If you can save your work when all about you
Are losing theirs and blaming it on Bill
If you can trust your scripts when others doubt you
And keep well hid your list of those to kill

If you can wait and not salute (three fingered)
Just ‘cos the hour glass goes on too long
Or being “blue screened” not give way to hating
Though many times the urge is oh so strong

If you can dream of programs that are faultless
Yet rise each day to play the usual game
If you can meet with Smartsuite and with Office
And treat those two imposters just the same

If you can watch the programs you have crafted
Abused by users till they creak and fall.
Or watch the project you have worked for, shafted
Because the poxy budget was too small

If you can make one heap of all your hard disk
And risk it all without a backup done
And lose, yet start again with Windoze boot disks
And never breathe a word when you get home

If you can force your PKZIP and DRIVESPACE
To get you through long after disks are full
And so work on when there is not a byte free
except a net-based drive somewhere in Hull

If you can meet accountants yet stay human
Or meet the boss, nor lose the techie touch
If office politics can never hurt you
And users count on you but not too much

If you can keep the unforgiving system
To 40 hours a week of uptime run
You are the guru, one in fifty million
And what is more, you’re in demand, my son.

May 2001

We get what we deserve in this country, we really do. Witness the Top Gear two-year-old car satisfaction survey when the most griped about cars are also the best sellers in the UK. It seems we don’t have the bottle to put our foot down, and so we put up with stuff that’s nowhere near what we want.

It’s a standing joke the way folk in restaurants in this country mutter about their food not being right, but when asked “Is everything OK with your meal?” they smile and nod politely. A friend of mine is an occupational therapist and sets up people having disabilities with equipment they need to get on with life, so she’s very good at spotting ergonomic nightmares at 100 yards. She is not small, but is certainly not as big as the average man, and explained to me that nobody makes cars that are appropriate for folk that size. Little things like sun visors that don’t come down far enough, high rear windows that she can’t see out of when reversing, and seat belts she can’t reach back to get hold of when her seat is fully forward.

And did you ever see a crash test performed with a dummy sat as far forward as a small driver will sit? (Come to that, why don’t the dummies in the official regulatory tests wear clothes?).

Totally inappropriate too are small cars for the ‘urban driving’ they claim to have been designed for. Aerodynamic styling is wasted on a town car. Why not make child locks that can be operated from the dashboard (for the ’school run’), and why have a rear offside door at all (it opens into traffic)? If the keyring is going to open the doors by some clever non contact system, then how about a hands-free self opening and lifting tailgate for when you have armfuls of shopping? Rounding the corners off the front of a modern car could knock eight inches off the effective external turning circle, improving parking no end, but I guess that just wouldn’t look right, would it?

It is said that women have the major say in more than half of non-fleet purchases, and yet this lack of user friendliness towards the average height female driver suggests either appalling market research ­ or a realisation that UK consumers are quite happy to get what they’re given.

I find the same thing in engineering sometimes. I often wonder just how bad a supplier has to be before some people will stop using them. When you meet a supplier consistently bad, it’s clear that they are only in business because some people continue to accept that level of service. Then you’ll hear those people say “yes, but they’re cheap”.

Perhaps it would be better to think of the extra spent with a company that gave better service as an investment in the future quality of our supplier base. It is worth the extra to get the message across to the bad supplier. If we want the market to evolve, the unfittest mustn’t survive.

I’ve always liked helpful guides. So here, for the first time, from a machinery designer, is the definitive guide to buying automation.

1. Show the supplier who’s boss. Refuse to discuss the project over the phone so that they have to travel the length of the country to find out what you want isn’t the sort of thing they do.

2. Never do a justification calculation in advance. Leave it to when the guy is on your site until you figure out that vision-guided robots in a clean room cannot payback a quarter of an operator four days a week.

3. Jazz up the spec. If you need 5 units per minute, round it up to 10. It raises the game a bit and leaves you with some in hand. Later, when you decide the price is too high, you can always change your mind.

4. Never actually write down your requirements in one place. A scattered series of faxes and phonecalls are far more fun and should stop the supplier ever really getting an upper hand.

5. Try and avoid the issue of an acceptance test. If you keep saying “we’ll know when it’s finished”, then you can guarantee hours of entertainment further down the line as people argue about what on earth the machine was supposed to do.

6. Never tell the supplier the true timescale. If you want a figure in next year’s budget, then swear blind that the order could be placed next week. Get it off your desk more quickly.

7. Supply the best product you have for their development work. Only bring out the really representative parts (the ones with the wildly varying dimensions) on the day of the signoff trials. That means the parts you supply at first should be either moulded parts all from the same impression of the tool; or with any other process, all parts from the same batch; or with strip and roll fed material, from an entirely different supplier.

8. Change who’s responsible at your end halfway through the contract. Appoint a recent recruit and cunningly change the spec at the same time. It’ll be weeks before they spot that one!

9. Make them hurry, right up till delivery day, then tell them suddenly that you aren’t ready to receive it yet. Keep ‘em keen!

10. Refuse to make operators available for training. When the machine is due, try to send the guy who will have to maintain it (and all the future operators) on a long course which starts on the day the truck pulls in.

11. Burn or eat all the manuals and troubleshooting guides the moment the installation team leave. Sack anyone who attempts to read them.

12. Crank it up to 20% faster than specified operating speed at the first opportunity to see what breaks.

13. Make sure you get a home phone number. Only try ambitious running conditions or preposterous throughputs when the supplier’s tech support guy is asleep, and when an operator known to his friends as “animal” is on your nightshift against his wishes.

September 1997

1. Since we actually handedover the money for the 3D solid modelling system, the suppliers have been far more attentive and responsive.

2. Setting up a new computeris so much easier than you expect it will be, and takes so much less time.

3. When you add up all the tolerances,it turns out we actually have less of a problem than we thought we’d have.

4. I love these computerised phone systems- you know exactly what you’ll get and don’t have to talk to a real person at all.

5. The job has turned out to be easierthan we thought; we’ll deliver your order early and charge you less.

6. It’s great to receiveunsolicited promotional emails. They keep you abreast of things without you having to go to the effort of seeking out magazines to read.

7. I wish I received morecolour glossy brochures in the post. I love the time spent opening the envelopes, figuring out what on earth I ticked to bring this upon myself, then wrestling with the guilt of wanting to bin them despite thinking that somehow they might maybe be useful to me despite being, on the face of it, utterly irrelevant to what I’m doing right now.

8. You (or your automated fax system)seem to have accidentally sent exactly the same order to us twice, but rather then idiotically ship you the goods twice, we have applied a bit of intelligence and recognised it was a mistake on your part. You’ll only be receiving one set of goods.

9. I do love websites which take ages to load.I always know when the modem downloads upwards of 500K before I get to see anything that I’m in for a real graphical treat that will make all the time waiting for it well spent.

10. I’m just calling to saythat since you are due to deliver our machine in less than a month we’d like to make absolutely no changes to the spec.

11. I think that Microsofthas demonstrated a very good grasp of security technology over the years and am utterly happy about the idea of US Naval vessels being fully controlled by Windows 2000. Would an accidental missile attack on Redmond be classed as ‘friendly fire’?

12. I love meetings.They are usually so productive.

13. I do wish I could get coffee at homethat was as nice as the stuff out of the machine in work. And where can I buy those great cups that seem to be made out of 5 micron copper foil that help warm your hands so well.

14. I think the Windows technical support modelis fantastic. Simply include ‘OEM’ in the product serial number and then make your distributors take the flak for every design fault (or charge 40 quid for a bit of phone support). I wish more manufacturers did that.

15. It’s far too much to askthat small planes flying to the Rhein industrial area from Luton at 7am should have power sockets at each seat for laptops. After all, there’s no reason to believe that a significant proportion of the passengers are anything other than holiday makers.

16. These days they makekids’ TV programmes far better than they used to. I remember the Clangers and it wasn’t a patch on the Pokemon cartoons.

June 2001

Implementing new ideas with as little fuss as possible is really, really important. Here, then is the definitive guide to bringing a new technology or software package into the workplace. Really.

1 Kill as many birds with one stone as possible. Bring in voicemail, email, hot desking and 7.30am workforce callisthenics all on the same day. That way, the disruption will be over as soon as possible.

2 Always assume when planning that the technology will succeed first time. If you are putting in a switchboard that allows customers to dial straight through to the extension of their choice, sack all but one of the switchboard operators five minutes after the system goes live.

3 If training is required, give it well in advance so everyone has been trained well before D day. Ignore people who say that this means folk will forget it before they get to use it, or that you won’t know if the training is no good until it’s too late. They’re obviously just wimps.

4 Always believe what it says on the software box about hardware requirements. If it says 4Mb memory is the minimum, then that’s all anybody needs. Any more should be regarded as a taxable perk. Any hardware upgrades required should be planned for the day before the software is installed. The IT department will cope.

5 When assessing hard disk space requirements, assume that all the old redundant packages will be removed before the new ones are installed. Tell anyone who suggests it might be wise to keep both old and new live for a trial period that they are being defeatist.

6 There is a pecking order for new hardware. Secretaries only need to view the boss’s diary, email, keep multiple word processor documents open and enter data into spreadsheets or presentations when asked, so a 4Mb 386 is quite sufficient for them. We engineers on the other hand need to be able to see Dilbert on the web, so we need far more Intel inside. The boss, who can’t figure out how to turn his on, needs a Unix workstation capable of finite element weather prediction.

7 Act completely surprised when the supplier with the pushiest salesforce (to whom you awarded the contract) suddenly disappears off the face of the earth after you make the final payment. Allude to a rigorous supplier selection procedure which you used but give no details to anyone insolent enough to ask.

8 When you’re deciding whether or not to upgrade, never even try to attach any tangible value to reasons like ‘we have to because all our customers will have the latest version’. Just leave these claims unquantified (like you do with anything to do with quality or strategy) and they will remain unassailable. You didn’t get where you are today by questioning the sanity of being at the ‘bleeding edge’. Nor did Bill Gates.

9 Don’t take any lip from the IT department. Everyone knows that the latest software is easier to support than the old stuff they’ve got used to. And what’s all this fuss about needing a faster network ? It’s only a load of wires…

10 Remind everyone that any grief the new system is allowed to cause just detracts from the end of year profit-share. It’s in their interest to cope. (Whingers !).

April 1998

They used to say that 75% of UK VCRs still had 00:00 flashinginstead of the correct set time. As Vic Reeves so rightly points out, 87% of statistics are made up on the spot, but I think we’d all agree that most technology out there is massively under-used. A classic example is Microsoft Office, whereby the majority of users now on Office XP could barely claim to have used more 20% of the ’special’ features of Office 97.

It’s interesting to compare off the shelf equipment with special purpose equipment in this respect. You might imagine that a special purpose machine wouldn’t be packed with features no one needs, since added features cost money, and the idea of a bespoke is that meets the spec and no more. Since a special purpose project costs more, you might expect that it would have a budget that would support a bit of genuine training. And as the equipment is targeted at a particular area in the customer’s organisation, you’d expect it to be used by a small number of people, so it should be possible to train them all to use every bit of it.

On the other hand, the user manuals and training aids supplied with a standard product can be amortised over a lot more units than the ones that are put together by a typical “special projects” company, and a product company (with its sizeable marketing department) probably has access to more suitable tools for preparing such literature or training aids.

So here’s a question: is a typical bespoke machine, in general, used by better trained staff than a standard machine?

I think there is a feeling across high and low volume products that manuals are mostly of little use to the people who need them. This is because, in general, the user does not read the manual from cover to cover before starting, but expects to use it as a reference text when he cannot figure things out any other way or when something goes wrong.

So should the ideal manual be written entirely as a trouble shooting guide? Probably not, but it should certainly be searchable. A hypertext format, which includes a search engine is almost essential and this leaves very few alternatives for how you implement it. It would need special dedicated program that the user installs which IS the manual (with accompanying CD), a Windows help file, a Word document with significant inbuilt programming, and a selection of linked HTML pages with Java applets to provide the search functionality.

The great thing about the last of these choices is that it can be implemented on a CD and be totally platform independent, and it can also be implemented (with no changes) as an actual web page.

But is the CD really worth giving out at all? Could we publish all manuals only on the web instead? And I don’t mean as downloadable PDFs you print out – I mean things you have to read online.

Apart from the cost savings, the advantage for the supplier is that it allows access to manuals to be tracked. It is then possible to see which aspects of the manual are and are not being used (or which need improving). It would also possible to build a ‘trouble shooting flow chart’ implemented as a series of web page questions, leading ultimately to an “if all else fails” form that generates an email to the supplier.

June 2002

I have spent the last month getting to grips with my first Palm OS PDA.Two things have struck me: the first is that the pen-based interface I always swore I’d never touch is actually quite handy if you have to take notes halfway across a departure lounge while walking with a hands free cellphone; secondly, the PDA is pretty under-featured compared with my good old British Psion Series 3A. The built-in database sucks, the built-in ‘to do’ list is almost unusable, and there is no built-in spreadsheet at all (though, of course, all these things can be bought from third party vendors).

But the upside? Well, it has USB connectivity to my PC, and complete synchronisation with the company’s contact management software. This means that I have 5000 customers and possible customers in my pocket at all times, plus my synchronised diary, job list, and a passable email writing facility. In other words, it is supported.

Though all these things are just about available for the very latest Psion PDA, a lot of things aren’t – such as files downloaded from Mapquest.com after a route search, or the up to date American Airlines flight timetable.

Psion, you see, made a key mistake with its PDAs. Its software ran on its hardware, but no one else’s. Its forays into licensing the OS came too late, and the Palm OS pretty much stole the market.

Some believe that the Microsoft mini OS that runs on machines like the iPaq will eventually mean the end for the Palm OS. I don’t think so. Palm has learned the lesson that for a computer to survive, the OS needs to survive, and that’s far more likely if it is licensed widely to other hardware manufacturers. And if Psion learnt this too late with its PDA business, then it has at least shown that it’s learned the lesson with its Epoch OS: hence the Symbian venture to put Epoch on the world’s cellphones.

But here’s a question: why hasn’t the same lesson been learnt in the industrial control world? Why hasn’t the “Palm OS” of PLCs been born and taken over? And before you say IEC 1131, that’s a standard, not an OS. I’m thinking more in terms of choosing hardware on reliability but knowing that the software platform will be what exactly you know and love whoever you buy from. And you’ll immediately know how to program it. Of course you’ll have a range of software options to control your hardware: there’ll be lots of IEC1131 compatible ladder-based offerings, and maybe a good Java implementation or two. And you’ll write code for one machine, and use it for many machines, even if the next customer specifies a different hardware vendor.

I think that Microsoft has tried to muscle in on this with its WinCE forays, but could you see yourself actually controlling a line of presses with a WinCE based controller? The OS I’m after needs to come from someone with a reputation for understanding real-time control, and must be rock solid. Maybe QNX or Microware could pull it off. But to do so they need implementations on all major players’ platforms, plus a good load of applications to launch it with. Perhaps the industrial market just isn’t big enough for anyone to dare.

And while the various vendors still sell on being different, perhaps there’s too much of a gulf between what the users really want, and what the vendors are prepared to offer.

July 2002

We really do get the service we deserve. In the same way people used to always vote Conservative for tax reasons after telling the pollsters they would vote Labour, purchasers say that service and supply reliability is their primary criterion and then go out and buy the cheapest.

The problem is, people don’t vote with their feet. If your cellphone company doesn’t give you the service you need, then you should walk. Vote with your feet, but make very sure to tell them why, and make sure the people who matter hear it. It doesn’t matter that the other phone companies are no better. They are probably no worse. The key is that the first lot have lost a customer and you have told them why.

In many markets we are faced with oversupply coupled with a lack of time, and that can lead us to make purchase decisions on almost arbitrary criteria. Well, it may not seem arbitrary at the time, but it often can seem that way in retrospect, when you realise just what the differences between the suppliers really are. We might not like to admit it, but deep down we know that it’s often true.

If you’re going to make an arbitrary decision, make the most of it. Pick a criterion that you feel strongly about and use it to narrow the field.

Hate what’s happening in Tibet? Tell your distributors you won’t buy products that are made in China. Care about the planet? Tell them you only buy from companies with a published environmental quality plan, and watch them scrabble for the documents. Care about exploitation? Tell them you need to see documents showing the location of manufacturing sites and assurances that there is no child or slave labour. Care about reliability? Tell them you want to see independent trial data before you believe them.

Over time this will actually have just as much effect as anything else to create the suppliers you want. But it will take a little bit of effort from you up front.

On an entirely different tack, here’s an interesting product that I think deserves a mention. ‘GoToMyPC’ is a system that allows you to connect to a PC from another machine elsewhere on the internet. Okay, so connecting to your desktop from another location is no so new, but what’s really neat is that you don’t need to have any software installed on the PC you’re connecting from. You just need a java enabled web browser, meaning that you can ’sit at your desk in work’ from a cyber café anywhere on the planet. It could be a Mac or Windows or Linux machine you are actually sitting at. The connection has the same 128-bit encryption security you associate with a professional Virtual Private Network. Because it is only the user interface that is being transmitted over the net, you don’t need huge bandwidth to get real work done, and because no actual data is transmitted, you have not created a security risk.

More details (including white papers explaining the technology in detail) are on the website (www.gotomypc.com). The catch (or the advantage, depending on whose side of the invoice you’re sitting), is that rather than buy a piece of software, you get the software that goes on your desktop machine for free, but the connection is initiated via the GoToMyPC website and you pay them a monthly fee per PC that can be accessed.

One day we’ll all be able to do our jobs from the beach…

March 2002

This month I’d like to rant about software, because it seems to me that the vendors want to have their cake and eat it. You see, we don’t buy software anymore ­ we are granted a licence to use it, but it still belongs to them. And yet they don’t seem in any way bound by law to offer any meaningful support.

What’s more (and this really galls me) they can discontinue a product (such that if you want to buy a few more licences of a product you have been using for a few years and which works OK, you can’t) and yet if you copy it yourself they can still prosecute you.

I recently tried to buy a few more licences of a product that I use every day. They had discontinued the product (when it last sold, it was (UK pounds)55, I think) but they graciously offered to sell me a piece of paper which granted me the right to make some more copies and install them ­ at a cost of (UK pounds)99 per extra copy.

But I guess that’s just good business on their part. And I’ll let them off because (like a lot of DOS software) it’s fast, stable and compact.

What bothered me far more this week was when a certain (well known) suite of new office software failed to install. (Or rather failed to reinstall to the hard disc, having been un-installed because the ‘run from CD’ installation had been unusably slow.

It turned out that the grief stemmed from a known bug, and the helpful guys in the appropriate Compuserve forum got me going within a day. The forum is not staffed by employees of the original software company, of course. But there was no attempt by the software suppliers to tell me of this known bug either when I purchased it or when I registered it. And not everyone who gets a piece of (UK pounds)300 world leading software home has lifesaving Compuserve access and a day to spare.

What’s more, further chat with the folks in the Compuserve forum made it clear that I should really install the ‘Service Pack’, (which is basically an official bug-fix) if I knew what was good for me. Again, I have also never been told such a thing existed.

Seriously, hanging’s too good for them.

Now, it strikes me that by the end of 1999 we will have the Millennium bug, the Euro and probably a new version of Windows all vying for the ‘computer nightmare of the century’ title (and my money’s on Windows, incidentally). But the Millennium one could be the messiest, because of the number of people it takes by surprise. And as a recent newspaper article which I read pointed out, it’s no good fixing your own systems if all your key suppliers go belly up in January 2000.

So I guess that being ‘Year 2000 Safe’ will soon be as much of a key letterhead claim as ‘IS09000′. By the end of this year, I for one am going to start asking all my key suppliers what they are doing about it.

And I’m also going to start planning my own personal Millennium preparations ­ move all my money from the bank to under the mattress where it can’t be deleted by a system crash, stock up with firewood for when the gas and electric companies fail, start ‘running down the freezer’, steer clear of phones, planes, trains and automobiles (and lifts, apparently!), and most of all, whatever else I do, I WON’T PANIC!

Griff has been final testing for Y2K at GB Innomech and claims the only non-compliant kit is the answerphone.

January 1998

...If you are going to make a significant change to a product, then change the *$!o&* part number…

I have said a few times of late that I thought that the whole ‘business to business selling on the net’ thing was missing the point. And this week has proved me right.

Just moments ago I read an article in the IMechE comic that was intriguingly headlined “How engineers are buying on the Internet” but went on to point out that the vast majority of net purchases by businesses are for office supplies. Hardly the ’supply train revolution’ that we keep being told is gripping our industry.

The point is, with most engineering equipment, the actual ‘placing the order and paying’ part of the purchasing transaction is a trivially small part of the whole exercise, and one for which my companies have already developed perfectly good, efficient systems that are not aided or improved at all by having to do something supplier specific on a website.

It is in the specification and selection (and post sales support) of components, not the purchasing, that we’re looking for time to be saved.

To all you suppliers out there, let me say this, which as much emphasis as ink on paper can provide: my time costs money – if you save me an hour in design and specification time, then you can add half the saving on to your component price and still be cheaper in real terms than your competitors.

But this week’s horror story bears a particularly important lesson for all suppliers out there. I’m not going to name names (though you can bet I’ve considered it), but this is what happened.

A certain company gave us some drawings in DXF form of its components. We devised a way to design mountings. Expensive parts were ordered. When the component arrived, a pattern of three through holes (that we planned to mount the part with) were nowhere to be seen. It turns out that there have been paper catalogue updates since we first obtained the DXF file (more than one in a twelve month period, apparently), but no DXF file updates. In this case, we’d have been better off never having set eyes on the DXF file. Lots of time lost, parts remade, and a significant bit of redesign of the overall machine.

I guess the key point here is that you should regard all types of product information – whether they be downloadables from a web site or a printed catalogue – as equally important, because your customers certainly will. If you give CAD files to clients, you should try and keep the major updates in sync with your ‘official catalogue’. Customers who have seen the CAD files should be entered on an update list so that they are reminded by email when the files are updated. Someone should have the job of signing off the files as correct and taking responsibility for them – rather than treating them like they are the product of some sort of basement skunkworks by the ‘techies’ who never come out into the daylight.

And finally, please, please, please if you are going to make a significant change to a product, then change the *$!o&* part number! At least then people with the old CAD files will stand a fighting chance of guessing that there has been a change, or at the very least will order an out of date part number and be told by your company that there has been a change.

One day we will get there, I’m sure. The paperless design office beckons…

September 2000