cftopper.com

This Prototype javascript library that everybody is talking about looks to be fantastic but where the hell is the documentation? I mean apart from the Sitepoint article and a few others theres no official documentation (or even a promise of documentation en-route)!

I've just done a big of digging and have just found that Sergio Pereira has just written "Developer Notes for Prototype.js" which is an excellent introduction to the prototype library. In his notes, I found a good overview of the main reason I am thinking about using prototype.js at the moment - it's easy-to-use but powerful Ajax framework.

I have to say I am very impressed with the Ajax.Updater class which will just shunt a bit of HTML returned from the server into place on your existing page with almost no developer effect:

var myAjax = new Ajax.Updater(
'placeholder',
url,
{
method: 'get',
parameters: pars

});

I like it!!

I think the site administration for cfTopper is pretty sweet. I should point out that the CMS for cfTopper is build on Digital Crew's InfinityCMS platform yada yada... we've been continually developing it for a number of years, slightly improving it for each new client site we develop. InfinityCMS makes is easy to knock out a custom CMS for just about anything.

Here is quick screen of what the CMS for cfTopper looks like:

I had a slight pain in my stomach yesterday afternoon and thought nothing of it. That escalated into what felt like my stomach was digesting itself with all sort of funny sounds emanating from by body - don't ask. I'm told that I probably have a 24 hours bug. Feeling a little better today but still not 100% better - using this as a great excuse to work from home (ie. stay in bed).

There I am lying in bed drifting off to sleep... my mind wanders... I start thinking about servers and bandwidth (yes I know I am extremely weird) and it occurs to me that it would be relatively easy to share a load across multiple web servers by having one the master server send a redirect header to another server to share the load.

Effectively we could have a number of different servers running distinct web sites. Then when one server is not busy it could help out another server by hoarding and serving the most-used simple (images, CSS etc.) content from a local cache.

A rough gist of how it might work:

1. All servers in the cluster stay in touch informing each other when they are not too busy.
2. A server that is busy will ask the other servers to help it share the load, the other servers will respond by saying OK, and then downloading and caching some simple content - CSS, images etc.
3. The busy server can now send simple redirects headers back the web browser informing the browser to fetch the content from one of the other servers. Tada! -  -  the non-busy servers will help serve the content - simple load balancing.

OK, somebody's going to get back to me a say that there will be a slight delay caused by the redirect header but come on we're talking milliseconds and the end would certainly justify the means.

Is it feasable?... well it could be implemented by developing a simple ISAPI filter. I don't think the actual programming would be nearly as easy as the principle, there'd be loads of factors to take into consideration, cache management, cache size, most effective load-balancing algorithm (round robin etc), inter-server communication... but it'd be a pleasure to solve those problems.

The end result is that a company who own 10 servers serving completely different websites can force those servers to help each other out greatly reducing the strain on any one server... sure they can't all be busy at once!

You thoughts would be appreciated. I may be crazy... I'll go away and program the comment part of this site now so you can have your say (check back if its not done).

What a start to my blogging... I just accidentally deleted my first blog post that I made last night. and no, I checked, there was no backup. How depressing. I'll go rewrite it this evening. Too much work to do now.

Update

Amazingly I did find a backup I made last night just before shutting down the laptop... the original post is back in place. All is well.

It was a long time coming but I've gone and jumped on the blogging bandwagon.

The purpose of this blog

I guess I should write about what the purpose of this blog is and set some goals... Well mainly this blog will be about ColdFusion Application Development interspersed with my crazy ramblings.

I'm going to also make an effort to contribute to the ColdFusion community by using this site to give out tips, tools and thoughts on all things ColdFusion.

I decided to program this blog from scratch instead of using an existing engine such as Ray Camden's BlogCFC. Why? Well, um, BlogCFC doesn't rock my boat is one reason, the other is that I wanted maximum flexibility. Or is that I just don't like using other peoples code. I don't know. Anyhow, it's done now. So far I've invested much more time developing the CMS for this site that actually creating the site. Already I'm feeling that the design is a bit old. I stumbled on a few great sites during the development of this site that impressed me lots... but I really did want to keep this site light and simple.

Some features I'll be adding

• Right Hand Side Column - I haven't done this yet because I want to do something slick here - probably flash.
• RSS Feeds
• Comments Engine
• Who am I blurb
• Fixing the shadow effect on post titles - Sam reckons the shadows could be much better *shrugs*
• Improving overall design - colours etc. Load of ideas, so little time. 
• Actually add good content - I have a good few ColdFusion tools I've developed over the years, some of which I find invaluable. I'll be sharing them all. Check back!

If you have any ideas, let me know - topperATcftopper.com.