Therefore I’d like to kindly ask anyone to sign up, even if you are not single, and answer the questions that are asked to you by the dating assistant.
The menu on the left is not active yet and you will not be able to use more than one photo.

If you are in a relationship or you want to withdraw your data later this year, please let me know. I would also encourage you that if you have a partner that he/she can also try the system. It would be nice to see whether you will be matched to your partner or not (matches are not yet visible to the user). Also I like to know any feedback you have, especially concerning the functionality or user experience. The project is almost finished and I am working on documenting the work. After the write up I hope i can use the feedback to improve the web site and to add more functonality.

The sign up page can be found at: http://www.aidating.co.uk/signup/

If you want to log on at a later stage you can do so at: http://www.aidating.co.uk/login/

Note that most effort is gone into the GUI and the matching algorithm and that contacting other users was not the primary concern during the period of the project itself. You can obviously expect this functionality soon since it is necessary part of every dating site. Also not that it is just a prototype and I therefore do not claim any specific results yet. You can leave any feedback regarding the site in this thread.

As an example I was working on a photo uploader. After all, what constitutes a good first impression? Is it intelligence, no. I rather think it is plain old attractiveness. Probably the reason why so many people fell in love with those überattractive celebrities. And probably also the reason why they have so many relationships that don’t work out (or is it just that the media is too much focussed on those break ups?). Attractiveness might constitute to a good first impression, falling in love, it does not help much for a long time stable relationship. Well I am loosing the main point now. The photo uploader. I was uploading a photo in the GUI, and Ithought it worked out. The next steps were to generate a thumbnail as well and to add a reference to the database. After generating a thumbnail I downloaded them to see what they looked like. Terrible! The image was completly distorted and not even recognised by some photo software. So I spitted through all the code. Seeing what could have gone wrong. Maybe not the right dimensions, a wrong setting that reduces the quality so much. I kept sending debugging. Till I downloaded the originally uploaded photos and realised they were as distorted as the thumbnails. So the problem was actually not with the thumbnails at all. However uploading and copying is quite basic in the back end. I wondered how this could have gone wrong. After all this code was highly similar with some online resources that described the process. So, I looked into the javascript for the uploader. Maing that javascript code was a pain after all. I looked into content-types and form encryptions. Nothing worked out. And then finally, Euraka!, I found out what was wrong. The ftp program downloaded the photos in ASCII mode instead of binary mode. No errors in the script at all. I lost at least a day for debugging something that was just working. I got the clue when I was reengineering the uploader seperately under the html root (where things are visible in the browser). And it just appeared to work. So I made it more similar to my real application and it was still working. So I knew I had too look somewhere else for the error.

The lesson is, we make mistakes. Always plan double the time you would rationally need to implement something. Some people might have said, just drop the photo uploader as a requirement. It has nothing to do with AI after all. Other people might say, test every aspect of whatevery you do. Testing is not always a bad idea. After all, I was testing and found out it didn’t work (or it did work actually). However testing everything you do, what some scientists do, takes at least 10 times the time you need for implementation.

But  happily it works now so I can put more effort into the main part. The matching process. The interaction between user and the questioning software agent. The fuzzy logic.

One in a hundred singles that had a partner once found it through online dating.

The reasons for singles to visit online dating sites are ‘curiousity’, ‘fun’, ‘to get to know someone’ and ‘boredom’.

A quart of the online daters that used online dating one or more times did register for a paid dating site at least once.

The majority of online daters describes their experience with online dating as ‘good nor bad’.

One third of the online daters managed to get at least one relationship through online dating.

Almost a quart of the registered online daters arrange a real life meeting after the first email contact.

According to four out of ten of the registered online daters that made a meeting in real life, the last meeting resulted in a one night stand or sexual relationship. In 7 percent this meeting resulted in a relationship.

Findings from: TNS Nipo, ´Online dating: aantrekkelijk of niet?’, ocktober 2004, for Parship.nl.

So I looked into authorisation systems. Some scripts are now secured using the apache authorisation module (.htpasswd). However for the user pages I needed an authorisation script that is secure and provides people with a login and sessions so they can stay logged in for some time. Many scripts were available but I decided to go with the Auth extension of Pear. For which an installation of Pear is also a prerequisite.  So today I installed PEAR and I am experimenting with the offered authorisation options and how to implement those. Good authentication is a necessity especially since AJAX requests are quite vulnerable to misuse. Tomorrow I will look into how to store the user data. PEAR is also used for the database abstraction layer and might be used for other functionality that will be implemented.

Creating a secure PHP login script
Creating a secure PHP Login Script
Pear

The business of finding that special someone has become ridiculously profitable for dating companies. And it doesn’t matter if you are a big fish or a one-man band.

Consider the proliferation of dating sites like PerfectMatch.com and MingleCity.com. TV commercials have been cropping up for Chemistry.com and, for a hot minute, I noticed that Dr. Phil had hung his shingle up at Match.com. Meanwhile, Marcus Frind, founder and CEO of Plentyoffish.com, has managed to take the free dating site he’s been running out of his apartment since 2003 to the top (Number one in Canada, number one in the UK and number three in the U.S.) using artificial intelligence. He probably figured, why buy the cow when you can develop an AI to run your site for pennies on the dollar? Frind, a 29-year-old computer programmer who appeared on The Today Show last summer, works a measly 10 hours a week and rakes in net profits of about $10 million a year. Clearly he’s struck a nerve worldwide. So, from what I can see, dating and the search for love represent a culture on the rise that doesn’t appear to be slowing down any time soon. (read further)

So it was very hard to come up with a name. The name should be original, easy to remember and it should somewhat cover what the web site is about. So many things came to mind and were usually taken. So for the moment I have the names “AI Dating” which covers what it is about but it is a bit tech savvy and I’ve got the name “Dating Atmosphere” which is not too bad but it is quite long for a domain name. Dating Atmosphere was actually derived from datingsphere, but that was taken as usually. With the sphere in the name I wanted to represent the idea in the web application that you might be in the middle and potential dates are in spheres around you based on functional proximity (match percentage). Possibly this will be shown in such a way in the GUI if there is time to develop it that way. This is one of the aspects that is novel and not seen on other dating sites (afaik).

So for the moment the dating site can be found at: aidating.nl (for a Dutch audience), aidating.co.uk (for the UK audience) and datingatmosphere.com (possibly for an international audience).

If you have any suggestions for a nice name that has not been taken, please let me know ;-).

Ever since the beginnings of the internet researchers have questioned its utility in developing and maintaining psychological healthy romantic and sexual relations. Advocates of the social presence theory and media richness theory purport that in leaner media individuals can be quite aggressive towards one another (e.g., in the form of flaming). In contrast, others believe that the lack of traditional cues in CMC can in fact be overcome and instead lead to more personal, intimate relationships.
As this paper will demonstrate, when we consider how beneficial it is to form relationships
online we also need to consider individuals’ characteristics (e.g., personality characteristics and physical attractiveness), the amount of time people spend online, the duration of online relationships, and how these relationships effect individuals’ offline activities and relationships. Overall, the view here is that online relationships can be empowering for many people; that is, cyberspace provides a unique environment for people to experience and learn about relationships and sexuality. (source: Computers in Human Behavior 24 2008)