Re: (not entirely...) OT: OPINION... chicken entrails, runic stones, and crystal balls... WAS CoBOL moved to OO

From: Doug Scott (dwscott_at_ieee.org)
Date: 01/02/04 

Date: Fri, 02 Jan 2004 13:12:03 GMT

Peter,

> 1. The purpose of producing a report is to provide "information".

Yes, and printing it is only one way of presenting the data. Your objection
seems to revolve around printing stuff, which I'll agree is largely
unnecessary. But the numbers still have to be crunched, and it's still more
efficient to crunch monthly figures once and store them for later retrieval
rather than crunch them whenever an individual wants to glance at them.

> 2. There are alternative ways in which information can be presented.
> .... Nowadays, a decently designed transaction system can reproduce a
> report from up to 7 years ago, in seconds, with no requirements for
> microfiche readers, storehouses or reinforced flooring...

True, and has been true for years. Again, printing should be distinguished
from processing. Presentation doesn't require massive database activity - it
requires the data to be there in a form which allows presentation - thus the
Data Warehouse.

> 3. Modern managers have grown up with spreadsheets and databases

Yes, this point is about hard copy print again. Same arguments.

> 4. Government (and similar Legal) requirements for information on duly
> filled out forms are being met by EDI, so there is no need to get a listing
> and transcribe the data to a hard form.

Agreed. But EDI is a /batch/ process.

> CEOs who want to know the top performing
> Branches World wide can get it on their cell phones with "bottom lines"
> updated every few minutes.

As at that instant? For a large company, the cost would be prohibitive for
such a fanciful thought. CEOs should be working on a three-month to ten-year
timeline, not minute by minute. That is the province of the sales department,
so it's much more localised. I know online insurance companies will vary the
rates by time of day depending on sales targets for the day, but to be
honest, people became slightly manic and that generated customer resistance.

Instant ain't good for statistics - you need some form of perspective.

> 6. Much of the current "Reporting Requirements" in Corporations are based
> around Batch Hard Copy

Well, that is certainly a surprise to me. Back in 1975 I justified the cost
of a County Council stock control system purely on the stationery savings. If
the rest of the world hasn't caught up, then I'm faintly surprised, and maybe
that's due to IT rather than the business.

> As these systems are replaced by more Network oriented solutions

Well, you're talking real time versus batch. I thought that argument was won
years ago. Let's focus our discussions:

1. The points you made against batch are really points against hard copy.
That is almost incontestable, I'd have thought, and I wouldn't dream of
arguing otherwise.

2. What you call "the network" isn't particularly new because the points are
based on real time processing, which will improve the timeliness of the data;
online access to the results improves the service to the users and is also
Goodness. But that has nothing to do with Batch, which is merely a method of
crunching the numbers.

Batch processing will survive because it's more efficient to extract periodic
figures once rather than many times. The /presentation/ of these figures is
distinct from extraction, and will change according to the technology
available. 

---
Doug
dwscott@ieee.org