Board index » delphi » Re: Borland Reports Net Loss of $5.7M

Re: Borland Reports Net Loss of $5.7M


2004-02-03 06:22:17 AM
delphi190
JQP writes:
Quote
Well, if open source really gives you a cost advantage, why are you
here? Shouldn't you be in a Free Pascal forum?
I use Open Source and I produce Commercial. I get a benefit from Open
Source so, in answer to your question, no :)
--
Derek Davidson
www.enterpriseblue.com
 
 

Re: Borland Reports Net Loss of $5.7M

Quote
Well, if open source really gives you a cost advantage, why are you here?
Shouldn't you be in a Free Pascal forum?


I don't think there's an OS development tool like Delphi (not even close),
or, a database like Oracle. But I also use PHP and Firebird, when they fit.
When it comes to webservers, Apache, OS, is a clear winner against IIS,
closed-source. Point is, there's place for both.
Besides, having a lot of OS components for Delphi in the NET (like those of
Jedi) can only make the case for buyng Delphi stronger.
Regards,
MARCO
 

Re: Borland Reports Net Loss of $5.7M

Cruxy writes:
Quote
But there's a difference... giving a few EUR for software is no
problem. But I won't give ~ EUR 1.700.- for a package from Adobe (I
could not make profit with it.). So I personally use Linux with
Gimp...
I'd rather pay the 1700 than have to put up with using GIMP.
When I tried it a year ago the interface left a lot to be desired. ie.
{*word*76}y awful
 

Re: Borland Reports Net Loss of $5.7M

John Kaster (Borland) writes:
Quote
So, where do you put the requirements knowledge base? How do you verify
the requirements w/ the consumer?
Quoting my post to which you replied:
"You have to be able to figure out where the real payoffs are, not the
bandaids the users always request."
--->"Then a requirements tool is a good thing for productivity, to
assemble and communicate information."
Quote
And have you ever looked at what CaliberRM does?
I never even heard of it til this thread...
--
"The system is less energetic when domains of opposite direction
alternate." - Dr. Memory
 

Re: Borland Reports Net Loss of $5.7M

Harri Vartiainen writes:
Quote
Anders Ohlsson (Borland) writes:

>54% of all errors are due to bad (*) requirements.
>82% of the cost of fixing all errors are due to bad (*) requirements.


What is the source of these numbers?
Harri, haven't you ever heard the observation:
A) The first 80% of the requirements require 80% of the allocated time.
B) And the last 20% of the requirements also require 80% of the
allocated time. <g>
--
"The system is less energetic when domains of opposite direction
alternate." - Dr. Memory
 

Re: Borland Reports Net Loss of $5.7M

Craig Stuntz [TeamB] writes:
Quote
One of the ways that CaliberRM helps with this is
by keeping track of the very complex links between
requirements.
Absolutely a problem. that is where tracking requirements in Excel and
Word breakdown.
Sounds interesting.
Even small projects have complex links between requirements.
--
"The system is less energetic when domains of opposite direction
alternate." - Dr. Memory
 

Re: Borland Reports Net Loss of $5.7M

Quote
>54% of all errors are due to bad (*) requirements.
>82% of the cost of fixing all errors are due to bad (*)
>requirements.

This is in line with my experience.
Doesn't anyone else find these numbers absurd? Not that they're not in
the right ballpark, but is this something that can really be expressed
to two significant digits? Are we sure it isn't 53% or 55%? The range
of complexity of software projects varies so enormously, the talent and
size of the project team working on them varies enormously, the kinds
of requirement methodologies vary enormously. Saying anything more
precise than "it is very common that over half of the errors found in
a large software development project are due to bad requirements" is
just silly, IMO.
-Jim
 

Re: Borland Reports Net Loss of $5.7M

Richard Grossman writes:
Quote
I never even heard of it til this thread...
Ah, the mighty Borland Marketing machine squeaks ever onward ... <g>
--
Derek Davidson
www.enterpriseblue.com
 

Re: Borland Reports Net Loss of $5.7M

Jim Gallagher writes:
Quote
Doesn't anyone else find these numbers absurd?
Of course. When I see numbers like these I am always reminded of the
saying "lies, damned lies and statistics".
You can invariably bend figures to say what you want. Healthy doses of
skepicism are the best way to approach them IMO.
--
Derek Davidson
www.enterpriseblue.com
 

Re: Borland Reports Net Loss of $5.7M

Jim Gallagher writes:
Quote
Saying anything more
precise than "it is very common that over half of the errors found in
a large software development project are due to bad requirements" is
just silly, IMO.
I think that is essentially what the figures are saying. Of course
there is a margin of error, and of course it varies with different
projects. I don't know why anyone would presume differently. You can
either accept the figures as a rough average, or you can demand
standard deviations, methodologies, etc., but either way the point is
the same: Incomplete requirements cost money, and this area is often
where a development team can make the biggest improvement in their
process.
-Craig
--
Craig Stuntz [TeamB] . Vertex Systems Corp. . Columbus, OH
Delphi/InterBase Weblog : delphi.weblogs.com
Want to help make Delphi and InterBase better? Use QC!
qc.borland.com -- Vote for important issues
 

Re: Borland Reports Net Loss of $5.7M

"Derek Davidson" <XXXX@XXXXX.COM>wrote in
message news:401f5c37$XXXX@XXXXX.COM...
Quote
Richard Grossman writes:

>I never even heard of it til this thread...

Ah, the mighty Borland Marketing machine squeaks ever onward ... <g>
See the latest issue of Software Development. There's anad for CalibreRM in
there.
 

Re: Borland Reports Net Loss of $5.7M

Jim Gallagher writes:
Quote
Doesn't anyone else find these numbers absurd?
You are welcome to do your own thorough study to refute the numbers.
I'll stick with the research that has been proven out over the last 30
years until you can show me better research that contradicts it.
Quote
enormously. Saying anything more precise than "it is very common
that over half of the errors found in a large software development
project are due to bad requirements" is just silly, IMO.
I am very interested in formal studies or research that contradicts the
studies we have found. I list some of them here for your convenience.
From James Martin (1979, but other studies have shown only minor
variance on these points), "An Information Systems Manifesto" on the
source of software defects:
56% Requirements errors
27% design errors
7% coding errors
10% other errors
From a U.S. Air Force Project, F. Sheldon, 1992
Reliability Measurement from Theory to Practice
"Note: Most Studies have shown that requirement errors typically
constitute between 40 and 60% of the total number of errors, while
design and code defects combined attribute to only a third of the
errors. that is a pretty significant difference."
From James Martin again on the cost of fixing software:
82% requirements errors
13% design errors
1% coding errors
4% other
And, drilling down into why changes are required for the code from
Easterbrook, et al:
16% Imprecise terminology
3% logical error
30% undocumented assumptions
27% inadequate requirements
24% traceability and consistency
To sum up: if you do not use a strong requirements gathering
methodology and a requirements tracking tool, you are being less than
half as effective as you could be, no matter what kind of developer you
are.
--
John Kaster, Borland Developer Relations, bdn.borland.com
Don't miss the best BorCon ever! info.borland.com/conf2003/
Add a feature/Fix a bug: qc.borland.com
Get source: codecentral.borland.com
 

Re: Borland Reports Net Loss of $5.7M

Quote
You are welcome to do your own thorough study to refute the numbers.
I don't think he's arguing with the general size of the numbers, just
the precision. For instance, your own examples have (slightly) different
numbers :-) No experienced programmer would argue that the general
thrust of the argument is incorrect - certainly it is all stuff I was
taught at uni, and experience has borne it out.
Quote
To sum up: if you do not use a strong requirements gathering
methodology and a requirements tracking tool, you are being less than
half as effective as you could be, no matter what kind of developer you
are.
I think you're over-egging the cake here a bit though. Some projects it
is counter-productive to get too picky about that as they are just too
small (I've just written a Palm program that took a week, all told. No
requirements bugs at all. Some requirements I disagree with, but that's
another matter <g>). But I agree that anything of a substantial size
would benefit from some measure of this. What size that is, is difficult
to judge.
Cheers,
Jim Cooper
_______________________________________________
Jim Cooper XXXX@XXXXX.COM
Falafel Software www.falafelsoft.co.uk
_______________________________________________
 

Re: Borland Reports Net Loss of $5.7M

Admiral Jake writes:
Quote
See the latest issue of Software Development. There's anad for
CalibreRM in there.
Although I totally didn't get it. Why is that person holding a
work of art? Does CalibreRM help you create more artistic,
masterpiece works of art? Does it prevent you from making frilly,
artsy-fartsy apps? What?
Oh - maybe it was supposed to be a painting of the guy, not this
other woman. But what the heck does painting portraits have to do
with managing software projects?
-Brion
 

Re: Borland Reports Net Loss of $5.7M

"Brion L. Webster" <XXXX@XXXXX.COM>writes
Quote
Although I totally didn't get it. Why is that person holding a
work of art? Does CalibreRM help you create more artistic,
masterpiece works of art? Does it prevent you from making frilly,
artsy-fartsy apps? What?
Since the work of art was not well-suited to the man holding it, I think the
idea was that CalibreRM helps keep you from producing beautiful but wrong
apps. It seemed pretty obvious to me.