These are my own personal views and not those of my company Phaedrus Systems see www.phaedsys.com which is where the full version of this column, with links etc, resides under the Documents tab.

 

Well that was the summer. I trust you all had a good break and relaxed, recharged the batteries.  The good news is that parts of Europe are coming out of recession so things are starting to move.  Of course autumn is the time for the trade shows.  To met people and see what is new. Humans need to network, but face to face not just twitter and facebook.

 

Even in a recession this industry moves faster than most and the shows and conferences are awash with new things, new presentations and even some old favourites like coffee and doughnuts!

 

I mentioned a new idea for beating the recession: Shoot the accountant! I have, both as a developer and recently as a tool distributor, had many discussions with Engineers and project managers about tools. Initially on feature sets and capabilities but eventually the conversation turns to costs. 

 

I don’t know how much this effect is seen in other countries but in the UK it is the accountants who not only handle the money but have a strong influence on how it is spent.  They define the structure of capital spending and what is on which budget and how it is costed. This has a serious effect on the engineering.

 

Several times recently development teams have been looking at tools for a particular architecture that they will use for the current and future projects.  They have done their homework and say that we can base a whole family of products for the next few years on this architecture and also save by reusing code, not just for the main sections but the libraries for controlling the peripherals too.  However the accountants want to cost each project as a self contained item. They do not seem to be able to say this is capital equipment like a lathe or desk.  Either that or buying capital equipment is so difficult no one tries.

 

This is most obvious where the choice of MCU is arm as the tools usually cover ARM7, 9 and Cortex.   ARM is the new 8051… If it fits the bill variants are likely to be used in other projects  for some time to come. Yet still the financial people want the tools costed per project… 

 

The other problem is different budget streams and cost centres.  A compiler may well cost 4K and another is 1K or in the case of some free.   However that is not all there is to it. I have several cases where it has taken the company 2-3K of time to get a “free” compiler to work and then on going costs to keep it running. There is the same problem with other things like USB stacks, OS, emulators and debuggers, in fact any tools or software.

 

However time and salary is on a different budget stream from tools expenditure. So one is still 4K and the other free, even though the “free” one has cost more in time and resources (Money).  More to the point in the current market the time is sometimes more important than the money.   Getting a robust project to market fast can make, according to some studies, over 30% difference to the profit.  So “saving” money on cheap tools or insisting they are costed against a single project is a flawed method…. Sometimes it can be fatal to the company.

 

I had a case recently where cheap tools were used simply because they cost less…  4 seats of them. However as the project progressed it was found the code was getting too big for the part and external memory was going to be required. Also execution performance was not good.

 

Out of curiosity one of the engineers revisited one of the “expensive” compilers they originally looked at.  The result was the code was not only 20% smaller but that much faster that they not only did not need the external memory but could move to a less powerful variant of the MCU which was cheaper.

 

As this was going to go into mass production in over 50K units per year the saving ram chips and a couple of GBP per unit on the MCU (not to mention time to design and test the more complex PCB) and the additional power required. It all adds up.   They moved to the “expensive” tools and saved a fortune…..

 

However, it had also taken them time to set up the less efficient system. That time they could not get back.

So when you are choosing tools there is a lot more to it than the headline figure for initial purchase. Also you have make the accountants see that many of these tools are for use on multiple projects. If you don’t others will and then you will never be able to compete.

 

The trouble is whilst engineers can do maths financial people do accountancy which is as close to 1+2 = 2 as F1 is to go-karting.  It is worth while Engineers understanding how the accounts and budgets work for their company. You need to talk to accountants in their language…   As a rule of thumb I tend to work on an engineer (in accountants terms) is worth at least 60 GBP per hour to a company.  That is pay, sick pay, holiday pay, national insurance, employers national insurance, pension, heat, light, desk, chair, administration costs etc)  Note: accounts need to be reminded they are a net cost. They cost money and produce nothing.   So a quick way to get more profitable is to sack them…

 

An In Circuit Emulator at 8,0000 GBP is worth about 3 man weeks. If it saves three of you a week each over the life of the project it has paid for itself. More to the point it is still there for the next project. Time, or getting a robust product to market is often as important as anything else. Saving a month of salary is one thing. Getting to market a month earlier (or not missing a deadline) may be worth far more.

 

I mention Emulators as they expensive but can do wonderful things like on target unit and system test if run with scripts… so after the first time you have push button automated unit testing suddenly the “far too expensive” 8,000 GBP ICE looks like “the bargain of the century”.

 

The caveat is of course as a tool vendor “I would say that…” However, I don’t sell ICE but I have used them often enough to know that despite the headline cost they can save a company, never mind a project. Also that “Free” can cost the earth or in the long run your job. So you need to understand the way the finance and project control works. You also need to work on solid engineering and reality. Not myth and hype.

 

An interesting item from the BCS cam my way in their “whitelist 23” podcast  that should spread a slot of FUD http://www.bcs.org/upload/mp3/whitelist-ep23.mp3    The BCS did a survey that discovered that most people (though hopefully not BCS members who have to work to a code of ethics)  said they would, if made redundant, walk of with internal company data. Whilst company data often only has a short self life of 3-12 months or is hopefully covered by a patent, in the current climate a competitor, or even a new start up, only needs 3-12 months to make the difference.  

 

At one time it was salesmen with the customer lists but now many engineers have as much equipment at home as at work and with USB sticks and mobile phones having gigabytes of memory managers should think carefully about who has the knowledge and data to make the company profitable and who is a net cost centre…..

 

On a final point some good news and a good news story for “Open Source” in general if not Open Source Software in particular. The Guardian (16th June) ran an item that A new hydrogen-powered car, whose designs will be "open source" under the “The 40 Fires Foundation” open source licence (www.40fires.org)   see , http://www.riversimple.com/  was unveiled in London. If it is “Open Source” and the designs available to all it can not easily be bought out as others have allegedly been. The Genie is out of the bottle and hopefully this 2 seat city car will be seen around the world.   Even if it eventually fails it might actually motivate the big….. err…. remaining automotive companies into producing far better and affordable ecologically sound urban transport. 

 

Author Details and contact

 

Eur Ing Chris Hills BSc CEng MIET MBCS MIEEE  FRGS   FRSA is a Technical Specialist and can be reached at This Contact

 

Copyright Chris A Hills  2003 -2008
The right of Chris A Hills to be identified as the author of this work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988

 

 

 

> Robert Pirsig on the
  Metaphysis of Quality