You may have noticed that Revyu has recently taken a bashing from the seedy end of the search engine optimisation brigade – people intent on abusing the site’s high ranking in the search engines to try and drive more traffic to whatever destination site pays their fees at a particular point in time. There’s no decent word for these polluters of the Web, who try and make a business by devaluing otherwise useful sites with links to products and services that most people don’t want or need. They’re abhorrent, and they’re taking their toll on Revyu, leaving junk reviews that have to be cleaned up by someone. They will be beaten in the end, but in the meantime we need to take a break from their onslaught and replenish our resources for the inevitable fight to come.

To this end we’ve temporarily disabled new registrations on the site while we clean out the garbage that’s been left behind.  Existing users can still login and add new reviews, and we welcome these thoughtful contributions as readily as always. There’s still a chance that some accounts have been created with junk reviews in mind but not yet used; these will be deleted as soon as possible after they’re detected. Our plan is to clean up the mess, take a break from the effort, and come back with a smarter system to prevent (or at least reduce) the kind of contributions that devalue the site for the majority of users. Thanks for your patience while we do this.

The number of reviews in Revyu has been steadily creaping up, and I’m really pleased to note that we recently passed the 1000 review mark. The 1000th review came from Revyu stalwart Paddy and was a less than glowing review of Google’s recently launched Chrome Web browser. As always, real opinions from real people motivated to write reviews make it much easier for the rest of us to work out what’s hot and what’s not. Here’s to the next 1000.

In terms of development, it’s been a quiet time for Revyu since winning last year’s Semantic Web Challenge, but there’s plenty going on behind the scenes, and much more planned for the future. Probably the biggest change has been that I submitted my PhD thesis and took up a new job in the Talis Platform Team, IMHO one of the hottest groups around working on Semantic Web technologies. My role is as a researcher, working on recommender systems and continuing the contribute to the Linked Data effort. Needless to say there are many nice parallels between this work and Revyu.

Usage figures on the site have continued their upward trend in the last few months, with more than 260,000 page requests last month alone, and on average almost 14,000 page requests per day so far this month. The continuing upward trend is helped by (or possibly a result of) an increase I perceive in the number of contributing users, with what seems to be a slowly rising proportion of contributions coming from people we don’t know. I’ve certainly had less time to contribute reviews, but this hasn’t slowed the rate of new additions, and stalwarts such as drewp are still going strong.

One of the other things we’ve been working on, which may have had an effect on traffic, is the improvement in our coverage of pubs, bars, restaurants and cinemas in London, thanks to a data dump from the Randomness Guide to London. Kake (who’s been a welcome contributor to Revyu) and co at the RGL were kind enough to give us a dump of their database, allowing us to import basic information about roughly 700 places in London into Revyu, ready and waiting to receive your reviews. In due course we’ll use the excellent RDF output from RGL to show you fancy maps of the places we’ve imported, just like we do for places in Milton Keynes (e.g. the Brothers Fish Bar).

So, that’s what we’ve been up to, amongst other things. In the medium term we’re looking at ways to improve the underlying infrastucture behind Revyu, so that we can host lots more data and loads more reviews. I don’t want to say too much at this stage until the plans are more concrete, but one element I’m really excited about is to prospect of better Revyu APIs, in addition to the existing SPARQL endpoint. We’ve also got plans to clarify the licensing terms of data in Revyu, so you know where you stand when you build applications on top of Revyu.

Needless to say, exciting times ahead…

We’re absolutely delighted to announce that at the 6th International Semantic Web Conference and 2nd Asian Semantic Web Conference (ISWC+ASWC 2007), held last week (11-15 November 2007) in Busan, Korea, Revyu.com was awarded first prize in the 2007 Semantic Web Challenge!

Quoting from challenge.semanticweb.org, “The Semantic Web Challenge offers participants the chance to show the best of the Semantic Web”. Challenge Co-Chair, Peter Mika added “this was by far the most competitive year, with the highest number of participants ever, and a significant improvement in the quality of the entrants“. Revyu was selected as the challenge winner from a record 23 entries of which 18 made the first round and five were shortlisted. The prize was awarded during the ISWC + ASWC 2007 closing ceremony, where delegates received a live Revyu demo.

This is great recognition of the contribution Revyu can make to a Web of Data and the hard work that has gone into the site, not least of which from contributors to the site. Thank you all! We’d also like to send our congratulations to the winners of second and third prizes, Potluck and the CHIP Demonstrator, and the runners-up GroupMe and iFanzy. Well done everyone!

October 2007, Revyu’s busiest month yet, with a total of 135,776 page requests. One interesting trend in recent months has been an overall decrease in the proportion of traffic accounted for by search engine crawlers. This is now down to around 25% of page requests, with Googlebot generating around three times as many requests as crawlers from the live.com and yahoo.net domains. We’ve also started to see more activity from archive.org and alexa.com crawlers, and a pleasing development is some more substantial crawling from Semantic Web services such as zitgist.com.

Top 20 Search Engine Queries bringing Traffic to Revyu in October 2007

• 4192 referrals: www.facebook.com
• 662 referrals: facebook.com
• 600 referrals: barn asia
• 341 referrals: www.facebook
• 153 referrals: kick and dicky
• 110 referrals: park farm bristol
• 109 referrals: cicciolina
• 105 referrals: costco milton keynes
• 97 referrals: wharfedale micro drive in ear headphones
• 97 referrals: 971 the river
• 96 referrals: brazilian club london
• 93 referrals: goosync review
• 88 referrals: barn asia newcastle
• 69 referrals: canon d400
• 69 referrals: wavendon arms milton keynes
• 68 referrals: the wavendon arms
• 66 referrals: barn asia newcastle upon tyne
• 63 referrals: goosync
• 62 referrals: las iguanas milton keynes
• 61 referrals: costco mk

The results are remarkably similar to last month, with regulars such as facebook, barn asia, park farm bristol, goosync and 971 the river well represented. The notable new entries are cicciolina and the recently opened Las Iguanas in Milton Keynes.

There are too many queries at the end of the long tail to report in this blog post, but I’m entertained to see someone searching for “dodgy restaurants and fast food” – I can’t think of a better use for Revyu than that. Unfortunately the first result on Google for this query is a link to the Revyu page about the excellent Indian Fast Food Restaurant and Takeaway, Cotham, Bristol. This is a great example of how keyword-based searching will never be the answer to all our information seeking needs. For anyone looking for excellent indian food in Cotham, Bristol, rest assured that the “bring your own booze” Indian Fast Food will also serve you up a great meal.

Continuing the Doodz-n-Chyx-inspired series, here are the…

Top 20 Search Engine Queries bringing Traffic to Revyu in September 2007

• 3971 referrals: www.facebook.com
• 613 referrals: barn asia
• 146 referrals: www.facebook com
• 136 referrals: barn asia newcastle
• 107 referrals: www.facebook
• 85 referrals: brazilian club london
• 80 referrals: kick and dicky
• 77 referrals: jbl platinum series
• 76 referrals: goosync
• 76 referrals: park farm bristol
• 75 referrals: goosync review
• 73 referrals: 971 the river
• 70 referrals: wavendon arms
• 63 referrals: wharfedale micro drive in ear headphones
• 59 referrals: barn asia newcastle upon tyne
• 54 referrals: park farm bristol airport
• 51 referrals: park farm parking
• 50 referrals: wavendon arms wavendon
• 50 referrals: wavendon arms milton keynes
• 44 referrals: review anything

I still don’t understand why so many people search for a Web address like www.facebook.com, but I guess for some people Google is the Address Bar. The list is also oddly similar to last month. Partly for that reason (although partly because of time admittedly) I’ve not linked each query back to the associated item on Revyu. Would I just be contributing to a “rich get richer” situation? Does that matter when it’s a zero sum game Revyu vs other sites, not one entry on Revyu vs another? Probably not. I’ll chew this over and decide again next month.

Again the long tail is interesting (perhaps next month I’ll publish the bottom 20). My personal favourite is “quality of kite sold at aldi”. Just 1 referral, but glad to see Revyu being used for what it was designed.

Today (at last!) I upgraded Revyu to use version 0.9.5 of the RAP library, the RDF API for PHP on which the site is built. This latest version includes a major rewrite of the SPARQL engine for database storage, giving major performance improvements. By my rough calculations the Revyu home page now loads around 5x faster, and you should see performance improvements right across the site.

The weekly email update I receive about new content on the awesome Kewl Doodz’n'Chyx blog includes a short list of search terms that have brought people to the blog in the past week. Inspired by this I thought I’d do something similar with Revyu and publish a list of the top search terms bringing people to the site each month.

You folks all work hard to write reviews but never have any idea if anyone’s actually reading them. Well the good news is that people are. In fairly impressive numbers. The list below shows the top 20 search terms bringing people to site from external search engines in August 2007. In that month the site received an average of 3765 page requests per day, and passed through the milestone of half a million total page requests since launch in November 2006; not bad traffic from a relatively modest 475 reviews I reckon.

Top 20 Search Engine Queries bringing Traffic to Revyu in August 2007

• 633 referrals: www.facebook.com
• 374 referrals: barn asia
• 217 referrals: barn asia newcastle
• 144 referrals: www.facebook
• 137 referrals: goosync
• 83 referrals: goosync review
• 76 referrals: 971 the river
• 74 referrals: helix apartments seattle
• 67 referrals: wharfedale micro drive in ear headphones
• 62 referrals: jbl platinum series
• 62 referrals: helix apartments
• 57 referrals: park farm bristol
• 54 referrals: hotel particulier montmartre
• 52 referrals: the garage wolverton
• 49 referrals: brazilian club london
• 45 referrals: review anything
• 44 referrals: kick and dicky
• 44 referrals: wavendon arms
• 41 referrals: www.facebook com
• 40 referrals: park farm parking bristol

It’s pretty interesting to see what’s on people’s minds, and I hope that these figures give you some encouragement and assurance that your views are actually read. For those Web2.0 fanatics among you, or those of you whose reviews aren’t covered by the search terms above, be assured that there’s a really long tail on this data (i.e. there are many queries that have led to traffic on Revyu just one or two times). Right at the end with just one referral is “fratboy tags”. I can’t think why.

Chuffed! Revyu has been shortlisted for the 2007 Semantic Web Challenge. Every year the Challenge (which is co-located with the International Semantic Web Challenge) seeks to recognise some of the best work in the Semantic Web world. From a record 23 entries Revyu has made shortlist of five, with the winners to be announced at ISWC2007 in Busan, Korea. Thanks to all Revyuers who have helped get us this far

In fixing the content negotiation earlier in the week I managed to break a couple of other things on the site. Oops. Sorry.

Happily these are now fixed, but if you ever spot any bugs on the site, please don’t hesitate to contact us.