Sleep Centre, Security, and Search

Recently we rolled out a handful of notable changes on this web site, I rarely talk about the web site but today will be an exception.

Recently we rolled out a handful of notable changes on this web site, I infrequently talk about the web site but today will be an exception.

We have improved Sleep Centre with a new structure and layout. We have begun the transition to promoting secure web links by default, and we have a new site search.

Heart carved into a window shutter is almost as old as the company. We registered the domain 15 years ago and have been growing and improving the site ever since. It was not even our first domain, as DssW was already on the web when the company was incorporated.

The web is constantly changing and never finished.

During that time we have shifted between many underlying approaches and tools. Creating, publishing, and maintaining a site of this size requires dedication. The trend in our tools though has always been clear; a preference towards hand crafted pages and approaches, assisted by scripts and small tools. Monolithic frameworks and applications that hide the detail are not helpful in the long term.

We are always refining our collection of tools. Most are home grown and reflect our very specific needs.

So what have we recently improved?

Sleep Centre

Sleep Centre is a large archive of usenet threads about energy saving, power management, and all things battery related. The conversations focus on Macs and are a treasure trove of questions, answers, and insights.

The archive is ageing. The face of usenet has changed dramatically. A while ago, we stopped adding new conversations. Now we aim to keep the collection online, searchable, and accessible to all for as long as possible.

To continue this we recently rewrote the tools that help us collate and convert the mass of text based conversations into web pages. The first results of the rewrite are now published. Links have been changed to be more readable. Automatic redirections are in place to ensure existing links do not break. has been available securely via Secure Sockets Layer/Transport Layer Security (SSL/TLS) for many years. Nothing has changed with regard to our certificates or security.

What has changed is subtle but important. We have begun promoting the site’s canonical links using the encrypted varient. Well formed web pages include a canonical link. This link tells search engines and web tools where the page’s canonical location on the Internet is.

Previously, our links started with the traditional http://. Now they start with https://. The single extra s in the protocol tells the search engines to prefer, and link to, the secure edition by default.

Security by default is critical in our products. Increasingly security by default is becoming important on the web. Critics question the additional computational burden and potential to slow the web down; we feel the impact is worthwhile. When you browse and download from, you should be able to do so privately and securely.

We have a new site wide search. The search box in the top-right of the page now returns results from a private Sphinx Search powered engine.

It is going to take a while to refine and embellish the results to the standard we want but already the results are accurate and an improvement over the previous search.

Previously we used a wonderful perl based solution called Fluid Dynamics Search Engine (FDSE). Looking back through our records I discovered we bought our licence in 2004 and have been using it ever since. Only a year later, after we bought our licence, FDSE stopped being improved but the core problem was solved and the software was reliable.

Given FDSE’s reliability why change? Firstly, FDSE is no longer being improved. That presents the possibility of security risks. This is not the main reason to move.

What drove the change to Sphinx Search is that our expectations of what a site search could be have changed. We wanted a faster, less resource intensive, and a more flexible search.

As a welcome bonus, Sphinx Search is open source and we have the chance to delve into the code, should the need arise.

Never Finished consists of many thousands of pages and we know the site will never be finished. There is always more we can document, always more we can improve, and the web will always define new expectations to keep us busy.