Re: ACT ANNOUNCES: GNAT 3.11p now available!

["news.oxy.com" <Vladimir_Olensky@oxy.com>]

dewar@gnat.com wrote in message <78m3sj$3a8$1@nnrp1.dejanews.com>...
>In article <78kk1d$a6u$1@remarQ.com>,
>  "news.oxy.com" <Vladimir_Olensky@oxy.com> wrote:
>> I hoped that GLADE would come with GNAT 3.11p as ready
>> and easy to use option of the GNAT 3.11p installation
>> package (at list for Windows NT users) that can be
>> installed as one of the available options during
>> installation. Instead it goes separately and you need to
>> mess up with additional Unix utilities to install it. I
>> think it is not very convenient for Windows NT
>> users and so it does not help to promote Ada.
>
>Well this is not a feature that our customers have required
>so far, and so it is unlikely we will do it in the near
>future. However, Vladimir, how about putting your energy
>into useful work instead of writing complaints. Nothing
>like fixing things yourself. If you can create a nicely
>packaged working binary release of GLADE for NT or other
>targets, we will be glad to put it in the contributors
>section at cs.nyu.edu!
>


First of all I should mention that I am not complaining at all.
There is nothing to complain about.
I am glad to have new release of GNAT that is better then previous.
I am just tying to point out things that to my opinion are not very
convenient to the "end-users", especially in Windows NT world. Such
"end-users" may consider Ada and existing Ada tools as building blocks that
can be used to build systems of different complexity where software
components are just few building blocks among many others.

May be in Unix world it is usual practice to force "end-user" to overcome
some difficulties and barriers to keep them fit but in Windows NT world it
is different. Just think over why Microsoft won the battle in the corporate
PC desktop market (Windows NT and Microsoft Office). They are just oriented
to the "end- user" and he/she feels that this is true. There may be
discussions that MS Windows is not the best operating system at all and that
there are better systems. But Windows NT itself and set of tools for
business applications are oriented to the "end-user". The fact is that the
game on PC desktop market is over. Remember destiny of the OS/2, Word
Perfect, Lotus 123, Quttro-Pro and many others.

Ada will have success in non-military areas in Windows NT world if and only
if it will be "end-user" oriented so that any individual (professional
programmer or occasional programmer) that wants to create something under
Windows NT will have ready to use building blocks. Here I would like to
quote myself from one of my posts: "Imagine that you need to build a house
(application) and not just a shelter but a very good house. You have set of
tools (compiler system) - hammer, knives, shovel, axe, nails, glue e.t.c.
whatever is needed for building house. But if there is no industry that
produces building blocks and materials then each person who wants to build
house have first to manufacture them by himself/herself. I hope that this is
clear to everyone how long will it take to build a house and what it will
look like"
From that point of view (a lot of people have the same opinion) nothing in
Ada world (in Windows NT  area) comes even close to Borland Delphi. I will
just  mention that  recently I explained to my 15 years son how to build
simple database application using Delphi and he did this in 30 minutes. Is
that possible now using Ada? I think the answer is obvious. This is also an
answer why it is difficult to make Ada more popular in non-military area.
If there will be ready and easy to use products then there will be more
success.

As a matter of fact your choice is your choice but do not forget that users
are making their choices also. If they have something that more easy to use
and that have a large number of different building blocks for their needs
they choose that product even if it will be based on more less perfect
programming language (see note about ACE below).

I know at least one Ada company that did the right step in right direction,
namely RR Software with their CLAW product. Of course this is not Delphi but
at least something. And I have already ordered it.

Here you suggest me to put my energy for creating public binary GLADE
packaging for Windows NT. If I would be full time professional programmer
and had enough time I would probably do so but I am belong to the class of
occasional programmers.
This should be done by software company professional programmers and not by
enthusiasts.
Of course I will deploy GLADE on my machine but only because I have been
waiting for it quite a while. But many others just won't do that. They may
think ;" Why  should I overcome any difficulties. They do not want to make
it easy for me so I will use something ready to use." As a matter of fact
Delphi itself has communication components and many others are available
from 3d party companies.

I am telecommunications engineer and have a lot of other complicated things
to do in the my telecommunications area. So I want ready to use tools and
building blocks when I need to do something "soft" for my needs. For many
years before I joined to OXY in 1993 I was involved in the Russian manned
space program dealing with top level complex onboard and ground system
development (telecommunications and control systems design, testing, launch
preparations, flight operations e.t.c.) - analyzing system approaches,
solutions, defining system requirements (to hardware and software
components) , intersystem interfaces, systems integration, supervising
development as a whole and developing some components by myself, working
closely with programmers groups, doing some programming by myself and many
other things. One of the important part of such work was finding and
outlining weak sides of the design and pointing out how to make them strong.

I see how well Ada is suited for complex system development but if there are
no ready and easy to use building blocks. Any system designer may choose
something (based on less perfect language) which has already  whatever
needed (when they are not obliged to use particular programming language).
Why should we have to reinvent the wheel if something is already exists (may
be less perfect but it is already at hand).
When I am talking about something I talking from system designer point of
view. Many people here are pure programmers and as such they have somewhat
more narrow point of view on many things and there is nothing wrong in that.

You mentioned that some of these issues were discussed to death in some
other forums and that there are nothing new in what I am talking about. I
suppose that most of these discussions were between pure programmers and Ada
enthusiasts and not many of the "end-users" and system designers in general
were involved in these discussions. Anyway I will check this.
As for me I would like to have perfect RAD system based on perfect
programming language.
Quite natural desire, is not it?

As for system packaging one of the ways widely used in NT world is
incremental value-added packages to add specific group of features to the
already installed products. Such packages may be aimed to different groups
of users. This decrease size if the initial distribution and gives a lot of
flexibility. You can have several such value-added packages for GNAT.

I have a feeling that at least two years have been lost. When Ada was
sponsored by US DoD many people thought that will continue forever and such
mentality played it's negative role. Now Ada software companies have to seek
for new sponsors outside military. And they may appear if needed "soft"
blocks will be developed. As example of that is Adaptive Communication
Environment (ACE) created by Douglas C. Schmidt which is Associate Professor
and Director of the Center for Distributed Object Computing in the
Department of Computer Science in Washington University in Saint Louis
 http://siesta.cs.wustl.edu/~schmidt/ACE.html ). Extremely good toolkit that
was built using not perfect Ada but C++. This is what was really needed for
industry. Just look at sponsors list and the list of companies that are
using it as well as what have been built upon it (e.g. TAO - Real Time
CORBA ). In addition this is free software. I wish to have it written in Ada
95 and be available as one of application building blocks within Ada RAD
environment. There is a great need for some kind of universal communication
software components.

I have a dream that Ada software companies will be oriented not only to the
already formed Ada user groups but to the potential users in many different
areas outside it's initial domain.
Windows NT world is one of such groups.
Some others are:
IBM AS/400 (to give an alternative not to COBOL but rather to the RPG an
CL) - of course if IBM get interested.
Palm Pilot (Ada has there some potential as GNU is already there and no
Microsoft).

Finally I would like to mention once more that Ada software Companies are
making their choices and potential users/customers are also making theirs.
Very often their choices are not the same.


Regards,

Vladimir Olensky
(vladimir_olensky@yahoo.com)
(Vladimir_Olensky@oxy.com)
Telecommunication specialist,
Occidental C.I.S. Service, Inc. ( www.oxy.com )
Moscow,
Russia.