You can make up lots of silly examples, but the facts of how requirements work does not change. If your customer gives you requirements that are not complete, you always have the option of adding your own requirements. That does not change what requirements are and how they work.
yes it's quite amazing to follow this discussion and see this irrational logic, I don't know the origin of it, perhaps their upper level management/lawyers likes the lack of firm requirements, if something goes wrong they can put the blame on the customer spec.
But I think you have the answer to your original question, to many engineers the lack of firm requirements is seen as an asset, not a liability, so no attempt is made to correct the situation.
It depends only upon how the cost of a human life compares to the other costs.
Come on, get real. Every practical project requires at least 3..4 times more time and money to get it perfectly straight.
We spend a great deal of time in training
You may like it or hot, however the current situation is what the market is asking for. I completely agree to your point, however all of us have to deal with the reality.
Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
It can sometimes be left to the engineer if the engineer has specialised knowledge of the application domain. An engineer knowledgable of the application domain will already be familiar with the legislative requirements for products sold into that domain and would also know the standards to which he must work.
It would inded be a folly for the client to assume the engineer might know the application domain without first establishing the engineers capabilities and knowledge first.
It is not acceptable for applicable safety standards not to be met in any product. The fact that we still see such incidents and recalls is a sad reflection on the way some industries approach the issue of safety or management of safety related products (a large percentage of total products).
[%X]The legal (and legislative) circles will have at least as much bearing on the implied requirements as the "good engineering practice" has on the engineering requirements. Some of the clients policy of selection of supplier companies would, I submit, establish whether or not an engineering concern had the required capability and management procedures in place. Once established does this need to be in a specific requirements document. By virtue of their policy they have already preselected a shortlist of suitable engineering companies to supply them.
I don't recall in which country you reside so the laws where you are may be different from where I am (UK). The UK Consumer Protection Act has some real nasty teeth in it for those whose products cause injury to the general populace. I think the designer of an electronic piece-part of a certain brand of washing machine has found out to his cost.
-- ******************************************************************** Paul E. Bennett .................... Forth based HIDECS Consultancy ..... Mob: +44 (0)7811-639972 Tel: +44 (0)1235-811095 Going Forth Safely ..... EBA. www.electric-boat-association.org.uk.. ********************************************************************
I care for the outcome of my work a great deal and I think I know where Rickman is coming from in his "all encompassing requirements" requirement. I have even ripped up the clients original requirements document and sat down to re-write the whole thing (usually a much slimmer document) on which we have the essentials covered and can be assured of the eventual outcome in terms of functionality, maintainability, safety and environmental impact. It does not take a lot of verbage to get a decent requirements document together just clear and precise construction.
Management care about the bottom line and engineers need to be able to speak to management in the terms the really understand. Bonuses, after all, tend to depend on overall profitability.
I do not deny my training (it is much broader than electronics hardware, software and simple mechanics). I put quite a bit of it into practice every day.
[%X]
I would be very concerned about this myself. Any change to any part of the system (and test procedures are part of the system) has to be fully reviewed in the light of the requirements - safety and otherwise). It may have been correct that the test procedure was changed but unless a very thorough review of the change was undergone then I would consider it delinquent to implement the changed procedure.
-- ******************************************************************** Paul E. Bennett .................... Forth based HIDECS Consultancy ..... Mob: +44 (0)7811-639972 Tel: +44 (0)1235-811095 Going Forth Safely ..... EBA. www.electric-boat-association.org.uk.. ********************************************************************
There are requirements or other guidelines regarding the power consumption?
You'd have great fun with the standard construction contract requirement that all work will conform to the best practices of the trades involved. If it weren't for such ambiguous requirements, contract specifications would become impossibly large and the cost of your house could double.
So you think it is better to have requirements that have no meaning? I know some friends that were screwed on a construction contract of their new home. The builder specified "builder" quality and they did not like the painting that left brush strokes showing and other visible flaws. The contractor insisted that was what they agreed to. Obviously they needed requirements that were clear and understandable. It may be common practice, but that does not mean you will get you what you want.
It's just a matter of practicality. If your friends had specified every conceivable detail concerning the construction of their house, it
would have never gotten built, or would have cost a fortune.
We just had a custom house built. I doubt that our painter (who did not speak a work of English) would have agreed to a whole lot of "garbage" in the specification. He did an excellent job though, and we got exactly what we wanted.
You have to allow for different ways of doing business, especially when it comes to specifications. If you're buying a coffee pot, you don't write a 30 page specification (unless you're the US Air Force), before you drive over to Walmart.
-- Hershel
Yes, you are right in that there are many ways of doing business. But I don't think we can do our work using engineers who don't speak the same language that we do. I don't see how not speaking a common language helps you to get your doors painted correctly.
No, we don't write specifications for our coffee pots, even though it might have helped us get a better one. The unit we have brews individual servings of coffees or teas. If you are a tea drinker right behind a coffee drinker, you stand a good chance of getting coffee flavored tea.
It seems that even for something as simple as a coffee brewer, detailed, clear, thorough requirements help produce the correct product!
The courts seem to think that they have meaning.
Your friends didn't know how to deal with contractors who try to slip shoddy workmanship by them. Refusing to pay is the first line of defense.
"Builder quality" is a legal term?
I know the courts have language to deal with non-payment on a contract.
I read both of them as, "if it's over 1 kilogram you don't get paid, and if it's less than 1 kilogram you don't get paid more."
The only general rule of requirements-based projects is that exceeding the requirements without a distinctly negotiated reward is - or should be - a firing offense. You're giving away money when you do that, and spending time that you know will be precious enough that you could lose everything at the end of the run. So it's not necessary to say that it's okay or unnecessary to do a better job. But when you're writing requirements for a supplier, you write them with such ambiguities, because maybe you can induce them to give you those freebies.
I.e., when you see requirements like that, you're being cadged for a bonus. I don't think many people who write requirements really understand that; they just want to have a better thing made. But if they do understand that, it's righteous to be irritated that you're being asked to do more than you agreed to be paid for.
The technical problem, here, is that the second clause of both of those statements is not a requirement, and therefore does not belong in a requirement. So where do you put them?
A good requirements document has a section that gives a wordy, hand-waving description of what is meant, and then lists the actual requirements that must be met.
E.g., a made-up example:
========================================= Section X.Y.Z: Chassis
Description: The chassis ties the wheels, body, and motor together. It needs to be made of light materials and constructed for strength rather than inertia or style.
(sketch of chassis in situ with other parts)
Requirements:
---------------------------------------------------------------------------------- The chassis shall [Req. X.Y.Z.1] be less than
1 kilogram in mass.---------------------------------------------------------------------------------- [Req. X.Y.Z.2] Deleted.
---------------------------------------------------------------------------------- When force of 45.4 newtons is applied in an inward direction to each of the front arms of the chassis, the arms shall [Req. X.Y.Z.3] deflect not more than 3 mm.
---------------------------------------------------------------------------------- =========================================
You can infer a lot from the descriptive part about the item and its place in the system. There's no need to say where to put the wheels, for instance, if there's a picture in the description and a general requirement that the thing has to work and look like what it usually looks like. Though if you're subcontracting the chassis, every dimension will be specified in the subcontract requirements.
I threw in the trace tags and a sample of a changed req. for grins, or maybe because of PTSD...
In which case they're usually gibberish and you need to make tactfully sure the customer knows that they may not have specified things correctly or clearly. Getting buy-in for the "okay, just do it as you see fit" requirement is solid gold for a level-of-effort project, and a request to deliver half the value on a fixed-price project.
Which is why quality is all about feeding back information to the farthest reach at the earliest opportunity. Reqs aren't a goal, they're a communication mechanism in a process that has as its goal a product with a certain purpose and quality. But the way some managers treat requiremnts as biblical texts you'd think they'd never bollixed a delivery before.
--Blair
That might be true, but how much would it cost? How long would it have taken to come to market? Why did you buy the coffee maker in the first place - it's obviously satisfied the coffee maker maker's main criterion, to sell coffee makers. Not every need is satisfied by a detailed set of specifications.
-- Nobby Anderson
I was referring to "best workmanship practices of the trade" and similar standard clauses. "Builder quality" is specific to your friends' case and may indeed have no meaning, legal or otherwise.
Quite true, but they also understand withholding payment for non-performance and shoddy workmanship. Ambiguous portions of contracts are weighted toward the party not preparing the contract.
I most ceratianly can, and I do.
So?
When I worked at Mattel, we had the following requirements for products sold for infant use. (#2 was and still is unwritten but I assure you that it is a hard requirement just the same.)
[1] No part is small enough to cause the infant to choke to death [2] Must not be made out of the skin of endangered Pandas. (Thanks to David M. Palmer for the example.)Both are required. Do either and the entire production run must be corrected or destroyed. And both are important to Mattel.
"Tickle me Elmo" was tested for choking hazards, but was not repeat not tested to make sure that the plush cloth wasn't made out of Panda skin died red. There are an infinite number of requirements that are never tested for, and this is one of them. In the real world, whether something is tested for depends not only on whether it is important/required, but also on the odds of noncompliance.
I also worked on the space shuttle ("17 inch disconnect redesign for the main fuel tank, among other portion). The shuttle also has an unwritten requirement that the seat covers and spacesuit padding not be made out of the skin of endangered Pandas. And again, nobody tests to make sure the Shuttle is Panda-products free.
I can also call you someone who confuses opinions with facts... :)
That's true, but not everything needs to be verified.
Actually you said that you have to specify what is important, and then you assumed that everything that is important needs to be verified. Both statements are incorrect. Not using Panda skin as spacesuit padding is important, but it doesn't need to be specified or verified.
-- Guy Macon
You just called a perfectly valid refutaion by counterexample "silly". dismissing it with a bit of handwaving and once again claiming that your opinions are "facts."
If you walk away from this discussion with a changed attitude, you will have a lot more sucess in engineering and in life. If you insist on being ineducable, it will bite you in the ass again and again.
I advise less effort trying to "win" and more effort trying to learn.
-- Guy Macon
I suspect that the Air Force Coffee Pot spec was reasonable for something that has to survive rapid decompression and inverted flight without scalding the user...
-- Guy Macon
There are an infinite number of ways that a product can be unsafe. You appear to believe that they must all be listed in the requirements. Plush toy fur and spacesuit padding can be soaked in nitroglycerin and thus made into a high explosive, or they can be washed in water containing rattlesnake venom. Do you expect Mattel and NASA to test for those substances?
And of course, you can produce a design that met those requirements and still not be doing your job as a designer. That is what Rickman isn't getting. He would also say that an unwritten requirement isn't a requirement.
"Log" meets Mattel's specs for a toy. ("It's big, it's heavy, it's wood!" so no choking hazard, and only the deluxe version is carved out of sequoia.) So it meets the requirements, but it is not a good product.
If you were to add the requirement that it be Fun!!! then Rickman would object that this is not a valid requirement, because it cannot be objectively verified. And even if it could, Mattel should fire the designer if he made something that was merely Fun!! since it didn't meet specs, and also fire any designer that dared to make something that was Funny-Fun-FUN!!!!! since he didn't renegotiate with the customer for additional payments for exceeding the requied funicity.
Why dye it? There is such a thing as a red panda.
A few years ago you could buy stuffed toy cats at {Wal,K}-Mart, made in China. Very soft and cuddly. How did they make and sell them for so cheap? Well they had a good way of manufacturing cat-shaped pieces of fur--the main processing step involved emptying out the rest of the cat. The toys were withdrawn from the Mart shortly after someone in the US discovered their production techniques.
-- David M. Palmer dmpalmer@email.com (formerly @clark.net, @ematic.com)
Not using
sure it does, and is, you just seem not to be aware of it, look up the spacesuit specs, in the restricted materials or prohibited materials area, it will specify materials that can't be used in the suit, in addition all other materials furnished in the suit or used as part of the process of making the suit must comply with various regulator acts, including hazardous waste disposal, environmental, health, safety, and transportation compliance, panda skin would not make it through this (illegal to transport out of the country of origin for one thing)
you verify it not by test, but by listing all the parts in the suit and show that they all are compliant with the above regulations
typically a company will have a pre-approved list of materials that can be used to minimize this task, I'm sure Mattel does, and NASA does, so you try to make a product with these pre approved parts, if not the material must be verified by your parts procurement department before it can be used in the product, and proof of such is supplied to the customer
frankly I'm surprised you don't know this.....
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.