Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #32704

Re: Software's Weak Link is Code

Show all headers | View raw

On 7/26/2014 5:42 PM, Celeros wrote:
> On 2014-07-25 09:16 PM, rickman wrote:
>> On 7/25/2014 10:22 AM, Roger Levy wrote:
>>
>> Funny, over the years I have come to realize that nearly all of my
>> thought processes are graphical in nature.  I excel at things that can
>> be manipulated as an image in the mind.  I find that I can do lots of
>> math in my mind by forming pictures of it.  To me 7 * 24 is an image of
>> blocks of 25 arranged in two rows of four (100 per row) with one missing
>> and one square missing from each block of 25.  I can "see" the missing
>> block of 25 and the seven missing squares so 7 *  24 = 200 - 25 - 7 =
>> 168.  To me *everything* is visual.
>>
>
> Funny indeed! I can totally identify with that rows of blocks mental
> picture for visualising numbers and operations on them. I'm going to try
> that approach a bit more and see where it leads me....
> For me, even just trying to remember 8 items or so I would place each
> item in it's own box, with two rows of 4 boxes, or for 15 items 3 rows
> of 5 boxes.
>
>> I struggled a bit when I had to start designing FPGAs in HDL rather than
>> schematics.  I found written code to convey little understanding or
>> form.  In fact, the indentation and white space are the most important
>> part of code to me... the "picture" of structure of the code.
>>
>> Yes, diagrams for complex things can be complex, but only if you don't
>> manage them properly.  Just as good code is hierarchical and regular, a
>> diagram of that will be well presented and clear.  I always start with a
>> block diagram even if it is just in my mind.  So continuing with a
>> diagram would be second nature to me.
>>
>> I think perhaps Dirk is right in that it may be time to consider what
>> visual programming can do for us.  It has been used in signal processing
>> for some time now and I believe there is a tool from National
>> Instruments which produces code for data acquisition and control.  These
>> are mainstream products which have been successful for some time now.
>> Yes, it is likely overdue to look for better ways to program.
>>
> I agree that for engineering type, very hardware bound systems, where
> everything (inputs, outputs, transformations) can be described
> completely (and mathematically even), that for those systems it should
> be possible to code by diagramming and evolve the system by alterations
> to the diagrams (but admittantly this is not my field, so I'm probably
> talking nonsens!).
>
> Where I have my doubts is when the system being developed is at a higher
> level, more business system like, where UML and ERDs can describe things
> and perhaps speedup the initial state of the system, but any further
> development, how can that be accomplished by diagramming?
>
> If it is possible to stay completely in the diagramming realm, then yes
> perhaps, but how would things work when at some point the ability of the
> diagram to express a specific meaning breaks down, and the generated
> code needs to be altered - those alterations per definition cannot now
> go back into the diagram.
>
> It almost feels like diagramming can only aid comprehension and speed up
> development, but who knows, maybe that is just a view that stems from
> the fact that diagrams usually appear in specifications, and then is
> translated into code, and then perhaps updated rarely to reflect
> subsequent changes in the code.

Yeah, sometimes even things that you think are inherently graphical, 
like schematics break down.  Board level schematics for modern devices 
with hundreds or even thousands of pins portray nothing of the flow of 
the design.  Usually each device is broken into a set of blocks, each 
block containing an interface or few.  They are placed next to the 
device on the other end of the wire... or sometimes not, the other 
device is on a different page of the schematic and connected by a 
reference to the net name and page number requiring a multi-sheet search 
to find all the components on a net.

This is no different from a net list and often portrays the information 
in a manner more difficult to understand.  The only reason people think 
net lists are difficult to work with directly is because they are 
usually generated by machine without human oriented formatting and 
organization.  If they are written for human consumption they can be 
just as good as a complex schematic if not better since everything on 
one net is listed together.

-- 

Rick

Back to comp.lang.forth | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Software's Weak Link is Code visualforth@rocketmail.com - 2014-07-23 09:59 -0700
  Re: Software's Weak Link is Code Roger Levy <roger.levy@gmail.com> - 2014-07-23 10:32 -0700
    Re: Software's Weak Link is Code JUERGEN <epldfpga@aol.com> - 2014-07-23 11:40 -0700
      Re: Software's Weak Link is Code Celeros <kerneels@gmail.com> - 2014-07-25 16:10 +0200
        Re: Software's Weak Link is Code Roger Levy <roger.levy@gmail.com> - 2014-07-25 07:22 -0700
          Re: Software's Weak Link is Code Roger Levy <roger.levy@gmail.com> - 2014-07-25 07:28 -0700
          Re: Software's Weak Link is Code Paul E Bennett <Paul_E.Bennett@topmail.co.uk> - 2014-07-25 18:06 +0100
          Re: Software's Weak Link is Code rickman <gnuarm@gmail.com> - 2014-07-25 15:16 -0400
            Re: Software's Weak Link is Code Paul E Bennett <Paul_E.Bennett@topmail.co.uk> - 2014-07-25 22:19 +0100
            Re: Software's Weak Link is Code Celeros <kerneels@gmail.com> - 2014-07-26 23:42 +0200
              Re: Software's Weak Link is Code rickman <gnuarm@gmail.com> - 2014-07-26 22:44 -0400
        Re: Software's Weak Link is Code "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-07-25 16:45 -0400
          Re: Software's Weak Link is Code Celeros <kerneels@gmail.com> - 2014-07-26 23:08 +0200
    Re: Software's Weak Link is Code "Alex McDonald" <blog@rivadpm.com> - 2014-07-24 00:04 +0100
      Re: Software's Weak Link is Code Roger Levy <roger.levy@gmail.com> - 2014-07-24 08:25 -0700
  Re: Software's Weak Link is Code "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-07-24 03:26 -0400
    Re: Software's Weak Link is Code Mark Wills <markwills1970@gmail.com> - 2014-07-24 01:32 -0700
    Re: Software's Weak Link is Code Unknown <dog@gmail.com> - 2014-07-26 08:17 +0000
      Re: Software's Weak Link is Code "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-07-26 05:39 -0400
        Re: Software's Weak Link is Code visualforth@rocketmail.com - 2014-07-26 17:37 -0700
  Re: Software's Weak Link is Code anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2014-07-24 13:45 +0000

csiph-web