Is this really impossible??

[oldschool]

Hi folks,

I am trying to design a new feature for a java-based product and I
have been told by our engineering team that what I want to do is
impossible. Now I'm not new to programming but almost all of it (~20
years worth) is in C, so I know things like this can be done but I'm
not is a position to make a statement w.r.t. java. Can anyone please
help and say at least if there is a solution??

The problem:
I am trying to design an overview window (called a dashboard) which is
basically just a container for simple graphics (line graphs, pies,
etc.)
It has 4 quadrands divided by framesets (moveable), and is initially
empty.
This dashboard is spawned off by a master application. Now what I'd
like to be able to do is simply right-click on a graph displayed in
the master application where I then get the option to send the graph
to one of the quadrants within the dashboard.

In this way, I can keep an overview of the graphs I really need to see
in the dashboard while still being able to leave that graph and
traverse through various menus and other parts of the master
application.

This essentially involves knowing the details of what needs to be
graphed or displayed, but, instead of simply displaying them in the
master window, they should be able to be displayed in a separate child
window (dashboard) and kept running in that window even when the
master process moves on to other things.

I'm assuming that each quadrant within the dashboard will have it's
own thread to manage it's own graph. Or maybe these threads can be
part of the master and only redirect. Either way, surely it's possible
in java for one process to redirect graphics to the window context of
one of it's own child processes?

I hope I explained everything enough to give a general feel for what I
need.
Is this asking for too much technically (i.e. impossible in java) or
am I being fed a load.

Any feedback very much welcome.

 

[shakah]

Sounds like a basic model/view/controller design, and of course it can
be done in Java.

Almost everything is possible in software -- I've found that when a
software person says something can't be done it either means "it's too
hard to do", "I don't want to do it", "I don't know how to do it", or
even worse "I think I know everything, so if I don't know how to do it
it must be impossible" (the last one is frightening to deal with).

 

[John C. Bollinger]

I am trying to design a new feature for a java-based product and I
have been told by our engineering team that what I want to do is
impossible. Now I'm not new to programming but almost all of it (~20
years worth) is in C, so I know things like this can be done but I'm
not is a position to make a statement w.r.t. java. Can anyone please
help and say at least if there is a solution??

At the level of your description (omitted), I don't see any particular
reason to think it wouldn't be possible. Perhaps your engineering
personnel are hung up on some of the details? As a last resort,
anything that can be done in C can be interfaced to Java via JNI, but
nothing you described made me think even that would be necessary.

 

[Malte]

Hi folks,

I am trying to design a new feature for a java-based product and I
have been told by our engineering team that what I want to do is
impossible. Now I'm not new to programming but almost all of it (~20
years worth) is in C, so I know things like this can be done but I'm
not is a position to make a statement w.r.t. java. Can anyone please
help and say at least if there is a solution??

Can't see why this should be impossible. Obviously, one would need to be
very familiar with drawing 2D and have an intimate knowledge of Swing,
but other than than...

 

[oldschool]

Can't see why this should be impossible. Obviously, one would need to be
very familiar with drawing 2D and have an intimate knowledge of Swing,
but other than than...

Thanks very much to everyone to responded to my post. It reinforces my
suspicions all along. I couldn't obviously write in every detail of
the requirement because that would have been an entire document, but
the underlying complexity is no more than as described.

This seems to be a case for 'gentle persuasion' and diplomacy as
opposed to any technical hurdle, which in a lot of cases is nearly
worse. At any rate, it certainly motivates me to learn java. Who says
you can't teach an old dog new tricks..

Thanks again,
Joe

 

[Chris Uppal]

Either way, surely it's possible
in java for one process to redirect graphics to the window context of
one of it's own child processes?

It may be this expression that is confusing the people you are talking to. I
doubt if you really mean that you want to have separate OS processes in this
architecture, but if that /is/ what you really mean then it is indeed
impossible.

(Meaning "as impossible as it ever gets in software" -- more effort than its
likely to be worth.)

-- chris

 

[Paul van Rossem]

oldschool wrote:




It may be this expression that is confusing the people you are talking to. I
doubt if you really mean that you want to have separate OS processes in this
architecture, but if that /is/ what you really mean then it is indeed
impossible.

(Meaning "as impossible as it ever gets in software" -- more effort than its
likely to be worth.)

-- chris

Agreed, I think a problem with these requirements (as stated here) is
that it contains a mix of real user requirements and implementation
details. These implementation choices might not be the best solution to
your problem, and therefore pose undesirable limitations. I can
understand that some engineers might have a problem with that.

But I understand this is only a small excerpt of your real requirements
document (which should only describe the problem from the user domain
and not contain terms like "frameset", "thread", "process")...

Hope this helps,
-- paul