Sunday, November 06, 2005

Web 2.0 podcasts, AJAX and all that...

This week I have been mostly listening to podcasts...

O'Reilly's Distributing the Future podcasts

For those of you in the dark about what a podcast is, a podcast is essentially a RSS feed which links to audio files rather than web pages. The podcast I have been listening to is O'Reilly's Distributing the Future, which as you might expect from O'Reilly contains some very high quality content.

I'm still in the middle of listening to the recent batch concerning Web 2.0 and the O'Reilly sponsored Web 2.0 conference. I may transfer the MP3's to my pre-iPod Creative MuVo 128Mb MP3 player but in contrast to the usual assortment of background music I keep on it, I really do intend to listen to these broadcasts carefully. Some very interesting points were made by Marc Hedlund in the Data for Web 2.0 entry. He was speaking about the merging of online and offline content. He mentioned Google Desktop which I have coincidently been looking at this week.

Google Desktop

Google Desktop's (GD) new sidebar provides some Apple Dashboard / Konfabulator like functionality. The real core functionality of GD is that it provides a search engine facility for the files on the local harddisk, the clever bit is that when I access Google from my browser I now have an additional options to search my local hard disk in the same way as I would search the web. This is partially achieved by the GD client software installing a "web server" on my machine for the local search. Marc Hedlund makes the point that many people do not understand that although they can search their local files via Google that data is not accessible to anyone else using Google!

Oracle Collaboration Suite

I attended a technical presentation given at my workplace by Oracle this week about Oracle Collaboration Suite. [engage rant mode] Now as the old saying goes 'when you all you have is a hammer, everything looks like a nail'. Oracle's Hammer is called the Oracle 10g Database and the nails in question include e-mail, calendar, voicemail, mobile communication, content management, search and file management. Apologies as I might be guilty of paraphrasing this in a quite cynical way! Oracle are completely open standards based and as long as you only every intend to use Oracle's email server, calendar server, LDAP server, application server, database, search technology, single sign-on mechanism, web service APIs, development software, and development tools (and a few other Oracle bits and bobs) you should be just fine!

Undeniably impressive was the concept of "workspaces" which looked like a quick way to get Yahoo/Google Groups style of web presence established quickly.

For quite a bit of the presentation I was told how I could upload my files using WebDAV to a backend database and this could then be accessed by others over the web. I got the impression that all we would need to do this was a hefty 10 processor server with a couple of terabyte hard disk (or 5 suitably configured 2 processor Linux machines), some extensive capacity planning, a cultural overhaul and complete organisational restructure and then the Oracle Collaboration Suite would be a very good fit for our organisation. [disengage rant mode]

To be fair I can see the some clear benefits of this approach but I couldn't help thinking that I could virtually already do some of this "on the cheap" by simply installing a web server on my workstation. It would seem that centralizing all our services on one server in the midst of the general trend to farming our processing to the client side is a bit like urinating into the wind.

The future?

There is currently a finite number of internet ports on my workstation, there is therefore a limit to the number of web servers I can reasonably install on my machine. This becomes especially critical if each client requires multiple ports for secure and insecure connections, etc.

There is a lot of potential for a "mashup" of the technologies using AJAX and Greasemonkey married to the Google Desktop style of local / remote web interaction. The fact that I heard Mitchell Baker talking about Greasemonkey in a "Distributing the Future" podcast and Mark Pilgrim seems to be writing a book about it fills me with expectation that it will soon become an important development tool that I can't ignore. There may well come a time when a local web server installation will become integral part of the web browser itself. If this web server contained a JavaScript scripting engine like Mozilla's Rhino, all the better. If the web is going to make a transformation from Web 1.0 where it is essentially a "Pull" (read only) to Web 2.0 where it will be more "Push and Pull" (read and write) medium then, to me at least, it makes total sense that we will have all have webservers on our workstations (even if this is less obvious than say a local install of Apache HTTPD is now). I hope that the need for this kind of technology is identified more widely soon so that a standards body can emerge and establish a multi-vendor reference implementation and common programming interface. Is this what Web 2.0 will become? I may have more insight when I've finished listening to the "Distributing the Future" podcast!

The future is here it's just not evenly distributed yet...


Unknown said...

A real mixture of United states technologies as well as Europe rolex replica is actually cause sufficient in order to diffuse any kind of question. Which means you understand the solution could use one that actually think about the reason why purchase a good IWC view? Many people choose Rolex fake hublot however, maybe you have asked yourself why purchase a Panerai view? This particular manufacturer appeals to everyone’s interest using its minimum as well as uncommon design. A few of the options that come with fake cartier wrist watches tend to be they've already big lustrous face, a good building and therefore are water-resistant. Right now for those who have obtained the choice to buy luxurious fake cartier view, after that the reason why purchase a good IWC view? IWC is recognized as since the the majority of unique rolex replica view producers generating mechanised procedures for each number of designer watches.