Embedded Systems Engineering Standards Column vol 12.6 September t2004 Show time. By Chris Hills

I have to give the usual disclaimer that that these are my own personal views and not those of the ESE Editor and publisher. or those of my employer at the time..... I now work for Phaedrus Systems Ltd!

There is a 75% cut down version for print in ESE. The full one is on my web site at www.phaedsys.com Which you are now reading.

 

A mixed bag this month after the relaxing two weeks in the Sumner! Testing is like art, THE Show, ESS and are you sending far more than you know in an email? Also the answer to the Dam busters riddle

 

I visited a client whilst on holiday last week… OK! I know what I said in my last column about taking a break!  What I did was a client visit at the other end of the UK (Scotland) and extended the trip by a few days to actually see the place I was at. You know what “holidays” business trips are... I once did a three-day trip to Paris and did not get out of the hotel where the conference was. Thanks to the Metro from the Airport to within 50 metres of the hotel I only saw 50 metres of Paris. Not what I had intended or what most non-business travelers think of business trips.  This time I spent four days actually seeing the countryside around and about the area. Though some of you might know that we have had some of the worst summer weather and flash floods on record in Scotland so it was hardly the relaxing scenic trip I was expecting and due to some unforeseen delays (floods and road closures) I missed the one place I wanted to see, Rosslyn Chapel.

 

The person I went to see at the client’s offices was a tester. He had a small team and they were looking to gets some tools and systems in place to test the systems they develop.  They were still using methods I spoke about as being a decade old in my Embedded Debuggers presentation five years ago. At least they knew, or rather felt, there had to be a better way.  Recently there have been several people call me and there have been several discussions on the embedded NG on the net about formalising testing. Unit testing and integration testing in particular.  The problem with embedded systems is that you have to test them non-intrusively on hardware where there is often no room or resource for test code. This is where simulators and Emulators come in. However they require careful setting up and a planed approach to testing to make sense of it all and really test things properly.

 

By coincidence I have been reviewing a book on testing (Software Testing Fundamentals by Marnie L. Hutchinson http://www.ideva.com/Software_Testing.htm. The review will be on the ACCU web site, www.accu.org, shortly) that has some fascinating information in it. The Author has been in testing a long time and runs courses. She also runs an on going survey and presented the results thus far in the book. You will be please to know that in the UK, unlike the US most of the testers are qualified (and to degree level at that) in computing or electronics and are part of a formal test team. Whereas in the US it appears a lot more “relaxed” though surveys are only as good as the size of the poll and the cross section you get.

The other very important point the author makes is, as an artist in a former career, that software is like art. However the point that most programmers miss is that whilst art is about the expression of ideas all artists, no matter what branch, have to work hard with a great deal on discipline on the techniques used in their art!  Dancers go to dance classes, actors to RADA etc. My son as a Graphic Designer spent three years at university learning techniques for his art. As he said flair and talent can only go so far and some times, apart from wasting a lot of time can give the wrong results.

 

A few years ago an acquaintance of mine in the print publishing world complained that with the advent of the desk top publishing systems (ie the Amstrad Word Processor) “everyone” thought the could turn out flyers, parish magazines and company advertising.  He showed me some examples and asked which ones I liked. I chose a few. Without exception they were all professionally done. I did not know why I did not like the others but he told me. It was the way some things had or had not been done that was correct in the professionally prepared examples but not in the others.  As he said these are techniques taught to professionals as part of their training. They are not something you pick up in casual use, unless it is a few decades of use.

 

In a similar vein the editor of a user group magazine, after a decade, got some professional production help for a year. He also reported that he was surprised about how much he didn’t know that he didn’t know.  We all knew something had happened as the magazine was a lot better but it vas very difficult to say why. We just found it more readable. It was all down to the techniques used in layout and the way the tools were used.

 

The author of the testing book said that at one company everyone started in the analysis department.  If you did not understand the domain no way would the let you write code for it. After awhile you could progress to programming (with the proper training or experience). The best programmers were promoted, eventually, to testing.  Testing was where the masters were.

 

Where am I going with this?  Testing should be done by your best educated, trained and disciplined people.  Those who have trodden the long path of learning the techniques, who have the discipline, who recognise solid engineering that is used innovatively from software that is working but “lucky”. I once spent several hours unraveling some “clever” software that had a subtle bug in it.  In the process we found some other hidden problems caused by sloppy programming that was masked by these “clever” techniques the programmer used.  A better programmer would have used clearer and clearer code rather than trying to be flash.  I have seen many “shorthand” and compact methods of writing C that take up less room on the page or screen used by these “clever” people where the compiled result is the same size as a longer source code version that is easier to read.

 

However rigorously testing is a discipline that takes time. Having readable source code and a system you can actually comprehend is one thing. You also need a good test spec and test automation.  I have a paper on automating testing and I am putting together a long overdue paper on Design and documentation

 

If your testers have been through the analysis and design phase they will be able to work with the designers to get a proper specification for the project and a test specification.   If you look at the V model, or better still the W model then you will see, contrary to most programmers expectations coding is a relatively minor part of the process…. Actually most case tool vendors who’s tools to auto code have been saying this for years. It’s just that their tools are not that good in the 8-bit field (yet) but they seem to be making inroads into the 16-bit field these days.

 

So for testing what you need is a good development system, I don’t mean an IDE, and system and design specifications to test against.  Where possible testing should not be done by the coders. If you have coders who look down on testing and think it is beneath them I would worry. There is a lot more to SW Engineering than coding. There is a company I know who has managed to automate their testing system using proper techniques and engineering. It took them a while but they have gone from 30% test coverage to over 90% test coverage and what is more it is fully automated. The “fun” work is now in the design and testing areas.

 

 The Show: The Embedded Systems Show is here again.  This time it is at at the NEC Birmingham.  At the junction of 2 motorways (M6 and M42  that link to the M1, M40 and M5), the main line to London and Manchester and an International Airport. For locals (Birmingham, Coventry etc) there are lots of busses.  It must be the easiest embedded show in Europe to get to this year.

 

Here is an overview of what's going on in the Exhibition Forum so far. Everything in this section is free and will be of practical use to engineers and managers if not on their current projects then on the next one.  For session abstracts please go to http://www.edaexhibitions.com/ess/conference.htm

 

Wednesday, 13th October
The Object Management Group presents: Open Discussion - MDA in Real Time: Use and Misuse
EEMBC presents: Processor Performance: Fastest Isn't Always the 'Bestest'
MISRA presents: The Launch of MISRA C2 (more on this later)
Interactive demonstration: Man-Wearable, High-Performance, Real-time Visual Computing Solutions

Thursday, 14th October
VDC presents: Selecting an Operating System…This should be interesting especially if you can get the speakers to hang around for the next presentation!
The IEE Presents: Panel Discussion: Why Pay for an RTOS?   I can see this one being fun!
BAE Systems presents: Automatic Code Generation - Case Studies.

Is the programmer dead? The keep saying this but the better these tools get the more auto-generated code there is.   See what the case studies show, does it work or only in some situations.

All these (and others) are in the free Exhibition Forum so you can wander in and have coffee and a bun, rest your feet and even join in…  Though I suspect the “Why Pay for an RTOS” might get heated. We, the MISRA-C working group, will be answering questions on MISRA-C2

 

The whole point of this Free Forum is to get engineers to mingle and discuss what is going on in a less formal atmosphere than a normal conference. This is why it is directly off the main show space.

 

 Many Engineers in the embedded world work in small companies where they are isolated from the rest of the industry… Actually it happens in some of the not so small companies too!  Going to a show like ESS (well there is only one show like ESS in the UK) lets the average engineer feel more connected and part of the world.  However managers always feel that Engineers who go to these shows will job hunt.  I have found that this is not the case. For every one that does (and lets face it if he is going to do that he is looking to get out anyway) the next 99 will come back excited at all the new ideas and things they have seen.

 

The other point that managers worry over is company secrets or project plans. In the last decade or two of attending these shows I have yet to hear of an engineer divulging the company secrets. What does happen is that they suddenly see something at the show that provides solutions for their problems. And it is just as often an idea as much as we need to buy xyz.  This relates to the 2 weeks in the summer column in the last issue.   Change is as good as a rest and in this case a change from the normal routine to ESS where ideas and discussions abound (as well as all the companies who’s tools you have only read about) is going to really boost your engineers. I was recently at a conference in Heidelberg where a delegate (from a major company) said to me that he got as much out of the discussions at breakfast, lunch and dinner with the other delegates as he did in the conference itself.

 

The other fascinating set of lectures is the PhD forum. Now I though that this would be a bit esoteric al, but no… Go and have a look in the ESS web site and you will find that about 80% of these “academic” thesis (thesis?) are being presented in co-operation with the company that sponsored them, this is Engineering research (not academic science) it is useful in the real world.

 

MISRA-C: Don’t forget the MISRA-C2 launch. I will have the first copies available to purchase (cash, cheque or credit-card) on the stand. You can pre-order your copy from and specify Misra-C2) but it won’t be dispatched before the 13th of October. It will be the same price as before £25 which is not bad as it is about 20% larger. You could of course come to the launch and meet the team, ask questions and see the presentation and the show in general.

 

Please be careful about claims by tools supporting MISRA-C (either 1 or 2) There been the odd “interesting” claims made.  Some Marketing departments push their luck.

 

In the mean time I have set up www.misra-c.org which will have information on C1, C2 and the test suite as well as (I hope) a web forum.  It should also have pointer to MISRA-C checking tools and MISRA-C aware compilers and resources. www.misra-c.org will also track the test suite progress.  Whilst it is an unofficial web site (eventually MISRA/MISRA will get their act together and an official site)  It will be possible to ask questions of the MISRA-C team, indirectly, and I will post an FAQ. There will be a MIStRAy-C2 soon as well (see http://mistray-c.phaedsys.org) . My only problem is some people think it is a serious guide (as they see much of the “rules” in code at their companies) rather than things to avoid!

 

Email is replacing most other forms of communication in the computing world, and in most parts of commerce too. Fax and letters have almost died. Like letters and fax email is “connectionless” you don’t need the other person to actually be present present when you send and email, unlike a phone conversation.  The other useful feature is you can send to several people at once. Conference calls never seem to work satisfactorily…. I SAID SATISFACTORALIY… hello? HELLO?

 

There are two major problems with email the first is obviously the spam (persevere with this as the second problem is also important). Either the junk advertising, get rich quick schemes or the viruses.  Earlier this year (2004) an acquaintance who works for a major ISP told me that at one point  90% of the email one month was spam!  At this point I go all conservative and suggest flogging is too good for them and we need the death penalty for spamming and viruses.

 

On average spam is about 50% a month, approximately. It is approximately because no one is sure.  What is spam? Everyone now has filters and firewalls etc but as many companies have found their monthly newsletter is not getting the response it once had. Product and seminar announcements are getting a lot fewer replies. People are so overwhelmed by the volume of spam that the useful gets thrown out along with the rubbish….  What I can’t understand is why the spammers try to defeat anti-spam filters. The filter is there because I am quite happy with the proportions/dimensions of my body (and my wife’s, unless they are suggesting that I need a boob enhancement!)  I don’t want to meet “local” girls from “The Valley” somewhere in the USA or Russian brides who are dying to meet me for romance (and a passport). As for various drugs, I am sure customs and excise will be more interested than I will when they arrive. For Spam I have no solution other than filter and delete.

 

The other major problem is volume in bytes (or octets probably) At one time all email was ASCII plain text.  Now, thanks to many people-using Outlook, html, rtf and other forms encoding of email are used. Why?  Send yourself an email in html and then look at the message size…. I regularly get emails where the plain text is 1k bytes but the html is 4-6kbytes!   Then there are the embedded graphics.  Not the attached documents and pictures that are often sent with useful information but the background graphic and the company logo etc.  Many emails are 80% html formatting and pretty graphics that serve no purpose.   On top of which they often get snagged by virus filters.

 

Several companies send me beautiful emails of many colours and subtle graphic effects highlighting their wears. When they get to my mail reader they are an unreadable mess of text and raw html.  On problem is that many, conscious of the size of the graphics will embed URLs so they automatically pick up the graphic from the web site. Just like a virus email does though it usually accesses more than an on-line graphic. NB I don’t mind the embedded URL’s that you have to activate manually. It is the automatic ones I don’t like.

 

Marketing like these emails as they look like the old advertising flyers they sent out on paper when the send them. If we could just get these idiots to go back to plain text we could cut 90% of the volume of these emails.  Just think how much stress this would take of you company email network?   

 

I have found a simple reply saying “Your email unreadable and was deleted. Company policy is to only accept plain text emails please do not send HTML, RTF Unicode etc” usually works. Try it.   I have found that many companies now give you the option of plain text.  

 

The bulk spam we will have to leave to the ISP’s but if we can get our own emails back to plain text we can probably decrease the volume of junk in our own company networks, if not the Internet, by a considerable amount.  

 

 The Dam Busters and showgirl’s legs…Well the prize for the correct approach height for the dam busters raid was claimed by Mike from a well known aerospace company “Up North”

 

Dear Chris,

With respect to the Dam busters, the original planned height was 160 feet, which was manageable, but proved to be too high. Barnes-Wallis decided that 60 feet above the water should just about do it – about half the wingspan of the Lancaster bombers. Pythagoras' spotlights had been developed during World War I to determine the altitude of a plane, so I think the music-hall scene may have been apocryphal. I certainly cannot say how many there were in the line-up. Incidentally, one of the planes got shot up and lost its comms & instruments so returned to Britain using its lights to fly low-level across the Zuider Zee. However, the alignment had been changed by the flak and the beams did not converge until the protruding bulge of the bomb made contact with the water! The pilot was able to drag the plane out of the water - losing the bomb - and return successfully to RAF Scampton. Cheers, Mike.

 

I also had an email on the subject from Murray Kopit , New Westminster BC Canada which surprised me as ESE is not distributed in Canada…

 

So we now know that the height above the water was 60 feet and the mugs have been dispatched. There are still some ASCII -HEX mugs for anyone who knows how many stocking clad legs there were on stage in that scene. Which turns out to have been gratuitous stockings according to Mike as the lights technique comes from WW1.

 

Anyway drop in and say “Hello!” at ESS. I shall probably be chained to the stand from much of it apart from being in the MISRA-C2 Launch. Otherwise any excuse to get off the stand and have a coffee and a bun any feedback or input would be gratefully received

 

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