Just some thoughts

Tuesday, February 21, 2006

Well, I've been thinking about lisp web-applications


I just read a bit of this email conversation. The conversation basically compared and contrasted the views of the RESTful school with those of the continuations school. Basically, the two participants agreed that RESTful architectures are good for web applications which act more like "collections of resources named by URLs", and that continuations are better for "applications which happen to use the browser for a GUI."


I am interested in using common lisp to create a weblog, and perhaps some other things. I started reading this conversation, because I am trying to figure out if I want to use TBNL or UnCommonWeb. I am particularly fond of google's use of asynchronous javascript in its maps, gmail, and feed reader. I'd like to emulate them as much as possible.


Asynchronous javascript (AJAX) maintains state on the client, and tries to respond to the user's actions using that state. Calls to the server may not be needed if the client-side state holds all the information necessary to handle a particular action. When calls to the server are necessary, request and receive a much smaller amount of data than would be needed for an traditional page view. These calls are much like a function call within a "traditional" application.


Before I read that conversation, I was leaning toward using UCW. Continuations seem promising in simplifying the task of controlling a web application's flow, but I think that will be done better with client-side javascript. I think that asynchronous javascript (primarily its decendants yet to come--think XUL) is the way of the future. In that light, I think I will likely be using TBNL.


Thumps with the "clue stick" are welcome.




Technorati Tags: , , , ,

Monday, February 20, 2006

Here's a great example of what a web page should be.

Robert Hahn has an amazing weblog. His content is great, and his html prowess is amazing. To be really impressed, leave a comment!

Meta-Level Compilation project at stanford

http://metacomp.stanford.edu/

Interesting... they seem to be using it most is as a sort of "lint-on-steroids."

But I look at this:

The goal of the Meta-level Compilation (MC) project is to allow system implementors to easily build simple domain- and application-specific compiler extensions to check, optimize, and transform code.

To me the words "transform code" stand out like crazy. I'm thinking that this sounds like lisp-style macros, although, it seems, with much more work involved.

, , ,

Tuesday, February 14, 2006

Ah! So that's why!

This article says that people have only a 50% chance of correctly interpreting the tone of an email. It goes on to say that people who read the emails feel 90% sure of their interpretations, and the senders are 80% certain that their reader will understand the email's tone correctly.

I've wondered about this, because I've a friendship go sour when the only communication between us was email. So to avoid similar things in the future, I'm going to pay much closer attention to how I word my emails, posts, and writing in general.