<span class="gmail_quote"></span><br><div style="direction: ltr;"><div></div><div style="direction: ltr;"><span class="q"><span class="gmail_quote">On 5/9/06, <b class="gmail_sendername">Lars Aronsson</b> <<a href="mailto:lars@aronsson.se" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
lars@aronsson.se</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Nick Hill wrote:<br><br>> If we are thinking of creating a system which can determine<br>> valid routes, we will need to determine valid routes for<br>> specific modes of transport.<br><br>Do it the <a href="http://Amazon.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
Amazon.com</a> way: People who have been here, traveling at<br>your speed and direction, have later ended up there or there.</blockquote></span></div><div style="direction: ltr;"><div><br><br>Yes, but I think the big issue that Nick is alluding to is the quality of the data and its source. We need substantial metadata to be able to use the track data for routint - if it is to be reliable. For example - we need to know if the data has been collected from buses in bus lanes, from couriers in vans or bicycles. From the E-Courier data alone we could get data from vans, cars, bikes and bicycles I think.
<br></div></div><div style="direction: ltr;"><span class="q"><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">For every point and speed vector (within some margins), you need
<br>at least 10 observations in order to make an accurate prediction.
<br>You need a lot of track data, such as we have for Oslo and London.</blockquote></span></div><div style="direction: ltr;"><div><br>I think we also need quality assesment of the tracks, maybe include DOP values and covariance in the metadata to give an indication of the reliability and accuracy of the data. I like nick's ideas re filtering out the crap data and averaging the good stuff. Time of day would also be a major factor. I system like this could have the edge on other routing systems as we would be able to continually asses journey time estimations based on, as you point out, the current time for the particular time of day.
<br> </div></div><div style="direction: ltr;"><span class="q"><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">(A Swedish web bookstore tries to make recommendations without
<br>this threshold. For many less common titles, it only gets
<br>ridiculous. People who bought "Practices of an Agile Developer"<br>also bought "The Concept of Anxiety" by Soren Kierkegaard,<br><a href="http://www.bokus.com/b/097451408X.html" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
http://www.bokus.com/b/097451408X.html
</a> )</blockquote></span></div><div style="direction: ltr;"><div><br><p><a href="http://pagead2.googlesyndication.com/pagead/iclk?sa=l&ai=BrF-vR4ZgRKP0O8LUkwGb9KSpBP20zwnVzuivAcCNtwHQ4TgQARgBIIaPgAIoBECKFkiGOVDw1PHeApgBpo8GoAHalYz_A6oBJ0FjY291bnRBZ2UxMjB0b0luZmluaXR5K0NvbnRlbnRPbmVib3hDVrIBCWdtYWlsLmNvbcgBAdoBMGh0dHA6Ly9nbWFpbC5jb20vNzlncW9wb2R1djd0ZDl5MXdwcHp2MWNibm80ZXF3M5UCP9MkCg&num=1&adurl=http://www.healiohealth.com/tek9.asp%3Fpg%3Dproducts%26grp%3D82" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
Magnet Therapy For Life</a> Heal pain, bruises, sprains, etc. Feel Better Now! 100% Guaranteed - The advert that gmail gave me from this email!<br></p></div><br>Nickb<br></div><br>
</div>