Bayesian Pointers

If you find Bayesian filters useful, here are two interesting open source packages I found, thanks to Stuart Langridge.

Reverend is a general Bayesian classifier written in Python.  If you are tired of hacking SpamBayes, you might want to give Reverend a try.

Orange is a component-based data mining software with a nice bayesian library inside it.  It is implemented in C++, but comes with Python binding and something called Orange Widgets.

OpenGL-based SVG Implementations

I visit Elliotte Rusty Harold's excellent Cafe con Leche Java and XML pages every other day to find the latest news on Java and XML.  From his XML page, I found svgl, a library that displays SVG pictures using OpenGL, taking advantage of the GPU.  At svgl site, I also found Smoke, a similar library that uses OpenGL to render SVG.  Both uses FreeType 2 for text rendering.  Frankly, I feel that SVG lost much of its initial momentum, thanks to lack of leadership from companies like Adobe, Macromedia, and Microsoft.

R.I.P Ruple

As of May 20th, Rogue Wave halted development of Ruple, a Java-based fusion of XML and TupleSpaces.  XML TupleSpaces is a nice idea, but difficult to sell.  If you are interested in TupleSpaces, here is a paper you might enjoy: Dynamic PKI and Secure Tuplespaces for Distributed Coalitions.

I am getting a dangerous itch to apply tuplespaces to web services workflow problems.  TupleSpaces are extremely powerful as coordination infrastures so tuplespaces and web services go very well together IMHO.

Python 2.2.3

Python 2.2 bugfix version 2.2.3 is out.  I probably won't be upgrading unless I run into one of 40 bugs fixed in this release.  Quite a number of my projects involve native code and, to debug those native code, I need to make debug builds of Python and Win32com binaries.  Too much work for too little return.  I'll wait for 2.3 final.

Killer Business Blogging Issue

Ray Ozzie raises what I think is a dark cloud looming over Blogland.

"By suggesting that employee blogs might be "officially sanctioned", or by in some way acknowledging that the corporation benefits from the blog, the company is arguably exposing itself to claims that it is contributorily and/or vicariously liable for any injuries the employee-bloggers cause."
"By providing the employee resources and active support related to blogs, if the blogger is ultimately sued for libel, the plaintiff may very well claim that the corporation is also liable … not a totally specious position." – Ray Ozzie's Weblog

<

p dir=”ltr”>His other post about Nullsoft's WASTE also raises a worthy point: importance of complacency-immune security.

Mental Cauldron

Although I don't think I have all the ingredients yet, I am throwing everything I got so far, including topic mapping-as-blog-contribution and Emergent Markup Languages, into the cauldron that sits on my shoulder and let it simmer for a while.  One thing that bothers me is the the difficulty with updating static blog pages.  Majority of blogs are static outputs from blog tools.  Inlining might do the trick, but there are UI, security, and style issues.  Oh, well.  I'll think about how I am going to set the table later.

iTunes Mustic Store Insights

A must read insightful article on iTunes Music Store by Scott Loftesness, a micropayment business guru and one of few people whose views I respect.  Scott starts from a consumer's view and then examines iTunes Music Store business from a micropayment business expert's point of view.  As I said, a must read if you are interested in the business side of iTunes Music Store.  What do I think of iTunes Music Store?  It's not just a hit, but a homerun of the decade.

Software as Metered-Service

Responding to my "Who will pay for software?" post, Marc Canter points to tiered price model mentioned in his "Pay for usage" post as a way possible solution.

I think a well-known variation, metered service, will be quite attractive if users are given free monthly 'casual usage' allowances (10 hours a month sounds reasonable).

Software itself will be free to download, but if the usage (measured by the software itself) goes over 'causual usage' allowance for the month, the user will be billed for the extra hours.  When the total amount reaches the 'street price' of the software, user gets to own the product outright.

Would I go for this?  For software that I want to use brief-but-intensely or for software I intend to use all the time, my answer is yes.  But for software that I intend to use to create non-commercial contents with, like Maya ($1999 retail) or 3DS Max ($3500 retail), it is still a lot of money and very poor ROI. 

I think we need multiple creative solutions that, together, will bring the magic back into the software business.

Who will pay for software?

On Saturday, "Dave" asked the question "Who will pay for software?" in a DaveNet article.  It's a tough question that raises a lot of issues.  Looking at those issues as a developer is particularly tough since survival instinct warps my judgement.  I feel that a user's perspective might shed more light than that of a developer.

I am artistic in a half-assed way, so I like to dabble with graphics programs like PhotoShop and Illustrator.  I use those programs frequently enough to buy them.  I also would like to do some 3D graphics and animations as well, but professional 3D tools are just too expensive for a dabbler like me.  So here are some questions from a user's perspective:

  1. Why do part-time users like me have to pay just as much as fulltime users?
  2. How can I pay for just what I use?
  3. Why can't I pay as I use the software?
  4. Can I pay when and if I profit from the contents I created?

One solution that might satisfy these questions is shared content ownership.  The idea is to provide powerful tools for free in return for sharing ownership of the contents created by the users using the tools.  Whether this business model will work or not is the big question.

As a user, I am willing to share ownership of contents I create with tools like Maya if I can use it for free.  As a developer, I am also willing to let people use tools I create for free in return for part ownership of the contents they create.  They were unlikely to buy my software in the first place, doing this will reduce software piracy as well as increase market awareness through peer-marketing.  Since most professional users will still pay upfront for my software to avoid sharing content ownership, I am not losing any existing revenue stream either.