I am attempting to define exactly why (or if) MS office tools are inappropriate for professional engineering work. I would claim that for any professional work, in science and engineering, there are some basic functions that should be easy to work with in simple direct ways. Revision control, finding differences in data and documents and merging document edits all seem to be central to tracking and managing all documents. Data oriented tools should allow clear separation of process and data, and easy means of iterating and comparing the results over differing data sets. These operations have pretty questionable support in MS tools. Are these core requirements in your view? Are there other differences that are worth discussing?
To Mark Haefner:
Just a note regarding engineering: the word is derived from the same root as 'genius', and the word 'engine' in English comes from the same root. In Latin languages what is called in English 'engine' is 'motor', or 'machine'. So it is quite confusing that in English the basic word 'engineer' means 'someone who operates engines'. The word 'ingineur' in French and similar words in the Latin languages are derived not from 'engine' but---as I indicated---from 'genius'. Today there is still confusion in English, but not in other languages. In France an 'ingineur' does not operate a locomotive. It is unclear to me why in English (in particular in UK English) the word engineer is common to mechanics and highly skilled engineers. But btw, in the times of Da Vinci someone who studied science was an engineer of constructions and machines as well.
the engineeering tools include some tools that use for practical and professional work in engineering, business tool is for the simulation and maybe including how much the cost of the project
Engineering tools could be useful to simulate real systems for example
One of the biggest, and frequently overlooked, problems with the use of business tools in engineering and science is that they lead to letting your software dictate how one analyzes data. Business tools such as MS Word, PPT, etc are fine for writing reports and papers, for presentations,etc., but when it comes to analyzing data, tools such as Excel are not a good choice. These business tools offer very little in terms of the data manipulation, analysis and interpretation that is necessary in science and engineering. I see it all the time in engineering students and junior engineers. They limit their analysis to what they can do in Excel, even if a more sophisticated approach is needed. Being limited by your software choice is not good engineering.
In my opinion; An Engineering Tool may differ in two aspects
1) Reliability and Security Metrics of the tool
2) Redundancy and options for Alternatives!
I think, perhaps naively, the difference between a professional engineering and a professional business environment is that the standards of traceability are quite different. In business if I can track the dollars, that is enough, maybe not best practice, but if an auditor can "see" the cash flow honestly they expect to do some detective work to divine methods and intent. For engineering there is a higher standard of traceability for:
- collaboration and development history, showing rigor of development
- documents supporting publication and precedence (rigorously dated)
- protection of IP and proprietary development
- for agency and standards requirements
Some of the issues are clearly shared in the business world; I would want:
- equal reliability in tools
- mobility and control over my data in both domains
- robust collaboration (tracking and merging support)
My point is that when I list the dozen characteristics that occur to me designing these tools, I weight the features differently in the scientific and business domains. I don't see in the business or scientific communities any such nuanced analysis when tools are actually chosen. Perhaps a more pertinent question is: Why aren't individuals listing what they need a tool to *do* before they choose a tool?
MSOffice is inappropriate for engineering as your coworkers may be forced to use different versions on different platform and cannot exchange the text without hidden reformatting and the software is not open source (which means that any outcome can not be traced back to code behaviour by the end-user.
As an engineer/scientist you want to do things reproducibly and do not like surprises.
Use a free open source tool that is accessible to all and available on many platforms. This is the reason why TeX and OO are more suitable for engineering; if you have a problem with those programs the open source community may solve your problem for the sake of all users.
In the long run you will be able to separate software flaws from engineering flaws better by using open source code.
I really think an engineering tool is a subset of business/office tools. It really depends how you define each. If you consider Visio an office tool, I consider Visio a very useful engineering tool. I believe the precise answer is in HOW each are used, not in WHAT is used.
I agree with most of what's been said above.
We engineers care greatly about (seemingly minor) details that non-scientific folks ignore. We've all probably found ourselves bending some common tool to our (probably unusual) needs.
Engineering (and other scientific) pursuits describe a point-of-view, an approach.
Thus, I'm not convinced that "engineering tools" truly exist; perhaps "engineering uses for tools" would be a fruitful approach?
Take Visio as an example. If I have two complicated architecture diagrams, one derived from the other, and want to review the differences so that I may verify the traceablity to my requirements and test documents, is there a programmatic way to do so?
Take Excel as an example. I have a complicated engineering tolerance analysis based on a measured test set, several pivot table worksheets providing statistical limit analysis over repeated measurements on multiple devices, ending in summary graphs ... then I update the data. How do I include and verify the new data? After I finish updating the sizes, references and calculations how do I verify the operations? How do I find *all* of the differences between old and new documents, in both the equations and data? For that matter, how do I verify the the original analysis, on a small, contrived test data set then replace the test data with measured data, without changing the system I just verified.
I know how to do all of these things in a multi-file system, in flat text files with data, program and mark-up available directly. I'm not so certain these are possible in the MS environment. The difficulty I see is in the *discipline* of engineering, in the need to provide for "proof" that requirements have been met and the need to rigorously document the solution.
Engineering tool is meant for fabrication and or simulation where as Business tool is meant for promoting Business and its activities
Engineering tool is intended to do designing or simulation of something novel which is more application- oriented like fabrication of photovoltic cell or antennae to harvest the Sun's energy and in fact this novel tool is being exploited by many business men like Building promoters or massive designing of industries. In fact, business tool exploits engineering tool for revamping or reorienting business promotion. Business tool as a matter of fact depends on engineering tool rather than vice-versa
Roland, in your praxis as an R&D engineer, how do you validate your design equations? By what process do you maintain and document changes in your design? Not to pick on you in particular, but many have the mistaken belief that because you can do something with a particular tool, it is automatically good practice to use it. Knowing a tool does not make the result provable, correct or good engineering practice.
For a different look at this problem, refer to the US Food and Drug website for the problems meeting V&V standards for Excel. Search for "Excel verification fda".
Depending on human actions for repeatability is not good practice, without a high degree of audit-ability, review and testing. None of which is easy in Excel, which is why standards agencies recommend against Excel in critical development applications.
I have found Google Docs to be a superior tool for collaborative engineering work, at least those parts of it that involve documents and spreadsheets. It provides an on-line archive of all changes to documents, with the possibility of reversion as necessary. Recently I wrote three papers with collaborators literally all around the world in this way - you can even have live commenting and change tracking with your co-workers.
In the software world, low-cost or free versioning and change tracking environments for distributed collaborative development are readily available - for example, Eclipse+SourceForge. Are such environments available for mechanical or electrical engineering?
Engineering tools are developed based on scientific principles and proved their utility in fabrication and or development of systems including software packages while business tools are outcome of experience in promoting business and marketing.
You should use, e.g. Excel only if you know why Excel yields 8 when you calculate -2^2+2^2
Excel especially is a widely adopted engineering tool for evaluating calculations. There are many others out there that many be more appropriate for the specific task eg Solid works for drawing and certain simulations and calcs, Matlab for extensive calculations and programming or Python for pretty much the same as Matlab but free.
@Karl: Not to be terribly nit picky but -2^2 + 2^2 = -4 + 4 = 0 :)
@Jesse: But not in Excel! In Excel -2^2 = 4 (that was Karl's point, I believe)
@Mike: I had to start up Excel to check that one. Yup. There's an EXCELlent source of errors if I ever saw one. I actually suspected that to be the case to be honest but couldn't believe they would ignore standard order of operations, I guess not.
It is interesting to note that on Excel, 2+2 is text. Exp(10) is also text. Anyone who would parse a complex mathematical formula, as default text (unless beginning with a + = or -) in a spreadsheet should not be trusted with any mathematics whatsoever.
I begin all equations with a + or - to save the idiotic extra key stroke Excel forces upon us to get an =.
I have frequently had D^2/4 auto corrected by Excel to be D^0.5, which it seems to do utterly randomly.
We used Excel to generate 2nd order regressions for our Forchheimer correlations for porous media. Excel had no problem generating negative terms. These physically meaningless negative terms were eliminated by dividing our measured pressure gradients by velocity to get linear equations following 'ancient' 1950's practice as used by the likes of Ergun. i.e. Excel 2010 is not as accurate as Ergun 1950 version. So much for modern computing power.
Your greatest computer is between your ears. You can not abdicate the responsibility for proper engineering or scientific practice to any program, be it Excel or some fancy FEM program.
No matter what software you use, you are responsible for the quality control.
Speaking about MS Excel ... anybody remembers Quattro? I am not sure how -2^2 was handled there ... but I do STRONGLY remember how terrible back-draw was to go from Quattro to Excel ...
... and again, which public-domain, or moderately (reasonably) costly SW for spreadsheet calculation would be recommended for more scientific applications?
In -2^2 + 2^2 a unary minus is used and in 2^2 - 2^2 a binary minus is used. A unary minus is evaluated before squaring. Hence programmers consider -2^2 + 2^2 = 8 as correct. But I doubt if normal people know this.
About Excel and as I matter of fact any programming environment. You can easily circumvent priorities by using parentheses. For example -(2^2) will yield -4 but -2^2 will yield 4.
Excel is a very good all around tool that can easily produce results and it can be used to produce nice templates for repetitive work. Still when you have to consider large amounts of data or you need to perform some serious statistical analysis I would suggest something more refined or something that is homemade.
Excel evaluates -a^2 as (-a)^2, that´s one of the problems. Other is the lack of precision in operations floating point arithmetic.
I found problems with excel when I tried comparing secondary variables where the primary variables were not synchronized. ie comparing displacements data taken from various different cameras. Common in impact experiments. In MATLAB its easy to change primary variable to a more convenient one, i.e 2 4 6 ms and then you can easily interpolate the experimental displacement data on these points. This allows you to compare what each camera is reading at each specific time point.
Most people are having a very abstract understanding about engineering. Actually Engineering from a historical point of view means "civil engineering". Nothing else. As soon you have this understanding, then you can answer the question very easy.
In my opinion, great problem in Excel is max number of data in graphs , 32.000.
See: "http://office.microsoft.com/en-us/excel-help/excel-specifications-and-limits-HP005199291.aspx" , Charting specifications Maximum limits.
@ Mark Haefner: The term engineering itself has a much more recent etymology, deriving from the word engineer, which itself dates back to 1325, when an engine'er (literally, one who operates an engine) originally referred to "a constructor of military engines. (According to the Oxford English Dictionary)
According Wikipedia: The earliest practice of civil engineering may have commenced between 4000 and 2000 BC in Ancient Egypt and Mesopotamia (Ancient Iraq) when humans started to abandon a nomadic existence, creating a need for the construction of shelter.
I used to build models with Excel and perform intensive calculations using Excel solvers, gool seek, Macros, Iterative solutions for many engineering problems, I did not recognize this problem, what when I just tried to check it myself, I was surprised. However. I discovered another problem with Excel 2003 regarding generation of uniform random numbers between 0 and 1. I used to generate 1000 numbers and came to a problem that I got negative values. When I went to the website on Excel I found that this problem was discovered my someone else and it was corrected in the version of 2007 and I test my model on that and it works perfectly. I just suggest to pose this problem on the Excel website to be avoided in newer versions.
I still do believe that Excel can be very useful for Engineering applications and it is very good tool for students and lectures to make simple and illustrative models in many engineering and scientific fields.
The scientific tools generally are high in libraries, examples and designs aimed at the production of new technologies. Have the tools for office as Office packages are important to organize and publish information obtained in the first.
To Mark Haefner:
Just a note regarding engineering: the word is derived from the same root as 'genius', and the word 'engine' in English comes from the same root. In Latin languages what is called in English 'engine' is 'motor', or 'machine'. So it is quite confusing that in English the basic word 'engineer' means 'someone who operates engines'. The word 'ingineur' in French and similar words in the Latin languages are derived not from 'engine' but---as I indicated---from 'genius'. Today there is still confusion in English, but not in other languages. In France an 'ingineur' does not operate a locomotive. It is unclear to me why in English (in particular in UK English) the word engineer is common to mechanics and highly skilled engineers. But btw, in the times of Da Vinci someone who studied science was an engineer of constructions and machines as well.
Just compare Matlab toolboxes vs Excel, Latex equations vs Word, and laugh! Additionally, there are engineering tools for statistical classification and pattern recognition like Weka, and Optimization tools like Cplex, to name but a few. They are just water for some engineering applications, but there is nothing similar in a MS office tool.
@Rafael Manory interesting etymology.
"engine'er" (c. 1325) originally referred to "a constructor of military engines." (Oxford English Dictionary) In this context, now obsolete, an "engine" referred to a military machine, i.e., a mechanical contraption used in war (for example, a catapult). Now according to Wikipedia, "ingénieur" is from old-french word "engigneor" which means "builder of war machines", i.e. the original derivations would be from approximately the same root. Quick and amateur sleuthing -
James, thanks for your sleuthing, but I think Wiki is wrong in the case of the French word. "Engineering" as an activity or as a field of study is "genie" in French. Simply taken, according to Occam's razor, this sounds as close to 'genius' as one can get. Other words with a similar root are 'genial' or 'ingenious' and so on...Of course the word may have been imported already for builders of war machines, but that's what a genius is for...:-) amateur linguistic sleuthing too...
Hi, I would add that in French "Génie" means an army specialty dealing with mines and trenches, like GENIUS not like ENGINE. Now, engine has clearly the root in genius, not the reverse. So, the etymology given by Dr. Manory is correct.
Now, in order to address somewhat the programming tools as engineering tools, of course, price matters.
It is normal for a commercial package like MATLAB to be friendly and provide nice GUI and plots. However, when looking for the scientific and historical background, the content of the package should be a wrapper (metalanguuage if you want) to the numerical methods in the netlib repository.
Nowadays the FORTRAN sources are usually available, something unconceivable before 1993. So, it is a good idea to have an old solid FORTRAN 77 at hand, get used to it a little bit and eventually use it in case of need.
In data processing there are many small or unusual little problems which make the life miseraable, such as random numbers generators for programming languages not designed for "number crunching". There is an attempt to offer a desktop helper with such modules (including F77), of course not so large as the whole netlib.org or sourceforge.net, at http://www.editassociates.com . The concept should help the scientists without an interest in IT or programming issues.
- Badges
- Science topic