Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: Machine vs. Human [response to Mr. Wright]

[Subject Prev][Subject Next][Thread Prev][Thread Next]

Mr. Wright,

You are understanding exactly what I am looking for.

Why don't you expand your FEA argument into ACI, AISC, or AASHTO?

 

[In reality, many decisions are made by non-technical staff.

For example, when I need to purchase a engineering software in my company,

 I have to make IT engineer and network administrator understand what and why I need it in additon to  boss.

These people do not have any knowledge at all about the structural engineering.]

 

If some upper level people who do not have technical background believe a software can replace drafters

 and engieers in NEAR FUTURE (say, within two or three years) and decide to invest millions of dollars

 into software development project, what can you do as an engineer?

Do you think that their decision makes sense?

 

 

 

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

> I heard some softwares are trying to incorporate drafting, structural=20=

> analysis and=A0miscellaneous=A0functions.

There's such a thing as CAD software that incorporates FEA capability=20

and FEA software that incorporates multi-physics analysis and some=20

crude 'optimization.' Is that what you mean?

If you're talking about software that exercises intelligent judgment,=20

you'll get different answers depending on who you ask. Software vendors=20=

will tell you their packages do design and analysis, but that's an=20

overstatement, usually because they don't understand the scope of the=20

kinds of judgment calls required for engineering design. For example=20

some FEA software may be able to run iteratively and find 'optimum'=20

proportions of a flat plate end closure for a cylindrical pressure=20

vessel, but it won't tell you the reasons why a flat plate closure is=20

such a rotten idea. And it won't give you the least rotten alternatives=20=

based on things like material delivery and shop labor constraints.

In short there's a fair amount of software that will do a lot of the=20

arithmetic thaqta most engineers do and the same package that does _one_=20=

kind of arithmetic can do other types. What software cannot do is apply=20=

the arithmetic intelligently nor can it draw conclusions from ambiguous=20=

results the way the human mind can. Maybe someday when someone figures=20=

out what real intelligence is, then perhaps we'll know something about=20=

artificial intelligence, and then maybe computers can help us with=20

engineering decisions that necessarily need to balance a lot of=20

off-optimum constraints.

Christopher Wright P.E.

 


--

___________________________________________________
Play 100s of games for FREE! http://games.mail.com/