I was asked to take a look at Clean Search which is a search engine that has been produced by NetSpark, which empowers "users to take control of their Internet experience, maintaining access to the abundant resources available online, while protecting users from inappropriate content." They go on to say "The company's unique technological approach automatically inspects and filters user content requests, evaluating the content within the page, in real-time, and removing any inappropriate text or image elements before delivering the remaining valuable content in a seamlessly reconstructed page. NetSpark offers the only filter capable of delivering such granular results, providing access to the maximum content possible."
You know what's coming don't you? You've read enough of my posts to work it out I'm sure. So anyway, I took a look. The best term to use when trying to trick a filter of this nature in my experience is 'dogging'. We start with a bunch of inappropriate adverts. (Women looking for male doggers, Cute local girls etc) We're then into hard core porn sites straight away with a single click.
Next search term up is 'blue tit' and I get "Sorry. We don't accept this search term on Clearch.org. Thanks for understanding." Easy enough to work out which word, but it would be helpful if they would say. This is the absolute classic failure of this type of engine. Any word that has a sexual connotation is immediately banned, irrespective of the fact that it may have many innocent uses. This immediately renders the search engine totally, utterly useless. We also end up with a weird situation where 'breast' isn't acceptable, but 'breast cancer' is ok. So why doesn't 'blue tit' work? 'Breast tissue' fails as well. Am I seriously expected to play 'guess the naughty word' before I can use this search engine? Really? There are lots more examples, and I could go on (not bdsm but sadism is ok) but I'll stop at this point.
Let's look now at the image search function. This is where things get really weird. I kept my 'sadism' search and ran that within images. I wouldn't advise you do the same, unless you have a strong stomach, as some of the images weren't exactly family friendly. Hell, they weren't anyone friendly! I tried the 'breast cancer' search again, and got acceptable results - mostly diagrams or images of pink ribbons; fair enough. Again 'breast' by itself didn't work. 'Librarian' was ok, but I was very interested in the related searches option, including such delights as 'hot librarian' and 'Pantiless librarian'. I followed a few more links of this nature and eventually ended up with 'Mammaries'. Oh boy, what fun I could have now (errr, the fun is less pervy than it sounds...) because while I couldn't search for 'breasts' I could search for 'massive mammaries', 'super huge mammaries' and 'giant mammaries'. However, 'big tits' was an entirely unacceptable phrase, as was 'blue tits'. 'Flashing' gave me lots of pictures of the type that you see on a roof, but there was also an image of a young lady in a state of undress. Another good tip - the latin words for various activities also brings up images that really shouldn't be there. As for the images themselves - I guess it depends very much on what you would regard as being acceptable. It's fair to say that I only saw a few pornographic images, and there were certainly images that I would have been very unhappy for a child to see.
Am I being unfair? No, I don't think so. If an engine sets itself to say say that it provides clean searches and it's name 'Clearch' is based on the concept, it's only fair for it to do what it should. I would certainly expect it to do as well as Google safesearch - but one of my image searches returned what was clearly a sexual act, and the same search at Google didn't. Of course, the complaint could also be made that I went out there looking for that sort of image, and it's a fair point - with a search engine that states that it does something particular, that's what I look for; if a search engine said it did excellent work in the field of medicine, that's what I'd test it for.
The bottom line is that this isn't a 'clean search engine' and to say that it is - well, they're just flat out wrong. It's also not a very good search engine - if there's no consistency with search terminology that's poor. If it blocks words or phases which are totally innocent, that's worse than poor, it's downright incompetent. Feel free to check it out for yourself of course, after a while it's quite fun to see how you can confuse it, especially with latin or slang terms (bristols are fine, boobs are not). My advice to these people - 'you don't have the first idea on how to create a search engine. Close it down and go away and do something else instead.'
Edited to add: If you're taking a look at the search engine now, it's changed, quite dramatically. I think they worked hard on it over night. The search option now links directly to a Google custom search and they've just gone straight across to the Google page for the custom search. Image search has also changed, and there's no reference to the unacceptable terms - you just don't get anything now at all! They're also using the Google safe image search as well, so a search for 'blue tit' defaults to just 'blue'. My 'dogging' search does however still pull up some unsavoury images. So all told - if all that they are using is a Google custom search with a safe filter on, you may as well just go there. (Added 30 April 2013)
Clearly, they haven't a clue...
Meanwhile, here's a picture I took while dog-walking today of a blue tit (honest)
http://imagingstorm.co.uk/wing-mirror-blue-tit.html
See if that gets filtered...
Posted by: David Bradley | April 29, 2013 at 07:32 PM