Site Specs

I have decided against protecting these major directory indexes from view, see left, because a major part of this project was actually to show off the technology and architecture underlying the whole thing. Protected areas are protected, sure, but I invite anyone to navigate through the directory indexes to beg, borrow and steal code at their leisure - just as we have all done since the beginning.

The whole site, apart from dynamic elements via javascripts.htm, Blogger, Moreover and the Fluid Dynamics Search Engine (www.xav.com) are static, but are under the complete control of a set of virtual includes found in the /network/include2012/ directory. The whole site's color and look, though not it's physical HTML architecture, can be changed through the use of a handful of these includes.

The major plus in this architectural style is that I can add and subtract elements from every content cell on every page. Considering that the site, as of May 2002, had 191 content files, this is the best way (without a database backend) to control an ever-increasing content site. Certainly, these are tips and tricks I learnt while developing TIMEasia.com, TIMEeurope.com, TIMEpacific.com, Asiaweek.com, Who.com, FORTUNEchina.com and parts of CNN.com's Asia sites - all very, very large content sites.

- Content directories

There are a series of content areas where I can use specific header and footer includes to personalize said sections. You can browse the raw files in any of these directories:

/network/digitimes/
/network/pics/
/network/rev/
/network/revue/
/network/video/

Note the different header and footer includes at work on these pages. If you look at any source code within the site, I'm usually pretty pedantic about signalling includes with bloody big markers, so look for them.

- Home pages and the search structure

Having no idea day to day what could appear on the site, but knowing that as the site expands I wouldn't keep up a plethora of home and index pages, the whole site is built around the search engine. Instead of a static link, most navigation purely triggers a search for the term required, thus the traditional idea of a home page is automated. The only differences are the two actual home pages at www.adamconnors.net and /network/. These are static. Plastered with search triggers though, they perform a more dynamic function than straight links which will inevitably go out of date.

Links are thus usually coded in a style like this:

<A HREF="http://www.adamconnors.net/cgi-bin/search/search.pl?Terms=revue" CLASS="globalnav"></a>

Simple, huh? Also note, all static links, used in the architecture, are absolute to the /network/ directory.

- Content cell includes

A basic map of where includes power each page TK.

- Blogger

Using the ubercool free publishing system Blogger, available online 24/7, I can update Mental Notes whenever the fever hits. But apart from that, you may have noticed I've highly customized how the output from Blogger adds content to every page of the site, in any format I desire. This again is done through includes, this time highly complex. Probably so complex I could never recount how it was done. Large technical explanation TK.

- Moreover

Thanks to another free service, Moreover, I can serve headlines from any number of sources at the base of every page using includes. Though this draws heavily on the page load speed, because it needs to contact an external server on each pageview, I have instated one basic fix: separating the top tables from the Moreover feed include so that they load first. Thus, the top of the page and content will have loaded while Moreover is contacted for its content.

One fix TK is contacting my ISP and having them do a cron job on cacheing the news feeds once an hour, thus storing the headlines locally and speeding up load speed. This, however, I wrestle with because I actually want THE LATEST HEADLINES. So, it's a tradeoff.

- Features

A wicked little javascript I've been hunting down for years, it's available at /network/include2012/ in a file named javascripts.htm. Just drag the file onto your desktop to view in a text editor, or look at the source on any page. An image- and text-random swapping script, it's beautiful. You can plug in any number of variables. Through a system of includes, it is imbedded in each page either at the bottom of the page, or, as on the /network/ page, at the top. See the working test model at /network/features/random_feature.shtml.

Note, for all includes, a cascading style sheet (CSS) - see the file css.htm - for font and color usage is all centralized, thus the formatting will not appear when you view includes apart from their functional pages.

- ISP

The site is based on servers in Atlanta by Interland. I went for the very big boys, and for US servers, due to the fat pipes and reliability. If you're ever going to noodle with streaming media, you need huge pipes.

- Hardware and software

TK.

- Adam Connors, May 2002