MindFeed

I am not one to get annoyed easily. However today I spent an hour trying to figure out why my web template for my profiles section worked fine in every browser apart from IE. Now clearly I understand IE has issues and doesn’t like real code and over the years I have been quite understanding that Microsoft write bad web browsers that do not work with correct code. However today I found a new bug that I had not encountered before. By having two div’s floated left, the last letter of the text in the right most div was repeated by IE outside the div on a new line for no reason at all. My code after much checking was perfect. This baffled me for sometime. I then very luckily found that if you provide a comment before hand eg:

<!– Include other stuff –>
<div id=”profileColHolder”>
<div id=”profileThinColumn”>
<p>thiin</p>
</div>
<div id=”profileThickColumn”>
<p>thick</p>
</div>
</div>

This will cause IE to produce:

<!– Include other stuff –>
<div id=”profileColHolder”>
<div id=”profileThinColumn”>
<p>thiin</p>
</div>
<div id=”profileThickColumn”>
<p>thick</p>
</div>
</div>
k

Notice the k on the end for no reason at all. All I had to do was remove the leading comment so that it looked like this:

<div id=”profileColHolder”>
<div id=”profileThinColumn”>
<p>thiin</p>
</div>
<div id=”profileThickColumn”>
<p>thick</p>
</div>
</div>

And now the extra k has gone. Now forgive me for being slightly annoyed but why on earth can I not put comments in my code!? They are supposed to be IGNORED anyways! This is just plain stupid. Someone needs to invent Double Glazing and put Microsoft Windows out of business by fixing all the annoying bugs they have.

Finally got round to updating the live feed system with a new sexy Ajax version. No more horrid page refreshes and instead we have a smooth auto updating Google map with zoom controls and choice of hybrid, satellite, or regular map options.

MindFeed now has a fully functioning contact form. W00t! Send me some love.

I have also realised that the latest version of Franson BlueTools seems to crash the whole application upon discovery of certain Bluetooth devices. I am not sure why, so I have gone back to the previous version which is much more stable and works. However something to consider if anyone else out there is just starting out and is having issues. I have sent a request to Franson to look into this issue.

The mobile client has also been improved:

- Now has option to delete XML files after upload to avoid cluttering up and saving space.

- Can now specify output location (under options) of the compressed ‘upload.zip’ so you can locate it and upload it easier when submitting your recorded data. Some folders on WM6 are not accessible from a web form upload component!

It occurred to me last night that cookies can easily be spoofed. I am not sure why it has taken me so long to realise this but thankfully I have. Thus, to ensure someone doesn’t just create a cookie with someone else’s user name and use that to log on, I have created a system where by the value stored in the cookie is based on randomness and this is then matched with a value in the database.

The cookie also contains uniqueness ie if it was based on randomness alone, there may be the occasion where the same random number is generated and thus 2 users would have the same cookie which is bad.  Thus to ensure randomness it is paired with other details that are unique. The exact details can be found in my report.

Now that the mobile client is pretty much finished I decided to test it out again on a trip to London. I am glad to say it worked, though I ran out of money on my mobile phone for the live feed half way! Oops.

Some photos are below:

Today I have successfully configured Apache to use secure connections when sending data to and from the site. This allows me to send users data in a secure manner considering the sensitive nature of it. I have created my own signature that I have digitally signed myself, and thus browsers will pop up a message “are you sure you wish to trust this site” as it is not signed by an official certificate authority. However, everything works correctly, and this could easily be rectified by getting the signature signed by an official authority, though I do not have the money to do this currently.

I have also modified the mobile client to use https in all of its URL’s now too and configured PHP sendmail to work so I can set up the contact form on the website.

Today was spent filling in wordy pages with meaningful information. Now I can concentrate on the rest of the website such as the API, profiles, applications etc now that the groundwork has been finished for the rest of the site and is at a respectable state. Once again all pages are valid XHTML.

I have also started designing some of the databases required for MindFeed. Live feeds are now integrated with the database and automatic detection of current live feeds has been added so a listing of all available live feeds can be produced. Need to integrate this with privacy settings later.

Check out the latest version of the site at you guessed it http://www.mindfeed.co.uk/

Stuff to do in no order:

• Live feed map - change to ajax version rather than static image.
• Implement API
• Construct some example applications eg review uploaded trips, find stalkers, general stats etc
• Implement user profiles - allow users to upload photos and customise their profile withs some basic information.
• Implement user privacy settings
• Write basic help articles

I have spent the last couple of days tidying up the website (all XHTML valid) and implementing the registration form and login form. All is working as expected in all major browsers.

The registration form is easy to use and is fully secure against SQL injection and validates all input both using JavaScript and server side PHP. The user can see at a glance if input is valid with visual feedback via ticks and crosses to save them having to go back and resubmit a form that contained errors.

Now to add some meat to the website, content etc.

I have also made a start laying out the thesis and business plan.

• Now deals with all major NMEA GPS strings  Now have satellites, more speed updates, alternative latitude/longitude updates if other NMEA strings do not exist  and HDOP measurements are also available from many NMEA sentences too.
• The mobile client now detetects when it is possibly “indoors” and thus can calculate when to automatically start a new trip log upon the gain of a new GPS signal. To do this we utilise idle time from when we loose a fix, until we get a fix again eg > 2.5 minutes start new trip.
• Auto refresh of GPS device to account for buffering bug I have discovered. If the GPS receiver is indoors for too long, it buffers up useless data, thus when we go outside again it takes a while before we get valid GPS data and thus is not good for live feeds as it is always x minutes behind! Hopefully this is now solved by disconnecting and reconnecting to the receiver every 30 seconds to flush any buffering.

Asides from that I have also found an annoying issue with the Bluetooth GPS receiver I have obtained - it auto turns off after 10 minutes of idle use. This wouldnt be a problem usually but it defines idle use as not having a GPS fix thus if you are indoors for more than 10 minutes it turns off, which is annoying. Solution: Buy a GPS receiver that does not turn off automatically.

Today I have added the following features to the mobile client making it very nearly fully ready for deployment:

• Integrated time interval settings (from options menu) for GPS recording to file every x seconds or “All data”
• Application now gets set saved settings on startup and populates the options form.
• GPS accuracy signal bars have been added to show the Horizontal Dilution of Precision (HDOP) and Average HDOP values for GPS data. This also adds to the user feedback experience so they can see what is going on too.
• I can now ignore very high HDOP values to have a commercial grade GPS recording device that is much more accurate than before.
• Changed timer for live feed to initially be 10 - 15 seconds to allow page to initially load.

Once again, a good days work, just about to go out and test it to make sure I have not created any new bugs!