Lasipalatsi

Elements of Fiction Writing from Writer's Digest Books

Tue, 24 Jun 2003 19:43:31 GMT

Conflict, Action and Suspense, William Noble
Characters and Viewpoint, Orson Scott Card
Scene and Structure, Jack M. Bickham
Beginnings, Middles & Ends, Nancy Kress
Voice & Style, Johnny Payne
Setting, Jack M. Bickham
Plot, Ansen Dibell
Description, Monica Wood
Manuscript Submission, Scott Edelstein

Bookmarks -> Blog: Development Environment & Tools

Tue, 15 Apr 2003 03:53:00 GMT

I'm a pack rat. I collect gazillions of bookmarks, articles, copies from books, and PDF files. Today is about clearing out the flotsam of my collection of bookmarks relating to my development environment and tools.

This entry will be a work-in-progress until I've got everything cleaned up and captured here.

Tools & Packages

Mostly to Use in Development

Cooktop XML Editor
Marathon GUI Test Tool
ArgoUML

Mostly to Look at for Ideas

pyblosxom
Vellum
Flock RSS Aggregator in Java
Haystack Universal Information Client (MIT)
LifeMap Concept Mapping &c
eVisioner Rules, Version Control, Persistence Infrastructure, &c.
Blozom Mozilla-based publishing tool
WWWOFFLE World Wide Web Offline Explorer (Proxy)

Configuration Management

Repository

I have to decide between CVS and Stellation on this one...

Build

Ant
Gump - and/or - CruiseControl - and/or - Anthill Build Management Server
JUnit

Project Information/Metrics

Project Status: Defect Tracking / Enhancement Requests

I have to decide between Scarab and Bugzilla on this one...

Languages

Python

Java

Windowing / GUI

wxWindows (wxPython)

IDE / Interface Builders

Know and Love

Eclipse (PyEclipse)

Evaluate

Wing IDE for Python

Libraries: Java

Know and Love

TM4J - Topic Maps for Java
XOM XML Object Model

Evaluate

Libraries: Python

Know and Love

PyXML

Evaluate

Satine for Python - XML Data Binding
Xelf XOM Python Port

Libraries: C/C++

Know and Love

Evaluate

References & Documentation

Sam has a cool comment spell-checker

Sat, 12 Apr 2003 19:30:18 GMT

Go and post a comment on one of Sam's blog entries. Be sure to misspell a word or two. Coolness!

Papers that inform the design of the NewTool platform

Mon, 10 Feb 2003 04:08:34 GMT

Here are several papers (some via Lilia via McGee) that inform my thinking about the NewTool platform:

NewTool will support the knowledge process

Sun, 09 Feb 2003 18:32:00 GMT

Knowledge Work as a Process

Jim McGee says that the goals of most knowledge management efforts today harken back to Taylorism (scientific management) with onerous command and control ideas and language as opposed to an recognition that knowledge is embodied in humans and any "management" of knowledge must take this fact into account.

NewTool acknowledges the human synthesis of information into knowledge and of knowledge into knowledge and tries only to augment human capabilities to allow focus on the process of synthesis and creation:

This is a process that is fundamentally iterative. The loops in this process are feedback loops, not opportunities for streamlining. You don't improve this process by rearranging the steps or breaking them down into specialized tasks to be distributed. Nor are there opportunities to eliminate non-valued added steps. Improving the value of knowledge work calls for different strategies. Two that are worth exploring are to improve the infrastructure at the periphery and to eliminate friction. [Is knowledge work improvable?, McGee's Musings]

Stay Tuned

Thu, 06 Feb 2003 22:03:18 GMT

Whiteboard drawing of NewTool - an open source platform for Free Media and Microcontent + (universal) personal proxy/server + authoring and replication

A sampling of references/influences/thought-starters for those interested in talking about this:

Marc Canter: Marc is all over Free Media Management (née Open Media Management) - Free Media Management, Virtual File mngt vs generic Media Management, Jon hits a home run! (identity issues), Intro to an Open Standards Architecture
Les Orchard of 0xDECAFBAD: PersonalWebProxy
IBM Research: Web Intermediaries (WBI) - "Intermediaries create smart pipes" (they've got a whiteboard picture, too, and also prettier pictures showing the architecture and some scenarios)
Abe Fettig: HEP Message Server
Yutaka Sato: DeleGate: an application-level gateway/proxy with "transcoding" capabilities and origin (client) capabilities - see DeleGate: A Universal Application Level Gateway
The Intelligent Software Agents Lab: WebMate from Carnegie Mellon ("a personal digital assistant, is a promising solution to the problem of finding useful information among a sea of texts and other web documents")
Matt Griffith: Jog, the personal Google & Wayback Machine
P.A.W. @ SourceForge: The Pro-Active Webfilter is a filtering HTTP proxy based on Sun's Brazil framework
Phil Wolff of a klog apart: Comments on microcontent and converged clients in From .blog to converged client
Anil Dash: Introducing the Microcontent Client
Sébastien Paquet: Part 2 of Personal knowledge publishing and its uses in research (here is Part 1)
OSAF: Chandler, of course...
Eastgate Systems Inc.: Storyspace
The SpaceMapper Project: MN8 and DataStore
KnownSpace: KnownSpace Helium
evectors: RssDistiller
Ticluse Teknologi: The Snewp
D. J. Adams: Tinkering with RSS and NNTP (Adams is also the author of Programming Jabber - I've got my copy and Jabber is in the tool...)

Fri, 31 Jan 2003 04:01:39 GMT

Thoughts about Search Engines and Communication. When we do a search on the Internet we are seeking an answer to a question, looking for information. We want to make a query and get an accurate response, free of irrelevant and distracting information. The art of using search engines effectively is the art of framing a question. To frame a question for a search engine we must keep in mind that we are asking a machine for an answer, a machine that does not know the associative meanings that we take for granted when speaking with people. Look at the gif diagram (http://www2.eou.edu/~jhart/netcomdiagram.gif); it represents a simple view of Internet Communication, with just two people and two machines. P to P is person to person communication; M to M is machine to machine communication; P to M is person to machine communication; M to P is machine to person communication. Clearly the ways a machine communicates with another machine will be different than the ways a person communicates with another person. Similarly, a person to machine communication is different from a machine to person communication. The trick in all this to recognize that what works with one relational direction of communication will not necessarily work with another direction. Obviously, if machines communicate with people the way that they communicate with one another, then people will not get the message. If people communicate with machines the way that they communicate with each other, then machines will not get the message. Many of the challenges of the modern age of Internet communication are summarized in the directional lines of this one simple, communications rectangle (except the complexities are multiplied many times over when we add millions of people and millions of machines). As I understand it, the great effort now underway to construct a semantic web is an effort to enable M to M communications to utilize associative meanings; such semantic enhancements could yield emergent communications that are not now available between machines; those semantic communications could then facilitate M to P and P to M communications beyond the levels that are now available. (See http://www.w3.org/2001/sw/;"The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation." -- Tim Berners-Lee, James Hendler, Ora Lassila, The Semantic Web, Scientific American, May 2001.) One of the best ways to reveal the limits of current search engines is to start with the end result, i.e., begin at the end of the search rather the the usual start of the search. Begin with the answer rather than the question. E.g., if you know that you want to find the online instructional resource site MERLOT, you can see how effectively a search engine takes you to your sought-after target. In a typical search you can't know fully in advance where or what your target is, so you can't evaluate how effective or ineffective the search engine is in helping you find what you are looking for. A simple Google search for "MERLOT" works very well, bringing up contacts with the sought-after target in the first five listings before diverging into sites for Merlot wines. However a KartOO search does not succeed as well; it does bring up the desired taste.merlot.org repository address but most of the other listings are for wine sites. If the spelling of the search term is changed from "MERLOT" to "Merlot" something a person unfamiliar with the repository listing might do, then the KartOO search engine is even less successful at identifying what is sought. If a more advanced search technique is applied, using "MERLOT Educational Resources" as the search phrase, then the results are much more targeted, with all wine listings removed. But notice that I had already adjusted my search language to talk to a machine from the first search. If I do a Google search with the kind of query that I'd use with a person ("Find the MERLOT instructional resources site") then Google finds nothing. KartOO does do much better because it has been set up to accept plain language queries ("Find the MERLOT instructional resources site"?), but it does take 7 pages of listings to finally turn up the taste.merlot.org address; instead it brings up many sites that refer to MERLOT. The AskJeeves search engine also provides plain language search capabilities but found very little when give the "Find the MERLOT instructional resources site" instruction and did not locate the desired MERLOT address. Also notice that even in the plain language search example, I kept my language very simple so that I was still doing "computer speak" rather than "people speak." If I were speaking with a librarian or instructional support person I'd use a much more complicate set of instructions such as "Please help me find the MERLOT instructional resources site and all other repositories of online instructional resounces. I want to restrict the search to higher education resources and resources in English. I'd like to exclude all discipline-specific sites." Just for fun I tried this complete set of instructions on Google and did get a number of relevant sites, even though Google indicated that it "limits queries to 10 words." When I tried the same complex set of instructions on KartOO nothing was located. It's impressive to me that the Google algorithms were able to reductively cope with the the complex set of instructions and yield items of use--even though Google was unable to fulfill the complete set of instructions. It's also impressive that the advanced search tools in Google and KartOO and other search engines permit a searcher to construct a complex sequence of searches once the person learns the search protocol language of the software. I'm not sure what to conclude from this brief analysis, except to say that the field of knowledge management and the study of human/machine communication is very much a work in progress. What is encouraging is that the fields exist at all and that progress is underway. However we are a long way from the kinds of conversational communications with machines that are so popularly depicted in science fiction books and movies. [EduResources--Higher Education Resources Online]

Fri, 31 Jan 2003 03:33:11 GMT

Graph-based Data Models and Languages. (thanks Murray) [Semantic Web Blog, featuring RDF]

Wed, 29 Jan 2003 18:37:48 GMT

The Digital ID Federation Myth

The key to any federation is understanding who's in it and who's out. The Digital ID federation concept sounds attractive, but doesn't include the customers, whose voice and stake in the game are like American Indians in post-Civil War America. Just because the federation issues get ironed out doesn't mean they'll do us any good.

But were we to assume that everyone controls their own web space, we have the foundation of an authentic federation.

Self-hosted Identity

Ming discussed self-hosted identity on Monday, worth repeating verbatim:

James Snell talks about being in control of one's own identity and storing it on one's own site, like as part of one's weblog:

"A discussion on Sam's blog got me thinking about self-hosted identities. Ideally, I should be able to put together a file, discoverable through my weblog, and digitally signed with my private key that contains all of the personal information that I want to make public. When I go to any type of forum (like a weblog) or to a commercial site (like Amazon), if they want my information, they would do what Dave suggests and put a "You know me" button on their page. When I go to the site, I click on the button, the site asks me for the location of my identity file. They download the file and extract the necessary information."

And he follows up here and here . We need that, of course. I'm tired of having entered my information on dozens of different sites over the years, and it being mostly outdated and forgotten. Much better that it is on my computer.

This is a more sophisticated form of the federated ID solution we baked into our microeconomy. The first step in letting people control their ID is to bite the bullet and require everybody to have their own web site. That seems like a big step, but it's shrinking daily. Blogging is one of the best reasons to cross the website divide, and identity is pretty close.

Xpertweb users assume their transactions are as public as a public company's. If you want to do a transaction "off the books" you won't want to do it using your Xpertweb persona(s). But for most transactions, transparency solves far more problems than it raises.

The Xpertweb protocols have no need to expose the buyer's financial information. Payment is made after the sale, through a trusted third party managed by the buyer, since the final price is dependent on the buyer's rating of the transaction. The only data needed to start the transaction is how to get the product or service into the buyer's hands. This inversion of the transaction—caveat emptor becomes caveat vendor—solves most of the difficult problems of identity theft and its handmaiden, Digital ID.

So Xpertweb's ID need not be as complex as Snell's thorough treatment, but the approach is perfect. Maybe we can convince Ming or James Snell to help out on this feature for our open source microeconomy...

The key to Xpertweb's usefulness will be the ease of using the forms, and having all the buyer's relevant data filled in automatically is a great start.

Blogging for Dollars

An Xpertweb page is basically a web log that keeps track of your words and comments of course, but extended with a commercial form of highly structured trackback. Every time the buyer submits a form, any data saved on the seller's site is duplicated on the buyer's site, by the buyer's trusted script, in the form of an order confirmation page. Then, as the transaction progresses, the mirrored data store is enriched, culminating with each party's grade and comment, which is the point of the whole system.

In the agora, everyone can watch each other shopping. The citizens are on display like the melons.

[Escapable Logic]