Although this question was often a good indication of a reviewer’s overall opinion, there were a significant number of reviews where it didn’t work at all.

In order to address this, the numerical ratings (Performance, Reliability etc) are now also taken into account when deciding which emoticon to display for each review, and this seems to significantly reduce the cases where the emoticons don’t reasonably reflect the general sense of the review.

The good news is that the RAID array has done its job, no data loss so far, I’ve got completely up to date backups (which I know restore properly), and engineers are looking at the physical server right now.

As a precaution, new reviews, new comments, and the members area have been disabled till the problem has been fully resolved. Don’t want to risk losing anyone’s new contributions if things go horribly wrong. It’s also probable that the server will be offline for a period while hardware is replaced. Fingers crossed the faulty kit will be replaced soon, and we’ll be back to business as usual.

Things are getting more personal, and the volume of comments is increasing. Enormous amounts of my time are being spent trying to keep things vaguely civil, to the detriment of the rest of the site, and frankly my own quality of life.

The people involved need to accept that they’re never going to convince the people who hold the exact opposite views.

Starting from today, the following policy changes are being made:

There is going to be a very low threshold of toleration for comments that essentially repeat points made elsewhere on the site, especially if those comments are not about your experience with a particular model, but are essentially part of a wider discussion about politics or economics.

Less toleration for unpleasant comments. I’m not going to spend time editing comments that contain unnecessary jibes. Those comments will simply not be posted to the site. People can learn to tone their comments down, rather than having me edit their comments to make them acceptable. New visitors will get a little leeway on this, as they’re less likely to understand where the boundaries lie.

As a general guide, here are some things that examples of things that I consider acceptable/unacceptable:

Acceptable: Posting new information about an old recall

Unacceptable: Posting multiple comments about recalls already mentioned elsewhere on the site, which simply repeat information available elsewhere

Acceptable: Posting multiple comments that address queries raised by other comments about a particular model

Unacceptable: Posting your unsolicited views of a particular model across many reviews

Unacceptable: Any comments that directly or implicitly are negative about other contributors to the site. It’s fine to say that someone’s wrong, but not to suggest they are an idiot or a traitor.

I don’t expect to see essentially the same comments being repeated either across different reviews, or within the comments of a single review. We live in an age of search engines, and Carsurvey.org designed to be easily searchable. No need to repeat arguments ad infinitum.

I expect everyone who participates on Carsurvey.org to have a sense of respect for the other people who’ve taken the time to add their contributions to the site.

My hope is that these changes will improve the signal to noise ratio of Carsurvey.org, as well as the general atmosphere, without restricting the subjects that can be discussed.

As usual, complaints or feedback should be addressed to steven@carsurvey.org, or added to the blog comments on this site. Please don’t add comments about this to Carsurvey.org itself.

Steven Jackson, CSDO Media Limited

Full marks to SoftLayer for dealing with this issue much better than The Planet did when I had a similar issue a few years back.

Apologies to everyone for the outage, and if anyone has lost any reviews or comments, I’m very sorry about that.

Hopefully all will be back to normal later (probably much later) today.

Apologies for any inconvenience caused.

I’ve been aware for years that North American visitors were confused by the site structure, and I eventually came to the conclusion that it was better for the meaning of a year on Carsurvey.org to be inconsistent, depending on the region of the review.

For a North American review, the model year (where available) is used in preference to the year of manufacture. This isn’t the case for the rest of the world.

So a list of 2006 BMW 3 Series reviews will contain 2006 model year reviews from North America, and 2006 manufacture year reviews from elsewhere. Not quite the same thing, but for reviews from their own region, it should fit visitor’s assumptions, rather than trying to make them fit the site’s data model.

MotorcycleSurvey.com has also had similar changes made, but in that case, model year always takes priority over year of manufacture, regardless of region, as that reflects the way motorcycles are marketed.

• When it comes to browsers or operating systems, I don’t think Google care very much about market share of particular products, but they do care about the underlying technologies. The  Chrome browser has a relatively small market share, but it has had a disproportionate effect on the direction of browser development as a whole. There’s an arms race in areas such as JavaScript performance and HTML5 support, and the Chrome browser gives Google a tool to use to push the technologies it favours. Sure, Google would like you to use the Chrome browser, but if you choose another modern browser with similar features, such as Firefox 3.5 or Safari 4.0, I think Google still count that as a win. A similar argument also applies to iPhone vs Google Android.
• A small percentage of marketshare in either browsers or operating systems is still a big number of people. If Chrome OS gets 5% of the operating system market, and Google has say 200 people working on it, I bet that looks like a good investment.
• I’m often faced with casual computer users with old PC hardware. Windows is running like a dog on their system, and often they don’t have Windows restore media for a clean install. Many of these people just about accessing their email, Facebook and some online shopping, not video editing or photo editing. A web browser is enough for them. Currently their best option is Ubuntu (which is still quite heavyweight for old hardware), or some mini Linux distribution. These do work, but there are always some silly issues that spoil the experience. No doubt Chrome OS will have a cleaner interface, will be faster on old hardware, will be kept up-to-date without the user intervening (like the Chrome browser), and will be very secure. If old hardware is supported, it will be by far the best option for basic web users with old hardware.
• More and more people have secondary machines in their homes and offices, and while a web browser isn’t the only application they need on their main system, in many cases it’s more than enough on a secondary device, as long as the price reflects the limitations. Michael Arrington’s Crunchpad is a great example of a device that embraces this.
• In response to Linux netbooks, Microsoft have been selling Windows XP at a big discount to PC manufacturers, and have been quite liberal about the hardware specifications of this hardware. With Windows 7, it appeared that Microsoft wanted to tighten up the restrictions on netbooks, and sell the operating system at less of a discount. Chrome OS gives PC manufacturers a club to beat Microsoft with, and may well force Microsoft to discount Windows 7 for netbooks to a greater extent than they planned. I can’t see Google losing any sleep over this.
• The ARM compatibility of Chrome OS is a very big deal. ARM SoC (System on a Chip) are very efficient in terms of power, and are also very cheap. Sub £100 ($150) smartbooks (like netbooks, but not Intel based processors) suddenly look realistic. 1Gb of RAM, 4Gb of flash storage,  combined with an ARM chip, is already a very cheap platform, and is only going to get cheaper over time.
• If you can have a modern fast web browser on very cheap and efficient ARM processor, why not embed them in lots more devices, many of which already contain processors and RAM. Manufacturers could very cheaply add Chrome OS to a TV, a PVR, a games console, or a monitor. Spend an extra £20 ($30) on a monitor or TV and get Google Chrome OS builtin. The Nintendo Wii has an optional Opera browser, which is passable, but hardly a great experience, and it presumably wasn’t cheap to develop. Instead, just add a couple of cheap chips and you get a good web experience from Google, with the added bonus that Google look after all the updates and security.

In summary, I think Google aren’t going for Windows or Mac OS X head on, but just want more machines out there running Google friendly modern browsers at a low cost. In many ways, Chrome OS reminds me of Microsoft’s plans for Windows CE a decade ago. CE never really made much progress beyond Windows Mobile devices, but in this new world where the web browser is king, perhaps Chrome OS will have more success.