Could computers be designed better? Is it realistically feasible?

Question

An answer on the question at the following link explains that computers are 'badly designed'

https://answers.yahoo.com/question/index?qid=20150724201206AAdvWIU

That's what prompted my question. Not being a computer programmer or engineer, I come to this section for your opinions.

Answer 1

I read the answer from the other post and totally disagreed with him. His premise was from ignorance and he presented little to back up his claim.



I agree with the above posters. The OP is the other thread was frustrated because he was having trouble teaching his grandmother how to use her computer. Many people aren't computer savvy. Look at this board. They don't take the time to learn. But elderly have other issues. I work with the elderly, people with dementia, Alzheimers, etc, and they have their limits. Patience is most important and not speaking "tech speak".

Answer 2

Computer hardware is pretty good. It's very fast, very reliable and very cheap. Today you can buy a computer for a few hundred dollars that is many many times better than a computer that you'd have paid a million dollars for in the 1970s.



The problem is computer software, particularly the user interface - the bit that interacts with the user.



I think that too often error messages are drafted in terms of what effect the error has on the computer instead of suggesting what mistake the user made and how to correct it.



I also think that many users are unaware that usually there is a specific input trigger that causes the computer to do something. For example, in a command line program, nothing happens until the user presses Enter; in a GUI, nothing happens until you click on something and then release the mouse button (some users never cotton on to the fact that nothing happens until you let go of the button).



However, for an art that is no more than 70 years old, there is a lot of tradition in programming. For example, few people ever cotton on to the fact that the Apple key editing commands are based on the venerable emacs editor. If you're a programmer then it makes a lot of sense to import commands from another program, particular when it's one that you and many of your colleagues are already familiar with. But how to explain that to someone who's never heard of emacs?

Answer 3

Absolutely. Everything can be designed better.



Worth reading:

http://www.joelonsoftware.com/articles/LeakyAbstractions.html



I won't debate that technology is designed poorly.... I can't pass judgement since as a developer and electrician, compared to most people I'm an expert, even if I'm terrible.



Have you ever realized how simple concepts can be combined in patterns that result in complex and useful results? This is most evident in science and mathematics, which basically are devoted to these patterns. Some examples:

Biology studies the cell, the building block of life

Chemistry studies the atom, the building block of matter

Physics studies how objects move according to simple rules --- but result in complex and beautiful patterns

Mathematics studies formal systems, from which simple axioms bring useful(?) results

Computer science studies computation --- sometimes digital computation, in which patterns of connected switches are used to express a solution to _any_ solvable problem



No one claims that the deep results of biology, chemistry, mathematics and physics are trivially simple, but here we are parading computer systems as easy-to-use and understand.



Your computer is essentially billions of switches flicking on and off millions of times a second, special-purpose electronics powered by signals moving at nearly the speed of light --- so how do we claim simplicity?

The same way mathematicians do: we define things in terms of themselves and combine them in clever ways, hide that complexity behind notation and hope that we don't have to remember details about what lies underneath the disguise. Then we do it again, and again. Eventually you end up with a concept such as the "mouse cursor".



Unfortunately, the problem is that

The idea of a "mouse cursor" completely fails to consider the massive amount of information required to represent it --- any failure at all in a conceptual component below will lead to unforeseen behavior: not only that, but

1. Users make mistakes

2. Engineers make mistakes

3. Stuff just breaks, and

4. _Theoretical results never exactly match experimental results_



When a computer doesn't do what I want it to, it's _probably my fault_. Even if not, diagnosis means I might have to know a hell of a lot about what's underneath, since an error describing the mouse cursor doesn't necessarily tell me about what actually went wrong.



We can only simplify complicated things so much. The solution is to compartmentalize and abstract --- which has costs of it's own, in efficiency, complexity, and required knowledge.

Answer 4

I would not say that computers are badly designed. As said above, they are, as such, pretty reliable.

It is the software that runs on it that is a pain: too many "programs" running at the same time and not communcating together properly.

Take the normal PC, running windows:

First, a DOS BIOS program runs, reading a text file (not a program!) to get the "config.sys", for the CPU to know what it is connected to. Then the bios runs the autoexec.com, telling the CPU what program to load and run: it is instructed to run "windows", but the windows program calls numerous programs, to install "services", "drivers", applications, User Interface etc, etc. Then you open a browser...

ALL these programs run "at the same time", passing data from one to another.

If you open the task manager, you will see hundreds of "services" running!... totally useless!

Why would you need to have the bluetooth running if you are NOT talking to a bluetooth? The service runs "in case" there is a bluetooth around, taking some processor time that could be used more efficiently.

Why would you have the driver for the old RS232 port, if nothing is connected?

On top of this, do not forget that each program, service, application has been written by "someone else", and the developer of one application does not have a clue what these OTHER programs do (at least, not in details), hence, it is very easy to "pass the wrong data" which may result in incompatibilities... or crashes.

I would rather have an interface where I have to tell the computer WHAT to run, and run ONLY that. It will probably be 100 times faster. If I want to run a browser, I want only the browser to run and the internet interface, but even browsers are overloaded with other programs, other navbars, other add-ons! I would like the internet interface to be included in the browser, and not a pair of programs. I don t need the internet connected if I want to write a letter using words!... But it IS connected, and plenty of traffic occurs while I am writing my letter, traffic I don t care... and actually I don t want! I don t need my "word processor" to connect to internet to see IF there is a new version of Words. I don t want Words to check the internet to see IF there are "add-ons" I MAY want to install!

That is where "computers" could be better developed.

Answer 5

Of course they could be designed better. Just about anything can be designed better.



But some of the complexity comes from flexibility. When you have a very flexible, general purpose tool, it's going to be harder to use for one specific application than a single-purpose tool.



Example: To turn a half-inch bolt head, it's easier to use a half-inch wrench than an adjustable wrench.



My digital wristwatch is easier to use than my laptop computer. But both display time-of-day. Does that make the wristwatch better?

Answer 6

User-friendly is always a holy-grail thing, people has different mindset, a combination of how they absorb the information and limitations and how they relate tasks into understandable instructions. I might as well call that mixing of mindset the "schema" of a person. Something fitting and sensible to a schema may not be so to another.