving toward allowing the inclusion of any syndicated information feed. Yahoo has been moving in this direction for some time and in its new beta version of my.yahoo that was released in the past week it allows the users to select the feeds they would like in their portal, even from non-Yahoo resources. In the new my.yahoo any information that has a feed can be pulled into that information aggregator. Many of us have been doing this for some time with RSS Feeds and it has greatly changed the way we consume information, but making information consumption fore efficient.

There are at least three layers in this syndication model. The first is the information syndication layer, where information (or its abstraction and related metadata) are put into a feed. These feeds can then be aggregated with other feeds (similar to what del.icio.us provides (del.icio.us also provides a social software and sharing tool that can be helpful to share out personal tagged information and aggregations based on this bottom-up categorization (folksonomy). The next layer is the information aggregator or personalized portals, which is where people consume the information and choose whether they want to follow the links in the syndication to get more information.

There is little need to provide another personalized portal, but there is great need for information syndication. Just as people have learned with internet search, the information has to be structured properly. The model of information consumption relies on the information being found. Today information is often found through search and information aggregators and these trends seem to be the foundation of information use of tomorrow.

Gordon Rugg and the Verifier Method

In the current Wired Magazine an article on Gordon Rugg - Scientific Method Man (yes, it is the same Gordon Rugg of card sorting notoriety). The article focuses on his solving the Voynich manuscript, actually deciphering it as a hoax. How he goes about solving the manuscript is what really has me intrigued.

Rugg uses a method he has been developing, called the verifier approach, which develops a means critical examination using:

The verifier method boils down to seven steps: 1) amass knowledge of a discipline through interviews and reading; 2) determine whether critical expertise has yet to be applied in the field; 3) look for bias and mistakenly held assumptions in the research; 4) analyze jargon to uncover differing definitions of key terms; 5) check for classic mistakes using human-error tools; 6) follow the errors as they ripple through underlying assumptions; 7) suggest new avenues for research that emerge from steps one through six.

One area that Rugg has used this has been solving cross-discipline terminology problems leading to communication difficulties. He also found that pattern-matching is often used to solve problems or diagnose illness, but a more thorough inquiry may have found a more exact cause, which leads to a better solution and better cure.

Can the verifier method be applied to web development? Information Architecture? Maybe, but the depth of knowledge and experience is still rather shallow, but getting better every day. Much of the confounding issues in getting to optimal solutions is the cross discipline backgrounds as well as the splintered communities that "focus" on claimed distinct areas that have no definite boundaries and even have extensive cross over. Where does HCI end and Usability Engineering begin? Information Architecture, Information Design, Interaction Design, etc. begin and end. There is a lot of "big umbrella" talk from all the groups as well as those that desire smaller distinct roles for their niche. There is a lot of cross-pollination across these roles and fields as they all are needed in part to get to a good solution for the products they work on.

One thing seems sure, I want to know much more about the verifier method. It seems like understanding the criteria better for the verifier method will help frame a language of criticism and cross-boundary peer review for development and design.

Quick Links in the Side Bar is not Optimal

Paul wants to "set up one of those link-sidebar thingies again" for his quick link list. Actually I am finding those the side link lists, like mine cause problems for folks tracking referrer links back and for search engines. Context of the links is helpful, but so is being able to find the date and page where these links came from. The way Paul is doing his quick links now works well. I was able to point directly to these links, the links he make have context, even if it is only a list of links.

Quite similar to the Fixing Permalink to Mean Something post the other day, the links in the side bar are temporary. I find links from Technorati back to my site from some poor soul looking for what comment and link vanderwal.net had placed. These links do not have a permalink as they are ever rotating. I have received a few e-mails asking where the link was from and if I was spamming in some way.

Why do I have the quick links? I don't have the time to do a full or even short write-up. I clear my tabbed browser windows and put the items I have not read in full in the Quick Links. Some things I want access to from my mobile device or work to read the info in full or make use of the information. Other things I want to keep track of and include in a write-up.

The other advantage of moving the quick links into the main content area is they would be easier to include in one aggregated feed. I know I can join my current feeds, but I like the sites that provide the feeds in the same context as they appear on the site as it eases the ability to find the information. This change will take a more than a five or ten minute fix for my site, but it is on my to do list.

Chevy Redesigns with Standards

Chevrolet has redesigned with fully valid (one minor issue in the style sheet) XHTML (strict) and CSS. It is beautiful and wonderfully functional. All the information can be easily copied and pasted to help the discerning car buyer build their own crib sheet. The left navigation (browsing structure) is wonderful and not a silly image, but a definition list that is expandable. The style layer is semantic, which is a great help also (for those IAs who understand). Those of you so inclined, take a look under the hood as there are many good things there.

Fixing Permalink to Mean Something

This has been a very busy week and this weekend it continues with the same. But, I took two minutes to see if I could solve a tiny problem bugging me. I get links to the main blog, Off the Top, from outside search engines and aggregators (Technorati, etc.) that are referencing content in specific entries, but not all of those entries live on the ever-changing blog home page. All of the entries had the same link to their permanant location. The dumb thing was every link to their permanant home was named the same damn thing, "permalink". Google and other search engines use the information in the link name to give value to the page being linked to. Did I help the cause? No.

So now every permanent link states "permalink for: incert entry title". I am hoping this will help solve the problem. I will modify the other pages most likely next week sometime (it is only a two minute fix) as I am toast.

You Down with Folksonomy?

Gene supplies a good overview of Folksonomy, which is the bottom-up social classification that takes place on Flickr, del.icio.us, etc. It would be great to have a tool that could help organizations develop a folksonomy over time. Gmail from Google could develop a great folksonomy that could be overlaid on ones own searches.

Marry this idea with Paul Ford's "Google beats Amazon and eBay at Semantic Web" and you have a wonderful jump on the Semantic Web that is personalized. Take Gene's idea of building a thesaurus or crosswalk of terms within and across systems and things can really take off. There would need to be contextual tools added to handle the multiple definitions like Macintosh is a synonym for Mac, but Macintosh is an artist and a computer, while Mac is a computer, artist, and British term for raincoat (short for Macintosh). Hence the Semantic Web adds context to get these things straight.

UXnet Aims to Unite the Splinters

Having trouble figuring what group will help you in your carreer as as a web designer that keeps information architecture, usability, interaction design, experience design, etc. in your toolbelt?

It seems there is a group that has come togther to help be the glue and bring all of these splintered groups together. UXnet aims to be the glue that draws the groups together. Many designers and UX/IA/ExD/Etc folks are lost in finding one good home and one or two good conferences. There are many resources, too many is what much of these designers and researchers say. Many of us wear many hats and need a good cross pollination to get better.

I have hope that UXnet will help close the chasm that keeps everybody apart. There are representatives from many groups as a part of the team pulling things together.

Profiled at InfoDesign

I am the current InfoDesign Profile - Thomas Vander Wal. This was harder than I thought it would be an many alternate answers ran through my mind, but I finally narrowed it down as much as I could. Peter has many other wonderful profiles and interviews at InfoDesign Special. I have been inspired and found new resources from these glimpses into other designers lives.

Gmail Simplifies Email

Since I have been playing with Gmail I have been greatly enjoying the greatly improved means of labeling and archiving of e-mail as opposed to throwing them in folders. Many e-mails are hard to singularly classify with one label that folders force us to use. The ability to drive the sorting of e-mail by label that allows the e-mail to sit accessibly under a filter named with the label make things much easier. An e-mail discussing CSS, XHTML, and IA for two different projects now can be easily accessed under a filter for each of these five attributes.

Dan Brown has written a wonderful article The Information Architecture of Email that dig a little deeper. Dan ponders if users will adopt the changed interface. Hearing many user frustrations with e-mail buried in their Outlook or other e-mail application, I think the improved interface may draw quite a bit of interest. As Apple is going this way for its file structure in Tiger (the next OS upgrade) with Spotlight it seems Gmail is a peak at the future and a good means to start thinking about easier to find information that the use can actually manage.

Web Standards and IA Process Married

Nate Koechley posts his WebVision 2004 presentation on Web Standards and IA. This flat out rocks as it echos what I have been doing and refining for the last three years or more. The development team at work has been using this nearly exclusively for about couple years now on redesigns and new designs. This process makes things very easy to draft in simple wireframe. Then move to functional wireframes with named content objects in the CSS as well as clickable. The next step is building the visual presentation with colors and images.

This process has eased the lack of content problem (no content no site no matter how pretty one thinks it is) often held up by "more purple and make it bigger" contingents. This practice has cut down development and design time in more than half and greatly decreases maintenance time. One of the best attributes is the decreased documentation time as using the Web Developer Extension toolbar in Firefox exposes the class and id attributes that provide semantic structure (among many other things this great tool provides). When the structure is exposed documentation becomes a breeze. I can not think of how or why we ever did anything differently.

Amazon Plog

Amazon is offering a "Plog" (personalized weblog) of offerings and order information as my front page to their site. I have a link to an order and offerings, which tell me what I rated or ordered in the past to get the offering.

I sort of like this front page as it has the info I am interested in, particularly why I am recommended a product and order info. I am not a fan of the "Plog" moniker. It is too much trying to "be" something, which it is not. Now if they could not return Dummies books when I search for DVDs or CDs.

Make My Link the P-link

Simon hit on plinks as an echo to Tim Bray's comments and variation on Purple Numbers (Purple Numbers as a reference). As I have mentioned before, page numbers fail us and these steps are a good means to move forward.

Simom has also posted in more plinks and in there points to Chris Dent's Big Day for Purple Numbers.

I have been thinking for quite some time about using an id attribute in each paragraph tag that includes the site permalink as well as the paragraph with in that entry. This would look like: <p id="1224p7">. This signifies permanent entry 1224 and paragraph 7 with in that entry. What I had not sorted out was an unobtrusive means of displaying this. I am now thinking about Simon's javascript as a means of doing this. The identifier and plink would be generated by PHP for the paragraph tag, which would be scraped by the javascript to generate the plink.

The downside I see is only making edits at the end of the entry using the "Update" method of providing edits and editorial comments. The other downside is the JavaScript is not usable on all mobile devices, nor was the speed of scrolling down Simon's page that fluid in Safari on my TiBook with 16MB of video RAM.

Why Content Managment Fails

Adaptive Path's Jeff Veen explains Why Content Management Fails. It comes down to a people problem in his book, which I agree with.

It also comes down to poor initial analysis, poor product choice based on the initial analysis, poor implementation, and trying to solve a people and process problem with technology, which often just compounds the problem.

Also take a look at Peter's comments on Enterprise Content Management. Peter is Jeff's partner and has some great insights that I have experienced also. The framing the issue as a technology problem is one of the common failures and difficulties I have run into in the past seven years dealing with CMS. It did not take me long to figure out it is an information problem, process, and mostly a people problem. I seem to continually deal with people that do not understand the variables in the equation.

In my current role I am always witnessing managers on the client side wanting the glitzy and having little and&047;or poor quality content. Just as a content management technology will not solve content generation problems or turn your ragged tabby cat into a beautiful tiger, having a beautiful site will not solve the lack of good content. Hiring technologists to solve information and people problems is pouring money down a hole. The approach to the problems will not discover the problems as the right questions have not been asked, the right discovery methods have not been used, the right analysis has not been done, the right deliverables are not produced, which does not lead to success.

Google is not my only search engine

Google has been letting me down lately. The past two months I have had too many irrelevant links or only a handful (when I narrow the terms) that do not have what I am looking for. Oddly I have Googled only my site and found the results where I mentioned what I was seeking.

I have been turning more and more to Vivisimo and DogPile for search instead. Why? Well they are both metaseach tools, Vivismo includes Google in what it searches, that search across multiple search engines and return them in one interface. These two services also have faceted filtering and/or categorical filters for the results. These facets greatly help filter out the junk. In short it solves the Paris Hilton site problem when you want a hotel room not a bimbo.

In the past I have tried Vivismo, but it did not seem to have enough depth, which has now been solved. Dogpile now offers a good breadth of search engines that seem to improve on the limited results I had been getting in the past. It is good to have options.

Tools to Manage Information On Your Personal Hard Drive

I have posted my thoughts on Tools to Manage Information On Your Personal Hard Drive for Mac OS X in particular. I have posted this on my Personal Info Cloud site. This is the first piece of content that I am not posting in both places. This may become a trend as I am spending a fair amount of time thinking through ideas related to the Personal Info Cloud in one place. The Personal Info Cloud has an RSS feed and I will be posting notices that new info has been added there as it happens.