vanderwal.net

XML for org charts and so much more

Thanks to Anil, I came across An open toolkit to facilitate knowledge extraction and document analysis, which contains a solid approach to org charts. I know quite a few folks that sing the "Org Chart Blues" on a daily basis. This is a solid step in the right direction. This document is the work of Gareth Hughes

Philip Greenspun also provides his Software Engineering of Innovative Web Services course materials online. The Problem set 4 is a wonderful section that covers metadata and its uses. As the overview states:

Teach students the virtues of metadata. More specifically, they learn how to formally represent the requirements of a Web service and then build a computer program to generate the computer programs that implement that service.

I not only found the Edison and the Big Thing, but the New York Times now offers signing up for narrowed news trackers. These e-mail alerts are set to keywords that have a corrolation to the article. This seems to be a nice easy step for user to set the alerts via e-mail. It would be interesting to know how well this service is used and received.

Metaphor of Attraction

Beginning with a discussion with Stewart on Peterme and the encouragement of Lane in another discussion to look for a metaphor other than navigation that could better explain what we do on the Web. Seeing Stewart walk by at SXSW after I had seen some of Josh Davis visual plays I combined the discussion with Stewart with the magnetic attraction Josh showed, which began my thinking about a metaphor of attraction. Magnetism seems like what happens when we put a search term in Google, it attracts information that is draw to the term on to your screen.

Come see where else this metaphor can go in this poorly written for draft of the metaphor of attraction. This is posted to begin a collaboration to dig back and move forward, if that is where this is to go. The writing will improve and the ideas will jell into a better presentation over the next few weeks.

InfoWorld has an interview with the executive director of the National Center for Supercomputing Applications (NCSA), Dan Reed discussing the computing grid. This is somewhat like having an electric grid but the commodity is data. The grid's biggest hurdle is interacting with differing data types for essentially similar data. This he explains is bridged using metadata. The metadata needs are enormous. Understanding the information and data is paramount to the task and getting the system right.

I have added two pages to help provide a guide for metadata usage. One page sorts categories by number of times the metadata definition has been used. The other page is an alphabetical listing of categories with their count. These two page builds took very little time to knock together (half an hour or so) and the value to me is much greater than that half hour used.

Okay, O'Reilly Net has been offering a lot of good resources of late and I am a little behind in catching up.

Those of us that need a light weight database for a small project and are using Microsoft Windows often turn to Access to perform the task. Steven Roman offers his tips in how to set up an Access database. The best tip is right up front:

Don't touch your PC until you have taken the time to rough out a design of your database.

The tips keep coming and many of them apply to any other database development. Once I learned to think of information and metadata in a relational database format (which also helps with setting up XML documents) application development became easier. Understanding a little database design will also help ask the right questions when setting up an application, database, and/or project.

This article also helps define the limitations of Access databases. Each database will have its own limitations or peculiarities. Knowing these differences will help think about the application, information, and how they can and will be used is helpful.

John Udell looks in to perl to create topic maps and bottom up taxonomy. Making taxonomy and topic maps easier is a great endeavor and quite useful.

DM Review provides a history of metadata and the future of metadata in one article, what more could you want. It is a good overview of the past tools and goals of metadata. Central to providing a solid information application is understanding the data and information we are working with. This understanding allows us to tie is to other information to give that data better depth and this is best done with metadata. Metadata help explains the data and allows for data to be tied (or joined) to other data.

An example would be putting together a quick view of all products sold in an ice cream truck. The driver has a small computer that captures all the sales he has makes throughout the day. At certain intervals he synchs his sales to a central computer so that he can easily stop back at the warehouse and stock up on items he is low on. The driver has sold 7 of ab4532 and 16 of tr78t in the past hour. He has had requests for pp887 too. This information is useful for the computers and the data has meaning, but it needs some metadata to know that he will need 7 of Abners Glow-cicles, 16 Tricky Rootbeer pops, a carton of covered paper cups. This metadata will help the person that will prepare these items for the driver's arrival at the warehouse. The "pp" indicates that a paper product is needed and the number following is its actual product code. The raw data now has meaning with tanks to the metadata.

IBM explains RDF in this wonderful introduction. [hat tip Nick at Digital Web New]

Content management is back at the forefront of every aspect of my digital life again. Content management revolves around keeping information current, accurate, and reusable (there are many more elements, but these cut to the core of many issues). Maintaining Websites and providing information resources on the broader Internet have revolved around static Web pages or information stored in MS Word, PDF files, etc. Content management has been a painful task of keeping this information current and accurate across all these various input and output platforms. This brings us to content management systems (CMS).

As I pointed to earlier, there are good resources for getting and understanding CMS and how our roles change when we implement a CMS. Important to understanding is the separation of content (data and information), from the presentation (layout and style), and from the application (PDF, Web page, MS Word document, etc.). This requires an input mechanism, usually a form that captures the information and places it in is data/information store, which may be a database, XML document, or a combination of these. This also provides for a workflow process that involved proofing and editing the information along with versioning the information.

Key to the CMS is separation of content, which means there needs to be a way to be a method of keeping links aside from the input flow. Mark Baker provides a great article, What Does Your Content Management System Call This Guy about how to handle links. Links are an element that separates the CMS-lite tools (Blogger, Movable Type, etc.) from more robust CMS (other elements of difference are more expansive workflow, metadata capturing, and content type handling (images, PDF, etc. and their related metadata needs)). Links in many older systems, often used for newspaper and magazine publications (New York Times and San Francisco Chronicle) placed their links outside of the body of the article. The external linking provided an easy method of providing link management that helps ensure there are no broken links (if an external site changes the location (URL) it there really should only be one place that we have to modify that link, searching every page looking for links to replace). The method in the Baker article outlines how many current systems provide this same service, which is similar to Wiki Wiki's approach. The Baker outlined method also will benefit greatly from all of the Information Architecture work you have done to capture classifications of information and metadata types (IA is a needed and required part of nearly every development process).

What this gets us is content that we can easily output to a Web site in HTML/XHTML in a template that meets all accessibility requirements, ensures quality assurance has been performed, and provides a consistent presentation of information. The same information can be output in a more simple presentation template for handheld devices (AvantGo for example) or WML for WAP. The same information can be provided in an XML document, such as RSS, which provides others access to information more easily. The same information can be output to a template that is stored in PDF that is then sent to a printer to output in a newsletter or the PDF distributed for the users to print out on their own. The technologies for information presentation are ever changing and CMS allows us to easily keep up with these changes and output the information in the "latest and greatest", while still being able to provide information to those using older technologies.

Seach Not and Find the Answer

Peter Morville explains why search doesn't suck, but is just not great. I completely agree. Search by itself misses much of the information, unless the site is well written (which provides a cohesive use of terms) or is augmented with metadata.

Let me explain, as Doug Kaye uses in his quest to find what is wrong with searching, a person six months or more ago could have been writing about IT as the possible wave of the future. More recently the same person could have been writing about Ginger. This past week the writer would have started writing about Segway. All were the Dean Kamen invention, but a user searching for a the breadth of our writing on Segway could easily miss our mention of IT or Ginger. The user would have to know to search on these other terms, if they did not they may not find our work. We loose.

This is where metadata helps out. If the information is tagged with a term that classifies this information or could have synonymous relationships established from that metadata item (personal powered transportation = IT, Ginger, Segway...) would greatly help the search. Most of us have been worked on projects that have had searches yet we constantly had users asking us were our information on "xyz" could be found, as they did not find it in the search and they know they read it on our site. That is a large persistent problem. Searching is not a solution only a patch that leaks.

By the way taxonomies can be fluid, they have to be as usage changes.

Triggered by Kendall Grant Clarks review of Scrollkeper on O'Reilly Net I finally got back to the Open Source Metadata Framework. I was following some of the early development on this as they were looking at embracing Dublin Core. The result is a nice malleable format that can be wrapped around many ideas. It is an essential component of Content Management metadata. Now matter how you twist CM these elements should be there in one form or another.

Peter Morville posts The Speed of Information Architecture cajoling IAs to slow down. He brings into play Stuart Brands ideas of slow and fast layers in society, which Peter does a nice job translating into slow and fast layers of IA.

Matt points out a Web-based bookmarking tool b.. This can be collaborative in nature if you should so choose.