drakard.com

Below is a short overview of what the site needed doing, and a quick skim through what was done. Both aspects might be expanded upon – certainly, some of the “what was done” side I plan on expanding into proper articles in time.

Before: The Existing Site

Amdram.co.uk is a venerable site in today’s internet, having been in existence since 1997, and over the years several different ways of storing content had arisen – from static HTML pages and old Perl-driven flat file databases to separate forum software and external Blogger.com posts. Going hand-in-hand with this legacy was that the information was not always structured consistently as navigation elements changed and were added, so while specific elements of the website remained lively, the bulk of the articles and features had fallen into disuse.

Although the site did deliver good search engine rankings (as it had been around long enough to build up page rank), SEO had not been a main consideration when the previous site was designed and a general facelift was also required. On a related note, advertisers had to be entered manually into the HTML and could not get much, if any, feedback from their ads.

Overall, a complete redevelopment and redesign was required, along with the preservation of eleven year’s worth of variously formatted data.

After: The New Site

For the relaunch of amdram.co.uk, we had the following conditions to meet:

• Migration of all existing content.
• A fresh look to the site and a level of organisation to the data contained within.
• Ease of use for the admins to edit content and generally maintain everything.
• To maintain the search ranking built up over the years, while ensuring that it can remain healthy for the future.
• Separation of advertising management from the actual content.
• To encourage users back into the bulk of the site, rather than solely visiting the forums.

Firstly, the existing site structure was analysed and broken down into rough collections of content. From this, we could draw up a site map in conjuncture with the site owner and start planning how we would classify the data in the new site and how the users could navigate through it.

After my latest vow of trying not to reinvent the wheel again, it was decided to look for an existing CMS to adapt to the needs of the new site and for that purpose WordPress was chosen as I had recent experience in customising it for a similar purpose. So before we proceeded any further, conversion scripts were written to parse the original data and automatically create WordPress database entries. Now that the existing content could be successfully imported into a default install, we could design and create the front end of the site.

A heavily modified combination of two existing WP themes sufficed to meet with the owner’s approval for the look and feel, but the speed was something of a concern. Since an off-the-shelf CMS must be generic enough to be downloaded and installed by potentially anyone on anything, it carries overheads (and other drawbacks) that aren’t suffered by a custom CMS.

Fortunately, there are ways of speeding up WordPress sites – and indeed, any site:

• caching common database queries to stop having to look up the same information,
• compressing the content so that the user doesn’t have to download large static files,
• having content expiration dates so that those static files don’t get downloaded again,
• caching the compiled PHP code to save the server working it out each time,
• using the WPcache plugin as WP can need all the help it can get,
• and removing unnecessary code in the WP theme since it is specific to just one site.

With these tweaks in place, the site was much quicker and bandwidth usage was reduced as a welcome side-effect.

Since I knew that the site owner was interested in developing additional websites (with different content) in the near future, decisions regarding advertising and media files in general had to be taken with one eye firmly on expansion.

For the advertising management in the new site, prior exposure to an existing solution was invaluable again. I had already been impressed by OpenX when researching it for a previous job, and here it was the logical choice as it was design to serve ads across multiple sites.

Likewise, ResourceSpace was already known to be a fairly useful way of managing an image library, making it easier to build up a central store of the type of images that accompany articles.

With a few other additions and alterations, amdram.co.uk was ready to be relaunched. Over the forthcoming months, feedback will be gathered from various sources and incorporated into the next phase of development.

Overview of the Systems Used

With the following four open source software packages, it is relatively straightforward to build a network of easily managed websites:

• WordPress – the heart of the matter. Replace with your favourite off-the-shelf or custom built CMS.
• osTicket – a user support ticketing system that may or may not see heavy use, depending on said users. Can integrate into your design fairly simply.
• OpenX – an advertising server which is, admittedly, a bit over the top if you’re not running it across multiple websites.
• ResourceSpace – digital asset management so that there is one central place to both manage and serve media files – again, overkill for single sites.

Tagged with: clients, development review

This entry was posted on Monday, December 1st, 2008 at 10:43 am and is filed under Articles. You can follow any responses to this entry through the RSS 2.0 feed.
Both comments and pings are currently closed.

Articles Articles on the Internet and other aspects of technology by drakard.com Amdram.co.uk Development Review A brief look at what was done during the redevelopment of amdram.co.uk and launch last week.