GNAT Support Costs 
Author Message
 GNAT Support Costs

Quote:


> > No language ever dies. There are even people today who use SNOBOL.

> Yes, old systems still need maintanence and Ada-Systems will need that
> quite long for sure. But all the young college students will jump on
> Java, C++ or whatsoever because this is where the new challenges are.

I think people use these older languages to develop new products too.

Quote:
> > What is missing for small projects that you can't get from another
> > company or the open source world?

> Is GNAT without support a alternative for small (commercial!) projects?

That depends on the project of course. I use unsupported software for
internal projects all the time (Perl, Tcl, gcc, OCaml, ...).

If I were working on a project which used GNAT and the cost of failure was
high I'd probably buy support from ACT. What kind of cost are we talking
about here? Tens of thousands of dollars per year? That's peanuts IMO.

Quote:
> The number of companies who offer low cost Ada compilers (validated and
> supported!) for the Unix world is quite limited, actually there aren't
> any. The cheapest Ada compiler for Unix that I know is still at least 10
> times more expensive than a good C/C++ compiler. Why?

I've found ACT support very good even for the public version of GNAT.

What would help Ada a lot more IMO is a hoard of hackers pounding out
useful Ada code.

-- Brian



Mon, 29 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> Only innumerous small teams and projects create a mass market which is
> what lacks Ada. Not concentrating on the market for small projects and
> companies ist the right way to kill Ada.

Just exactly what level of support would you expect from a company
concentrating
on the market for small projects?

Distribution?

Printed manuals?

Some of these services could be made available for a relatively low fee
outside of ACT,
if there is interest.  The problem comes when expectations are higher than
the services
offered based on a low fee.

SteveD



Mon, 29 Jul 2002 03:00:00 GMT  
 GNAT Support Costs


Quote:


> > No language ever dies. There are even people today who use SNOBOL.

> Yes, old systems still need maintanence and Ada-Systems will need that
> quite long for sure. But all the young college students will jump on
> Java, C++ or whatsoever because this is where the new challenges are.

> > GNAT provides a zero cost hobbyist solution.

> Which is of course a great thing!

> > What is missing for small projects that you can't get from another
> > company or the open source world?

> Is GNAT without support a alternative for small (commercial!)
> projects?

I know of at least one million lines applications (I do not know if
you call this small or big) developped with gcc without support
(as a matter of fact, they kept gcc 2.7.x for a long time because of
the upgrading cost; by the way upgrading a compiler is not something
done lightly in all compagnies I've worked for). Why would it not be
possible for something written in Ada instead of a mix of C and C++?

Quote:
> The number of companies who offer low cost Ada compilers (validated
> and
> supported!) for the Unix world is quite limited, actually there aren't
> any. The cheapest Ada compiler for Unix that I know is still at least
> 10
> times more expensive than a good C/C++ compiler. Why?

Market size can be a factor (development cost for a compiler is nearly
independant of the number of users); the computer vendor have an
interest to have a cheap C and C++ compilers available (and most
computer vendor are also compiler vendor so this increase pressure to
other vendor), validation also has a cost.

This is without speaking about the quality of support (from my
experience, I've better support from ACT as an unsupported user than
with some compiler vendors from which we have a support contract and not
the cheapest one...).

-- Jean-Marc

Sent via Deja.com http://www.deja.com/
Before you buy.



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> Well of course several companies in the past and in the
> present try to meet this need -- in particular Aonix has
> an inexpensive version of Ada 95 available for NT that
> will meet some of these needs, and indeed the public
> version of GNAT meets some of these needs at even lower
> cost :-)

Where is the low-cost compiler for Unix?

Quote:
> Note that Borland lost money trying to pursue this
> market for C -- they could not duplicate their previous
> success with Pascal -- and that is despite selling
> hundreds of thousands of copies of their C compiler.

Well, if I remember right Borland came into trouble because
Ashton-Tate was to big of a hunk to swallow. As you stated right Turbo
Pascal was a great sucess (I loved it!) and this proofes that it can
be done.

Quote:
> 2. We prefer to concentrate on what we are good at, which
> is providing high level support. There are many real
> success stories for Ada, and many of them tend to be
> large scale projects for which our style of high level
> support is ideal.

I think I quite understand your business model, it's fine - for you!
But I still want to do a small, inexpensive project in a small company
and I want to do it with Ada. I'm just sick of arguing against people
who claim that C++ or Java is cheaper.

AW
--
----------------------------------------------------------
Andreas Winckler
http://www.talknet.de/~andreas.winckler/
ICQ# : 28867366

"A good plan today is better than a perfect plan tomorrow"



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> What kind of cost are we talking about here? Tens of
> thousands of dollars per year? That's peanuts IMO.

Mass markets are always made from lots of "peanuts".

AW
--
----------------------------------------------------------
Andreas Winckler
http://www.talknet.de/~andreas.winckler/
ICQ# : 28867366

"A good plan today is better than a perfect plan tomorrow"



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Andreas Winckler

Quote:
>Only innumerous small teams and projects create a mass market which is
>what lacks Ada. Not concentrating on the market for small projects and
>companies ist the right way to kill Ada.

Not really. The only thing that will save ada is us creating useful ada code.

With ACT we have the best of both world. We get the compiler for free. We get
reasonable support (actually I think it is very good) through comp.lang.ada,

We can report errors to act. We may not get star treatment there, but I'm not
sure any other compiler vendor gives better support to their paying customers.

Greetings,



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Andreas Winckler wrote

Quote:
>I think I quite understand your business model, it's fine - for you!
>But I still want to do a small, inexpensive project in a small company
>and I want to do it with Ada. I'm just sick of arguing against people
>who claim that C++ or Java is cheaper.

Perhaps you should challenge their understanding of economics. It is people who
cost money.

You can divide ETS (Effort To Solution) into two critical components; money and
time (although everything boils down to money ultimately). Sometimes you can
spend money to lower the time component. If one chooses Ada over C one can
lower the time component. That lowers the total cost.

Perhaps you should try the following on them:
If I get a C compiler for $100 dollars per seat and an Ada compiler for
$10000/year a seat. How much would the C vendor have to pay me so that his
compiler would become cost effective?

Point is that Ada may reduce mthe ETS down to around 30 % of the C ETS. The
cost of a programmers dwarfs the cost of the compiler if he is as productive as
he can with Ada.

For C (probably as C++ and Java) it is customary to forget the phase which
comes just after you have typed in your program. It is the debugging phase. At
least for me, this tends to be noticable even for small programs. It is amazing
how many mistakes I can make in a few lines. When the program works, we
remember the time it took to type it in, but forget the time spent making it
work.

Greetings,



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> Perhaps you should challenge their understanding of
> economics. It is people who cost money.

I maintain systems with several 100 kLOC written in Ada - I know all
the positive arguments for Ada.

But here are the arguments I'm facing daily which all your arguments
do not address:

- I don't know Ada an I don't want to learn it
- Ada is a dieing language
- too little chances to make career with Ada
- Ada people are hard to find and expensive
- maintenance is paid by the customer
- Ada is too big and slow

By the way, right now I'm dealing with a core dump in an Ada process.
Quite bad, that I was always arguing that this does not happen in Ada.

What I really need is not ammuniton to argue, I have enough of it by
myself. I'm just sick of arguing alone against lots of C++ and Java
guys in our company. Ada is good but it's just being rolled over by
the reality, at least in our company.

AW
--
----------------------------------------------------------
Andreas Winckler
http://www.talknet.de/~andreas.winckler/
ICQ# : 28867366

"A good plan today is better than a perfect plan tomorrow"



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Andreas Winckler

Quote:
>But here are the arguments I'm facing daily which all your arguments
>do not address:

>- I don't know Ada an I don't want to learn it

That is a nice one to have on ones resume.

Quote:
>- Ada is a dieing language

As examplified by the increase in postings on comp.lang.ada

Quote:
>- too little chances to make career with Ada

The chances are steadily improving AND everything learnt can be re-used with
<your language here>.

Quote:
>- Ada people are hard to find and expensive

So are _good_ C++/Java programmers.

Quote:
>- maintenance is paid by the customer

???????????

Quote:
>- Ada is too big and slow

And C++ is not?? If your C/C++ code checked as much as the code generated by an
Ada compiler your programs would be glacially slow.

Quote:
>By the way, right now I'm dealing with a core dump in an Ada process.
>Quite bad, that I was always arguing that this does not happen in Ada.

Core dumps will always happen regardless of the language. Any Unix language can
avoid core dumps by providing enough signal handlers. Unhandled signals will
most likely cause a core dump. It is possible to get a core dump even if a
signal is handled, provided that the new signal arrives before the signal
handler is re-installed.

Quote:
>What I really need is not ammuniton to argue, I have enough of it by
>myself. I'm just sick of arguing alone against lots of C++ and Java
>guys in our company. Ada is good but it's just being rolled over by
>the reality, at least in our company.

The use of Ada seems to be increasing. Notice the steady rise in the number of
messages on comp.lang.ada. The reality is that Ada keeps attracting new people.

One reality is that the use of python is rising (Linux Journal, december 99).
Most people believe that perl is the big star in scripting. That may change
now. Many seem to find that Python is less work than Perl. How long will it
take before companies now looking for Perl programmers will notice this trend
and start asking for Python programmers? Once Python is on board then Ada
starts to look better.

As more people discovers that C/C++/Java is expensive and take a lot of time to
use, the use of languages like Ada/Delphi/Eiffel will increase. It is probably
just a matter of time before Ada people start to influence the Linux community.
When they do others will be attracted to the language.

How many do we need to get critical mass? Don't know. But the more that is
done, the easier it is for others to re-use.

Greetings,



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:
>... It is possible to get a core dump even if a
> signal is handled, provided that the new signal arrives before the signal
> handler is re-installed.

I believe that Unix bug no longer exists in most modern versions of Unix.

- Bob



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs


Quote:

> Is GNAT without support a alternative for small (commercial!)
> projects?

The only support option for MS Visual C++ is $150 per incident (and
notoriously bad). That seems to be enough for hordes of small commercial
projects.

--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html

Sent via Deja.com http://www.deja.com/
Before you buy.



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:
> - Ada is too big and slow

This is a referenced <<Ada gossip>>!
Speed and size depend on your compiler, the optimisations it provides
and the optimisations you are actually using, and how programs are
written...

--
Gautier

_____\\________________\_______\_________
http://members.xoom.com/gdemont/gsoft.htm



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> How many do we need to get critical mass? Don't know. But the more
> that is done, the easier it is for others to re-use.

Well, I try to do my very best to achieve this critical mass. It's just
anything but easy of a job with just 4 Ada developers in a company of
100 to keep Ada alive!

AW
--
---------------------------------------------------------
Andreas Winckler
http://www.talknet.de/~andreas.winckler/
ICQ# : 28867366

"A good plan today is better than a perfect plan tomorrow"



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

>Just exactly what level of support would you expect from a company
>concentrating
>on the market for small projects?

>Distribution?

>Printed manuals?

Amazingly enough, the main problem I've had with free software in
industry is the (lack of) up-front cost.

Example-
  A development group I worked with wanted to run some sort of Unix
  on their PCs.  The person assigned to look into it recommended
  FreeBSD.  The idea was shot down by management because you couldn't
  pay money to get FreeBSD.  The group ended up going with Linux
  because you actually could give somebody money to get a copy of
  Linux.  (Please no flames about Linux being better/worse than FreeBSD.)

--
=======================================================================
 Life is short.                  | Craig Spannring

 --------------------------------+------------------------------------



Tue, 30 Jul 2002 03:00:00 GMT  
 GNAT Support Costs

Quote:

> Example-
>   A development group I worked with wanted to run some sort of Unix
>   on their PCs.  The person assigned to look into it recommended
>   FreeBSD.  The idea was shot down by management because you couldn't
>   pay money to get FreeBSD.  The group ended up going with Linux
>   because you actually could give somebody money to get a copy of
>   Linux.

To refute this, I just pulled up http://www.cdrom.com/ , and in the
ephemeral World Wide Web today, from my machine, FreeBSD is shown
at the top of their "best sellers" box (followed by two Linux variants).

I have no way to verify that it is really a best seller, but it certainly
is offered for sale on CDrom.



Tue, 30 Jul 2002 03:00:00 GMT  
 
 [ 46 post ]  Go to page: [1] [2] [3] [4]

 Relevant Pages 
 

 
Powered by phpBB® Forum Software