O.O'Brien

The City is extremely built up, with skyscrapers and tower blocks in all directions. This means that a GPS receiver was always going to suffer, with signals bouncing off windows – the so-called “multipath” effect, caused by urban canyoning. I noticed this when carrying out early experiments to see if I could map the City from “first principals”, i.e. using a GPS. Walking past the Gherkin produced particularly dramatic effects. In theory, running should reduce the effect but it will still be considerable.

Nick Barrable has uploaded his GPS track from the City Race, and also drawn his actual route. The difference is striking – although some of the discrepancy may be due to the way the GPS track is georeferenced with the background map, I suspect a lot of multipath signals are happening too.

Nick’s GPS is sure he went around the other side of the Natwest Tower. It’s in blue, the actual route taken is in red.

In the narrow Venetian-style alleyways around Cornhill, the errors are large.

In the Barbican, the GPS track is way off.

City Race competitors may have noticed the unusual way that the controls were mounted. Controls going walkabouts is always a worry for urban events, and although the City is largely deserted during the weekend (apart from teams of construction workers and window cleaners, it seems) it was still a concern. The biggest worry was uninformed security guards and police officers, rather than vandals, destroying the controls. The City is a high profile and high security area, and beeping boxes of electronics, with flashing lights, attached to skyscrapers, would probably raise a few red flags.

At Oxford and Stockport, marshals were at every control site – the safest way, and a good way to explain what is happening to the public, but also a real challenge in getting the man-power out – non-orienteering helpers need persuading to sit out on a street for two hours in (possibly) the rain, while orienteering helpers are probably needed to run the assembly area. Non-orienteering volunteers were used at Oxford, and were given a free T-shirt as an inducement. At Stockport, the numbers of controls used were minimised. In both cases, the volunteers were highly visible, and I found myself generally spotting the volunteer and then seeing if I could see the control – perhaps not ideal if the volunteer is not on the feature itself.

A marshaled control in Stockport. Photo by orienteeringmccann.

In Edinburgh, metal cable was used to secure the controls to immovable objects. This was very secure, but made punching slightly difficult, as the cable went through the punch hole. Newer SI cards have a fatter tip and so would find punching even more difficult. In fact, you just need to hold the tip at the top of the hole for it to register, but competitors might still be put off by the inability to do a firm “punch”.

At the City’s test race, at Queen Mary University in July, I used thin (1.5mm) metal cable, and bolts, for the controls in Mile End Park, which were considered more likely to be vandalised. A number of useful points came out of this – firstly, the time taken to secure the wire was considerable – requiring a wrench to tighten the bolts. Also, pre-cutting the wire without measuring the tree/post diameter is not good. One of the controls was around a bush, that was too wide for the wire to fit. So, I ended up looping around a protruding branch. Some youth spotted this and simply looped the control off the bush and into the nearby lake. Thankfully this was retrieved by an orienteer, so didn’t affect the race too much.

My initial thought was to use thicker metal wire for most of the controls in the City Race, and have marshals for all the controls in areas likely to be popular with passers-by. Planning and controlling reduced the number of such controls from around 10, to 5 and then just one, which was good. However, on testing out the wire at Glover’s Wood, the previous weekend, we found that it was very difficult to get a satisfying punch, particularly with the fatter SI cards. The wire tended to bend down through the hole, as it was looped, rather than being flat against the side, reducing the space even further. Another solution was needed.

At the same time, Vince (the assistant organiser) suggested that instead of using time-consuming bolts, we use Gripples – these are metal blocks where, on putting the wire in, it won’t come back out – a bit like a giant cable tie. The grip is excellent, making it very secure. The advantage is that putting the control up is much quicker. The disadvantage is that the wire has to be cut to release the control at the end – there is no way to take apart the Gripple.

In the end, we used a mixture of 2mm and 3mm wire – we had a limited number of Gripples of different sizes to fit each gauge. During the planning phase, an estimated wire length needed, was made.

The breakthrough was discovering steel cable ties – these work much like the plastic ones, but are much stronger (and more expensive). I found a local supplier and bought 100 Hellermann Tyton MBT11S 30cm ties made of “Class 316” stainless steel, for £28. The metal is very flexible, so easy to loop, but also very strong – I’ve been trying to break a test one apart, using pliers and also metal-fatiguing it, and failed. They can also be cut with the same wire cutters that cut the wire, and are flat, so don’t obstruct the control hole. On a couple of controls where we ran out of Gripples, we linked together several of these in place of the steel wire. For all the others, the cable tie was looped through the control, and through the steel wire, to which was also threaded the control flag (hung down from the string), a plastic back-up pin punch, and a laminated, punched A6 notice explaining to passers-by what was going on, with phone numbers for the organisers.

The notice was particularly important as, at the Rat Race a couple of weeks before, a control was found by a policeman and removed – its whereabouts only discovered after the police phoned the SI unit manufacturer (whose phone number was on the control’s base).

The Gherkin control – this one did fall apart as the Gripple was faulty, so the wire has been wrapped around itself. Photo by Chris5aw.

We also stationed marshals at three of the earliest controls. These happened to be the ones I thought were most likely to be moved on by security guards, but, also being the first ones, the marshals could help out new orienteers with the way to the next control. Because the controls were still secured, the marshals only went out at 10am, and could go back once most starters had passed, rather than waiting until course closing time. Marshals were given bright red hoodies to wear, so that they looked official, and as an incentive.

On putting the controls out, we had two teams of two – one with pre-cut lengths and one using a roll of wire. The sites weren’t tagged – as, when visiting them the previous evening, they were either in parks with locked gates, or generally surrounded by loads of Friday night drinkers – in a couple of places, the street/alleyway concerned was completely blocked by them. So, the control teams had a manifest, with a photo of each site, the control description, an extra detail (e.g. secure to right hand side of the left bench) and the wire lengths. The two controls which we weren’t able to secure – one because of its location and one because of a landowner request – went out later, at 10am, as courses opened, and were then continuously manned.

The mechanism was a success, with no controls moved. The cables, cable ties and notices had to be discarded, but the Gripples, controls, punches and flags could all be retrieved for future use.

A few points for future use – the control hangers may want to use gloves in future. Most of the hangers suffered point cuts from the frayed steel cable – the way the cable is threaded through the Gripple meant that bare hands tended to get jabbed by the wire. Also, having teams of two was definitely useful – one person could assemble the items around the wire, while the other person gets each item and threads the cable tie. It was easy to get the control box the wrong way around, and difficult to correct it once mounting. Punching the control from the wrong end doesn’t work, so some controls will have needed to be rotated around for successful punching (although no one complained about this). Finally, one wire cutter is needed for each person collecting controls back in. We only had three, so we could only have three collectors. The lack of stakes makes collecting large numbers of controls easy to do with a small bag, though.

This control probably should have been a bit more accessible. Photo by maprunner.

Looks like the finishers’ refreshments tent has arrived, fantastic. 😉

Also, listen to BBC Radio London between 8:30pm and 9pm tonight – the organiser will be talking about the race.

There are five Oxford runners and five Cambridge runners on the Men’s Open course at the City Race.

An early varsity test – who will win? Which team will do best?

…and how will both teams do against ULOC?

With all three universities out in force, it could be a revival of the southern rounds of the Churchill Cup, which used to be a competition between OUOC-CUOC-ULOC.

This is one of the questions that the organiser asked me a few days ago.

Here is a short script, written in Javascript, which, when run from within Illustrator, will count and list the number of occurrences of each type of symbol (or SymbolItem) showing in the visible layers. As a bonus, it also counts the number of paths (i.e. areas and lines) although these are only broken out by the type of colour associated with each, and it also prints out the total area – which should be taken with a pinch of salt, as it gets messy with subtracted areas and areal holes. It is extremely basic, but I couldn’t find an equivalent script on a cursory trawl of the internet, so here it is.

(If you uncomment the commented-out lines, then you get an animation while each object is examined, however this does massively slow down the speed at which you get the result, so may be painful for large numbers of object. It does look purty though.)

//Count number of objects and areas/lines in AI CS3

if ( app.documents.length > 0 ) { 
noOfSymbols = app.activeDocument.symbolItems.length; 
inc = 0
symbols = {}
while (inc < noOfSymbols)
{
  thesymbol = app.activeDocument.symbolItems[inc]
  tname = thesymbol.symbol.name
  inc++
  if (thesymbol.layer.visible)
  {
    if (symbols[tname] == undefined)
    {
      symbols[tname] = {}
      symbols[tname]["name"] = tname
      symbols[tname]["count"] = 0
    }
    symbols[tname]["count"]++
    //thesymbol.selected = true
    //redraw()
    //thesymbol.selected = false
  }
}

noOfPathItems = app.activeDocument.pathItems.length; 
inc = 0
area = 0
pathItems = {}
while (inc  0)
    {
      area += thepathitem.area
    }
    else
    {
      area -= thepathitem.area
    }
    //thepathitem.selected = true
    //redraw()
    //thepathitem.selected = false

  }
}
  
res = "Count of currently visible items"
res += "nSymbols:nn"
for (var i in symbols)
{
  res += symbols[i]["name"]+" - "+symbols[i]["count"]+"n"
} 
res += "nPaths:nn"
for (var i in pathItems)
{
  res += pathItems[i]["name"]+" - "+pathItems[i]["count"]+"n"
} 

res += "Area: " + area
alert(res)
}

 
Here’s an example output:

Adobe’s AI CS3 Javascript Reference was very useful.

To answer the question… nearly 800.

Five days to go.

Hopefully it will be a bit drier than it was today.

Some building outlines from the City of London map that debuts in 7 days.

I’ve been using Google Docs and Spreadsheets to manage the planning for the City Race. There are some very “cool” features in Google Docs – one is the visualisation widgets for spreadsheets. They are extremely easy to use – just click and place them in, and they visualise your data, updating dynamically.

Here’s a quick visualisation of the locations of everyone who’s entered the City Race so far. This took me about two minutes to do – one minute to concatenate the appropriate address columns in the entries CSV file, and one minute to paste it into a Google spreadsheet, add the widget and watch it geo-visualise. The geo-coding from addresses to locations is all done on the fly.

Europe:

England:

(Images from Google Maps – Map data © Tele Atlas)

p.s. Here’s another Google visualisation widget, for the number of Men’s Open class entrants. That course is rather full now…

This afternoon, I was around, checking the control for the City Race. Rather than take pen and paper, which would have been the traditional thing to do, I took the iPhone.

Prior to going out, I created a spreadsheet in Google Docs. In the spreadsheet, I named the first workbook “Controls” and added a number of column titles in the first row, to it – Control Code, Description, Securing Details, Owner, Photogenic Potential – etc.

Google Docs are not yet editable on the iPhone, however it turns out that forms are – the forms are a simple HTML form that loads really well on the iPhone. I added my own Google email address as a collaborator for the workbook – under the “Share” tab, choose “Invite people to fill out a form”. This automatically creates a form webpage, with a field for each column heading – and adds a “Timestamp” column to the original spreadsheet, recording when each form entry was added. Then, by going to the iPhone and clicking on the emailed link, I can start entering the data, in a structured way.

I then cycled between each control point. At each one, I brought up the forms webpage, filled it in and hit “Done”. Then, I took a photo of the location, which the iPhone automatically adds geo-tags with its built in GPS, so that I know roughly where the photo was taken at.

On returning home, I now have a fully populated spreadsheet containing the details and observations that I made while out during the survey. No need to sync the iPhone with the computer – the spreadsheet is already on Google’s servers, each row having been updated straight over the 3G network every time the form was submitted. (Getting the photos does require a sync.)

One definite advantage of this method is you look a lot less suspicious tapping away on the iPhone screen, than furtively scribbing on a notepad as I have done previously. I might even try and finish the checks during a weekday, which would normally impossible due to the high number of suspicious security people and general passers-by – such is the City these days.