A questions and answers site for numerical analysis

It always surprises me that here are so few resources on the internet for numerical analysis. As a discipline it is incredibly popular amongst younger engineers, there are very few graduating engineers that I meet that don't want to get involved with some form of numerical modelling as soon as they leave university. It is also a discipline that attracts people who like answering other people's questions as well as people who spend a lot of time on the computer and like to be at the cutting edge of technology. So why the lack of resources? I'm not sure.

We do now have a chance to change this. One thing I can applaud the Stackexchange system for is trying to create communities, not just a simple website and now a proposal for a Numerical Analysis is making steps towards becoming a reality. It still needs your help though. Because the aim is to create a community, not just a Q+A site they need people to sign up before opening up the site. If you are into numerical analysis such as Monte-Carlo analysis or Finite Elements, why not take a look at the proposal and commit to supporting it in the future.

If you want to take a look you can do so here: http://area51.stackexchange.com/proposals/1907/numerical-modeling-and-simulation?referrer=7jKSnwANyc0

Digital data management for Search and Rescue

Search and Rescue following a major natural disaster has a wide range of threads, and one of them is managing data. With hundreds of rescue workers on the ground, all searching buildings, using up resources, and requesting assistance tracking all of the information can get very problematic. It is a task well recognised in the International Search and Rescue (ISAR) community and the ability to run a coordination centre is considered to be an essential of all significant ISAR teams.

Now running a coordination centre is much more than just data management, but you do have to manage your data. You have to know which team is where, you have to know which sectors have been searched and which have not. Not only does not knowing this mean that resources could be very inefficiently deployed but it can risk lives. You have to manage your data. 

In the team that I belong to we have known this for a while and have been searching for our ideal solution. For us, what would be brilliant is simply a few iPads, a few iPhones and one or two central machines. The iPads and iPhones can be take out on a reconnaissance and when the ISAR team returns to base camp they sync with the main computers and presto we have a complete log of our reconnaissance. We have always had a financial limit to this approach, so its not going to be something we can implement overnight, but equally we have always found a technological limit. Syncing is a real problem when writing applications and the effort we’d put into writing our own application just didn’t seem worth it. Well the good news is now there is great solution, Bento, Apples consumer database application now supports syncing with multiple devices. All we have to do now is create a database in Bento, sync it to our devices and we are good to go. 

With that in mind this is my first database. It is based on a standard form for logging building reconnaissance and at the minute is fairly crude, but it does work and makes managing the data coming out a search a breeze. Feel free to download it and use it yourself if it is useful. I am going to look at creating a few more forms in the future which might end up on here. It also tuns out Bento has Applescript support so for those members of my team who have a deep love of Google Earth I am also going to look at a few scripts linking Google Earth to the Bento database which I will post here was well. 

You can download the Bento template for the database by clicking here. One thing to note is that it currently does not appear to be possible to back up you Bento database on the iPad or iPhone by email the data. You should always have a backup of your data. 

Don't obsess about your gussets

I was out training with the team this weekend and we did a bit of training with timber shoring. One thing struck me during the training that I hadn’t noticed before. Many USAR workers working on shoring are obsessed with gussets. cleats and nail patterns to the point that they don’t think enough about the main purpose of the shore. 

There is a hierarchy to the importance of components in a shore. You could describe them as primary, secondary and tertiary. The primary components are those that fulfil the essential purpose of a shore. For a T-spot these would be the post and a pair of wedges. The post carries the vertical load and the wedges make sure the post is tight and vertical. This is all that is required for a simple vertical shore and provided it is done well it works. I know, because a simple post and wedges held up a building I was working on whilst there was a 6.5 aftershock. The key to this simple shore is quality, you need to make sure the post is vertical, and tight with nice square cut ends and is a good quality piece of timber with a low aspect ratio. 

A simple post and wedges is not that subtle though, sometimes you need the shore to do a little extra work and that is where the secondary components come in, the header and footer. These components contribute additional features to the shore, which are sometimes needed, and which are sometimes not needed. The header and the footer can provide functions such as a more controlled failure to the shore, as well as distributing loads better at the top and bottom of the shore. For these features, normally the quality is a little less essential than the primary features. They still need to be good timber, cut to a reasonably correct length, but if you are focussing too much on the secondary components and too little on the primary components then the overall quality of your shore will suffer.

The tertiary components include things like cleats, gussets and nail patterns. For most simple shores these are not essential. The shore will work without them, although they do contribute to the overall robustness of the shore. The problem is that these components are the most complex in the shore. They require the most thinking for inexperienced shorers so they have a habit of focussing too much on them. The reality is that the exactness of the nail pattern, the shape of the gusset does not matter that much to the effectiveness of the shore. The quality of the post and wedges though is essential.

When practising your shoring you need to allow team members to make mistakes with the tertiary components. If you teach them to obsess too much about the quality of these items then they will loose perspective about the essential components of a shore, the ones they must get right; the ones that are far more likely to result in the failure of shore. 

On the importance of handwriting in modern engineering

When I first started out as a very junior draftsman in a small structural engineering firm I got told how during the first 6 months of training as an architect you spend most of your time practising your hand writing. I’m not sure if this is still, or has ever been entirely true, but architects do have very good hand writing and you can only get it by practise.

So why is this at all relevant in the world of word processors and computer generated drawings? Why should we be concerned about teaching children to write by hand in the modern world.

Let me take a step back a bit. I’m writing this post the morning after a slightly tongue in cheek tweet by Frasier Speirs:

Fraser Speirs (@fraserspeirs) I started saying “handwriting is dead” in my presentations for the shock value. You know what? Nobody finds it shocking. Thu Feb 17 22:21:21 2011

So here’s the problem, people in education are saying hand writing is dead and yet when graduates leave university I now have to teach them how to write things by hand and do hand sketches. Surely in a world where everyone has a desktop computer far more powerful than anything available 10 years ago we shouldn’t have to do this. Surely tools like the iPad can fill the gap? Well, in my experience that just is not the case.

Modern day engineering revolves around the computer. Calculations, drawings, reports are all produced on the computer so why is it so important for engineers to still be able to work with a pen an paper. The first thing to understand about engineering is documentation and readability is critical. You still have to be able to refer to something you did a year ago and understand exactly what you did and why you did it. Even more importantly someone else has to do the same because we are always drifting from one project to the next, leaving our current work behind us. A quick calculation on the back of a napkin is absolutely fine in engineering, but make sure you keep the napkin and that you can actually read and understand what you wrote. Using the write type of pen on the right type of paper is not essential, being able to understand what you did is. The problem with poor quality hand sketches and poor hand writing is that no one can read them. It is just basic pride in your ability to communicate and all engineers should have that.

So if neatness and readability is important, why not do it on a computer why straight lines really are straight. The problem here is that most engineers think whilst they are doing. A classic way of thinking about a problem for an engineer is to sketch it and work out the problem whilst they gradually sketch it up. When in conversation with another engineer you will often see both engineers gradually building up a series of sketches that both help them think, and they ultimately form their record of the conversation they are having. The same goes for calculations, they are gradually built up with careful thought going into each line. Yes some work can be standardised and moularised, but thinking about engineering as a linearly process with a predefined route is a mistake and a very common source of errors.

The problem with using an computer tools is that they just are not suited to this way of working in every case. The reality of the engineering thought process is that it is hard. It needs concentration and careful steps, with care and forethought going into each step. My experience is that I see far more errors in work done initially on a computer than work that was originated by hand. There are many reasons for this but my opinion is that the two biggest are the mental effort it takes to use a computer for engineering and the problems caused by standardising too much.

To consider the effect of mental effort we need to understand what a good set of calculations might look like. Good calculations usually include text, guiding the reader through the thought patterns of the author, the calculations and formulae themselves, sketches explaining visually what is going on, and detailed referencing of source material, all tied together with quality assurance documentation. The problem is there is no tool that can do all of this. Yes, you could do this all in a word processing or spreadsheet document, even the sketches, but how much more mental effort and time does it take than if you just used an pencil and ruler. An the problem is that all the effort that is going into driving the computer is not being used in solving the problem. The problems I see when people originate calculations on a computer are basic ‘schoolboy’ errors that should have never been made, but you just are not as in touch with the work you are doing when you don’t do it by hand.

The standardisation of calculations and drawings is potentially an even bigger problem. It is so tempting to set up a spreadsheet that does every aspect of a complex calculation so you just change a few numbers and out drops the answer. The really great thing about this approach is if you can get your spreadsheet validated, not only do you get the answer quickly, but you know it has been calculated correctly. Or at least that is the impression it will give you. The problem though is that people just don’t check through the calculations. If your calculations are 10 or 20 pages long and you are being pushed for an answer urgently, it’s tempting to just make sure you have put in the right numbers at the top of the sheet and just to make sure you pull out the correct ones at the bottom. Everything in between should be correct because it hasn’t changed. In principal that is true, but all too often the assumptions and simplifications made in the middle of the sheet are forgotten and these can lead to massive errors when the sheet is used for something it was never intended for. Yes, calculations should be modularised, but they need to be broken into simple standard calculations where the author can self check the numbers at the start and end of each stage. This approach ensures that the author has the time and space, as well as the tools, to think about whether the correct answer is being derived at each stage.

So to me being able to work neatly and accurately by hand is important. The big problem though is actually not one of skill but of the lack of confidence that the lack of skill generates. It takes months or sometimes even years of effort to convince modern engineering graduates that sometimes the best way to do something is by hand. I hear all sorts of excuses of why they should be using AutoCAD for this, Word for that and Excel for the other. The problem is, what it boils down to is that they just are not comfortable with their hand writing and sketches. They feel like it is too scruffy, they are not capable of writing neat enough, that nobody can read their handwriting. It is a massive block that we have to work through and fundamentally it is just because they are no longer taught it in schools they way it is used. Schools do need to teach a lot of things in the modern world, it is a complex and involved place, but I really hope that they can learn from the real way that work should be done in engineering and keep hand writing and hand drawing as key skills that should be taught.

Edit:

I just found this interesting little snippet in a blog post on how and why software engineers use hand drawn diagrams when developing software. One issue raised by a number of readers of this post is that things have changed and I’m not repesenting the way modern tools are used by the younger generations. I think this post demonstrates this still is not the case even for software engineers. Skils in basic hand sketching are still essential in communicating your ideas effectively.

Token - a neat little app for iWork

I have just found a neat little application for users of iWork on the Mac App Store. Token is a simple quick launcher for iWork documents with the added benefit of autosaving documents and giving quick access to scripts. 

Take a look at if you are a heavy iWork user and let the developer know if you have any feedback.