Spherical Objects

When I designed the fencing tool I did not envisage very long fences. In fact it started as a solution to my frustration in fitting fences around buildings. Well one of my testers, Martin, used the tool to put a perimeter fence around a large airport. Some of the fences are over 2 km long. When adding an another joined fence (the next side of the airport) the technique is to use the end point of the last fence as the start of the next. This ensures that the new starts at the same co ordinates as the one it joins to. Problem is it didn't work! We had gaps and overlaps. I'm talking small errors here – less than one meter in 1 km but it sure shows when you want a joined up fence!

A bit of testing showed the reason to be a bearing problem. The tool took the direct bearing from start point to end point and used that to place each individual fence panel. Over short distances this does not matter – over long then it seems that spherical geometry takes over. When I changed the code so that the bearing is re-calculated each time a fence section is added then the problem goes away. Very long fences are therefore actually slightly curved but I'll give a prize to anyone who can spot the curve……

Documentation

LOM/SM 0.98 is in testing now. I have started to write new manuals. I suspect that writing the documentation will take as long as writing the software! At least it is started so maybe in a week or so the new version will be ready for release. I am pleased with the way the new version has turned out. Not perfect and I am sure there are still a bunch of bugs in there somewhere, but it is a step up from the last version.

Latest Beta in Testing

Hmmm two posts in one day.

The fill area tool now works by filling a user defined polygon with a selection of objects. A bug in the line object tool has been fixed. Assuming that nothing nasty is found in the next day or so I will definitely start on new manuals

Back to Work

Got back last night – cut my vacation short a day after the engine controller on my boat became faulty.

So I am back to work on the upcoming release of LOM/SM. Right now I need to finish the code for the new Fill Area Tool. This tool is intended to fill any polygon area with a random selection of objects. It is hoped to be more powerful than the last version which could only handle a rectangular area facing North.

After that there are a few bugs to iron out and then I can get on with the New Manuals

Vacation

I am going to be away from my computer for the next week. I will be on my boat. At this time of year the weather should be perfect – unfortunately no one told the weather so it is not that great. Anyway with limited storage space I had to decide between extra alcohol or my laptop – the booze won so I will not be looking at any code, emails, etc etc until one week tomorrow

Slow Progress

Well I have not made that much progress this week. I am on vacation next week and have a contract to complete (in my day job) before I go. I might get a bit of time tomorrow to complete and test the new fill area tool code which works on polygons. If not then It will be a week or so before I can get back to it. I will be our of internet contact – my boat does not have satellite communications 8-). I should be back in town on Sunday April 23rd.

Writing the Manuals

Well with some very valuable help from my friends Scott and Martin the next version is about ready for the manuals to be updated – or should I say re-written. It is one of the dilemmas of the software design world – function rich (and big learning curve) or keep it simple but with limited functions. LOM had its roots in simple but it is now quite a complex program – definitely not a learning curve free tool. Creating something of a swiss army knife for scenery design leads to lots of different functions. Figuring out the best way to describe how it works is a problem – go with a traditional manual describing what each menu or function does, or write tutorials. Also as a freeware product there can be more time in creating the documentation that in creating the program. Well I think it has ended up rather much as a traditional manual at the moment. My current thoughts are to treat the documentation as a series of tutorials on how to use the programs. This may mean that some bits get left out or possibly not covered in a logical order but Tutorial form it is (well right now anyway)

Positioning error is fixed!

This has been really bugging me so I went back into the library I created to access FS data via FSUIPC. I originally wrote the code when I did not have access to a 64bit integer data type and I found I was ignoring the last bit of position data.

With VS2005 I have the 64bit integer data type and when I modified the code the position now seems pinpoint accurate. The reason for the variability in position error was caused by me ignoring a bit of the position data which varied from nothing to a meter or so.

I am going to leave in the bias controls for objects but I do not need a project correction any more so that can go

Getting Closer

I keep finding more things to do with LOM/SM. Right now, with a little help from Gernamy, I am trying to fix the problem with different countries using different decimal separators.

I have been wondering for some time why there is a small offset between the placed object and the co-ordinates that are used to place it. I guess it must be somewhere in the way that I get and send information about co-ordinates. I thought it might be just me but a beta tester has also reported it. It doesn’t matter too much if you are placing larger objects but it is a problem in placing things accurately or attaching a fence to a telegraph pole.

Where my test site is (Florida) this error is about 95 centimeters in latitude, though I see a different error in different places. It seem consistent by location. At the moment I cannot find anything unusual about the data – I am using FSUIPC and the standard data. My current solution is to have bias parameters which can be set for each project. FS9 has a bias element which can be used to offset an object from its co-ordinates in all three planes. At first I couldn’t see much use for it but now I have added bias adjustment controls for scenery objects, windsocks, beacons, effects, fences and bridges. I can now place an object to within 1 cm in all three planes. I use the Finney CrossHair ‘aircraft’ for design work and this has graduated scales in all three axes so it is easy to see how much to change the bias controls.

I am also working on giving the user the ability to decide which point of an object should lie over the placement point so for example with a gate you might want one end to be on the placement point. By default placement points appear to be at the center of the object. The program now lets you select the center, center of the sides or one of the four corners. Not really tested it yet but it looks promising.

The other BIG problem are the manuals. I need to completely re-write them and can’t make up my mind whether to put everything into a tutorial style or stick with an ‘operations’ manual.