Re: Let's put this to rest

On Thu, 15 Jun 2006 10:19:02 +0100, "S Perryman" <a@xxxxx> wrote:

"AndyW" <foo_@xxxxxxxxxxxxxxxx> wrote in message
news:pt6292pkq8tgd04tajqng44679vlqvrcp9@xxxxxxxxxx

On Thu, 15 Jun 2006 08:53:07 +0100, "S Perryman" <a@xxxxx> wrote:

"AndyW" <foo_@xxxxxxxxxxxxxxxx> wrote in message
news:s39192p6h62sr9vopidhhep9lc3tlos779@xxxxxxxxxx

My question to this. How did you initially know that a stack was
required.

My answer to this : irrelevant to the discussion.

1. The "requirement" is to produce the "OOA" for a stack.
2. Not to argue the toss about where the requirement may have come from.

I would suggest to you that the term analysis generally means to break
something down into its parts. A requirement is simply a
characteristc that must be implemented in the final system for it to
be considered acceptable.

Requirements Analysis is the process of discovering requirements.
Usually by the breaking down of an existing system to determine its
characteristics.

Using that definition - you cannot already have an existing
requirement for a stack - you havint performed the requirements
analysis that determins that is a required characteristic. So it is
not irrelevant to the discussion if you are using the terminology
incorrectly.

Obviously my words about 1 and 2 have been ignored. :-(

The reason I ignored it, is that you had in my mind made an incorrect
assumption and to build a response to an argument based on an
incorrect assumption would have been a bad thing to do. It would have
just re-inforced what I thought was your incorrect view.

If you say that a requirement already exists for a stack, then by
definition there is probably no need to do any formal analysis,
because as I said above, analysis is the breaking down of a whole into
parts [to see how they work] - but you already know how it works,
because you just said there is a stack. Likely now you would just go
and design some solution.

So it was a self defeating argument

However....

Iin the other part of the thread I suggested that requirements
analysis was different from OOA (or normal analysis). In this case
one performs an elicitation of requirements and builds a model of the
problem space (as it exists).

So, one might have a piece of software, the requirement that was
elicited from the customer is that the piece of software needs to be
changed to implement a stack. One does not know 100% how the
software works. So one builds an Analysis model of how the software
works by breaking the original piece of software down into respective
chunks, looking at the code, then building a model. One now has a
requirement + a model of the existing system and can now procede on
working out a desired solution.

In MDA there is the issue of platform independence, so both the
analysis model and the solution model would likely be platform
independant and contain a generic stack. When the solution is created
for the target platform, one maps the generic stack to a specific
stack.

Think of it this way, the original model may be created by a system
architect, the solution may be done by a platform architect.


So

These two entirely different approaches are WHY it is so important to
give 'a toss' of when the requirement came from :)

Now, in both solutions different models of stacks may be used. In MDA
you have both the generic stack (think of a design pattern for a
stack) and the implemenation solution (for which the model maps to).
In the non-MDA approach, you simply have the implementation solution
of the stack.

Thats how I understand things.





.

Relevant Pages

  • Re: Lets put this to rest
    ... Not to argue the toss about where the requirement may have come from. ... characteristc that must be implemented in the final system for it to ... we have the "requirements" for a stack and we have a spec for stack. ...
    (comp.object)
  • Re: make test segfaults with "--enable-shared" on Python 2.3.3
    ... You don't mention what platform you are seeing this on. ... I know that there are platforms where the amount of stack space available ... Modules/_sre.c source file for more info. sre in 2.4 will be ...
    (comp.lang.python)
  • Re: Stacking Washer/Dryer trick?
    ... Are you familiar with the furniture/appliance slides that you can get ... Prolonged weight on the sliders can eventually damage the foam top. ... ground in normal usage should keep the platform from moving. ... Leave enough room behind the stack to tilt the platform backward just ...
    (alt.home.repair)
  • Re: doubt on regarding stacks
    ... >> One filthy trick you could employ (with all warnings set abut ... >> does indeed have a stack and uses it). ... > platform. ... > plz let me know the correct concepts. ...
    (comp.lang.c)
  • Fundamentally impossible to trigger EINTR in multiple threads
    ... I hope that the following statement is incorrect, ... EINTR due receiving an asynchronous signal. ... Other than pthread_kill there is no way to assure that a particular ... Unless a signal is processed on the stack of the thread blocking on ...
    (comp.programming.threads)