For Programmers: Free Programming Magazines  


Home > Archive > Cobol > March 2005 > Should cobol be behind bars?...









You are viewing an archived Text-only version of the thread. To view this thread in it's original format and/or if you want to reply to this thread please [click here]

 

Author Should cobol be behind bars?...
Kellie Fitton

2005-03-11, 3:55 pm

Hello Everyone,

I am trying to program some business graphics and need your kind
help with some key points. I am plotting a comparison Bar graphs
with six quantitative values to represent on the axis scale, these
variable data that determines the range of values on the scale are
dynamically obtained, therefore, the span of these numbers could be
low, medium or jumbo high. So, I would like to know if there is
some golden rules, when it comes to charting graphics with regards
to the following criteria:

Visual Clarity & Perception:

1). Should the scale have a starting-point of zero?
2). Should the scale have an ending-point higher than the maximum
value represented?
3). Should the scale interval equals the minimum value represented?

thanks a lot for your help. Kellie.

docdwarf@panix.com

2005-03-11, 3:55 pm

In article <1110558946.217050.53610@f14g2000cwb.googlegroups.com>,
Kellie Fitton <KELLIEFITTON@YAHOO.COM> wrote:
>Hello Everyone,
>
>I am trying to program some business graphics and need your kind
>help with some key points.


[snip]

>1). Should the scale have a starting-point of zero?
>2). Should the scale have an ending-point higher than the maximum
> value represented?
>3). Should the scale interval equals the minimum value represented?


Ow... what you're asking here is a matter of data presentation and
representation; people get advanced degrees in stuff like this. The
answer(s) to your questions are simple, though, and can be applied to all
three questions you pose:

It depends on what kind of data you are presenting and the kind of
audience to which you are presenting them.

Take your first question about a starting point of zero. If you are
graphing the distortions found in the roadbed of a large suspension-bridge
under varying weights then a good answer might be 'no'... a large
suspension-bridge's roadbed will show the more-or-less the same distortion
for zero, ten or twenty passenger-cars. Five hundred or eight hundred
cars start to might start to induce stress but under a certain point
charting 'minimal distortion' would be a waste of space.

If you are graphing changes in thread-alignment of a piece of string on
which successively greater weights are hung... then yes, zero is the place
to start; the difference a tenth of a gram makes here is significant.

Likewise, who will be reading the data? A general-practise physician
might be interested in variations across individual patients, differences
between two or three folks might be considered significant. An
epidemiologist might consider any population increment less than five
thousand to be insignificant. One size does not fit all.

http://www.gis-objectland.com/suppo...T_Diagrams.html
seems to have some good examples of different ways to chart data.

DD

James J. Gavan

2005-03-11, 3:55 pm

Kellie Fitton wrote:
> Hello Everyone,
>
> I am trying to program some business graphics and need your kind
> help with some key points. I am plotting a comparison Bar graphs
> with six quantitative values to represent on the axis scale, these
> variable data that determines the range of values on the scale are
> dynamically obtained, therefore, the span of these numbers could be
> low, medium or jumbo high. So, I would like to know if there is
> some golden rules, when it comes to charting graphics with regards
> to the following criteria:
>
> Visual Clarity & Perception:
>
> 1). Should the scale have a starting-point of zero?
> 2). Should the scale have an ending-point higher than the maximum
> value represented?
> 3). Should the scale interval equals the minimum value represented?
>
> thanks a lot for your help. Kellie.
>

You go all directions at once don't you :-)

You wont like it - OO :-)

I'm sure there are specific M/F examples covering this feature, but your
best bet, initially, is to look at a White Paper (on the M/F site)
written by Wayne Rippin. He does a neat job of showing how to link COBOL
to Microsoft Word or Excel.

Using OO COBOL method passing features you can make Excel do exactly
what you want. As a backup feature, from the IDE look at the Type
Library Assistant. You can extract a list of methods both passing and
returning features for a large number of products. If you do an extract
for both Word and Excel you will get two HUGE files/listings - from
memory something like 23,000 lines each, covering all the methods
applicable. I've never checked but it's a fair bet that a large
proportion of the methods required for Word and Excel are identical,
(e.g. Open, Close, Save etc.).

The point is, this is one where you don't re-invent the wheel. OK no
doubt, you can probably do it using your beloved APIs - but I'm thinking
the OO approach will be simpler.

It is not high priority to me - but I need some simple graphs - and
that's the route I propose to take, some time in the future - using
Excel. (Well of course - if you haven't got Excel on your machine
.......? Still that Type Library Assistant droplist *might* contain a
graphing tool that you want or are familiar with).

PS: Looking at your subject title - perhaps you should be :-)

Jimmy
Kellie Fitton

2005-03-11, 3:55 pm

Hi Jimmy,

I am still working and pondering our project
which will be ready very soon. Meanwhile,
my beloved cobol driving me nuts when it
comes to charting graphics --- I am using excel
to chart some graphs and trying to come-up
with a pattern that can help me determine my
graphics creiteria. So far, there is no golden
rule or standard that I can pinpoint from excel.

I donot need to outsource my graphic needs
to excel thru OO or even activeX programming,
using simple win32 API's will do the trick. ;--))

Kellie.

James J. Gavan

2005-03-11, 3:55 pm

Kellie Fitton wrote:
> Hi Jimmy,
>
> I am still working and pondering our project
> which will be ready very soon. Meanwhile,
> my beloved cobol driving me nuts when it
> comes to charting graphics --- I am using excel
> to chart some graphs and trying to come-up
> with a pattern that can help me determine my
> graphics creiteria. So far, there is no golden
> rule or standard that I can pinpoint from excel.
>
> I donot need to outsource my graphic needs
> to excel thru OO or even activeX programming,
> using simple win32 API's will do the trick. ;--))
>
> Kellie.
>

Well Doc has already made some suggestions. If you go into the MicroSOFT
site, using MSDN - you may well get some pointers there on presentation.
The only specific I recollect is where they make recommendations with
Dialogs and not using 'overkill' when introducing colour.

On the Answer Exchange - somebody had a query and posted a screenshot of
his Dialog using Dialog System. I was very impressed how elegant it was.
He had used some entry-fields enlarging fonts, and labels for his radio
buttons were discreetly coloured. A very pleasing presentation. And as
our good friend Michael says, (and he hasn't quoted it for some time
now), "It's the artist not the paintbrush".

The MSDN might point you in the right direction on graphing
presentation. Some googling might give you graphing tips as well.

Jimmy
HeyBub

2005-03-11, 8:55 pm

Kellie Fitton wrote:
> Hello Everyone,
>
> I am trying to program some business graphics and need your kind
> help with some key points. I am plotting a comparison Bar graphs
> with six quantitative values to represent on the axis scale, these
> variable data that determines the range of values on the scale are
> dynamically obtained, therefore, the span of these numbers could be
> low, medium or jumbo high. So, I would like to know if there is
> some golden rules, when it comes to charting graphics with regards
> to the following criteria:
>
> Visual Clarity & Perception:
>
> 1). Should the scale have a starting-point of zero?
> 2). Should the scale have an ending-point higher than the maximum
> value represented?
> 3). Should the scale interval equals the minimum value represented?
>
> thanks a lot for your help. Kellie.


Eek! We are not the one's to ask! I'd start elsewhere:

The American Statistical Society (if there is such a thing).
An editor of a scientific or mathematical journal
Various texts on data presentation (if there are such things)
College professors who develop articles for peer-reviewed journals



Joe Zitzelberger

2005-03-12, 3:55 am

In article <1110558946.217050.53610@f14g2000cwb.googlegroups.com>,
"Kellie Fitton" <KELLIEFITTON@YAHOO.COM> wrote:

> Hello Everyone,
>
> I am trying to program some business graphics and need your kind
> help with some key points. I am plotting a comparison Bar graphs
> with six quantitative values to represent on the axis scale, these
> variable data that determines the range of values on the scale are
> dynamically obtained, therefore, the span of these numbers could be
> low, medium or jumbo high. So, I would like to know if there is
> some golden rules, when it comes to charting graphics with regards
> to the following criteria:
>
> Visual Clarity & Perception:
>
> 1). Should the scale have a starting-point of zero?
> 2). Should the scale have an ending-point higher than the maximum
> value represented?
> 3). Should the scale interval equals the minimum value represented?
>
> thanks a lot for your help. Kellie.


I will toss my humble $0.02 in.

Only if it makes visual sense. If you have six bars that are between
700 and 900 then your scale should be something like 650 to 950.
Otherwise it becomes hard to tell the difference between bars.

Picking a scale slightly higher than the highest value and slightly
lower than the lowest value gives maximum difference between the bars,
As making the differences between things visually distinct is the entire
purpose of graphs, this is a good idea.
Kellie Fitton

2005-03-12, 3:55 am

Joe,

This is a good idea. the MS excel program tends to do
the same thing as well with every graph I created.

thanks, Kellie.

HeyBub

2005-03-12, 3:55 pm

Joe Zitzelberger wrote:
> In article <1110558946.217050.53610@f14g2000cwb.googlegroups.com>,
> "Kellie Fitton" <KELLIEFITTON@YAHOO.COM> wrote:
>
>
> I will toss my humble $0.02 in.
>
> Only if it makes visual sense. If you have six bars that are
> between 700 and 900 then your scale should be something like 650 to
> 950. Otherwise it becomes hard to tell the difference between bars.



Consider, too, graphing something like "Space Alien Abductions 1600-Present"

You'd probably ned a Log or Log-Log Y-Axis.


Sponsored Links







Also available: Server administration forum archive | Web Design forum archive | Software forum archive | Hardware reviews archive

Copyright 2008 codecomments.com