[GAP] Ada 2005 API

Riehle, Richard (CIV) rdriehle at nps.edu
Thu Feb 21 21:53:24 CET 2008


Explaining the ARM was once the province of textbook writers.  The
great works by Norm Cohen, John Barnes, et al served well in that
regard.   They included coded examples that made sense out of the
terse language of the ARM.

Unfortunately, with Ada's significant decline and the widespread 
belief that it is, if not dead, somewhat moribund, book publishers
have minimal interest in producing books in support of Ada.   The
availability of fully-coded examples, and clear, concise explanations
is a sad state. 

I have proposed an Ada 2005 update of my own book, Ada Distilled, to
several publishers only to be met with a total lack of interest. Most
publishers believe there is simply not a market for another Ada book.

For me, Ada has always seemed a good vehicle for teaching a wide range
of computer science and software engineering topics.   Even though the
Navy has abrogated its requirement for us to teach it at NPS, I still 
find ways to include it in some of my classes -- even using for examples
of solutions to computer science problems.    However, I am continuing
to fight an uphill battle at NPS.   The majority of the faculty has
no idea of Ada, and those that do have decided it is no longer relevant.

Even our DoD sponsors, the people sponsoring our research, seem to turn
a deaf-ear to any research proposal that involves Ada.   

The assertion that the ALRM is sufficient simply is not adequate.   We
do
need a coordinated effort to organize the libraries and associated 
documentation so those who have an interest in trying Ada will find the
resources for the language hospitable and easy to acquire.   At present,
unless there is some compelling reason for choosing Ada, the majority
of them find Java or C# to be a little more friendly.   As language 
designs, both of those choices are flawed, but they have great
libraries.
The quality of the libraries and corresponding documentation compensate
for the flaws in the language designs.   We need some similar model for
Ada, something that will be familiar to those accustomed to that kind of
tool and library support.  While some of this exists, its availability
is not obvious to the Ada newcomer.   AdaCore has done exceptional work
in this area, and they are to be congratulated for it.   Somehow it is 
not enough.  We need all the Ada compiler and tool publishers to pitch
in
and make this deficiency go away.

-----Original Message-----
From: gap-bounces at gnat.info [mailto:gap-bounces at gnat.info] On Behalf Of
John McCormick
Sent: Thursday, February 21, 2008 11:32 AM
To: GNAT Academic Program discussion list
Subject: Re: [GAP] Ada 2005 API

My students equate the material in Ada's annexes with APIs.  The ARM 
is poor at describing the operations.   For example, if you look at 
section A.4.5 of the ARM to learn about the operations available for 
unbounded-length strings, you find a package specification with NO 
documentation within it.  (My students point to this lack of 
documentation as a model for not documenting their own package 
specifications.)  Now there is a very brief section after the 
unbounded-length string package specification, but you have to go to 
A.4.3 to find the definitions of equivalent fixed-length string 
operations in order to understand the unbounded-length 
operations.  My students compare this organization to the Java API 
documentation and rightfully claim that it is far easier to figure 
out Java strings than Ada's.  So I would love to see a friendlier 
description of the annexes than that is in the ARM.  One that is 
closer to the online Java API.

John

At 12:36 PM 2/21/2008 -0500, you wrote:

>On Feb 21, 2008, at 12:02 PM, Edward G. Okie wrote:
>
> > Does anyone know of an Ada API specification anywhere on the web?
> >
> > My students who are learning Ada frequently ask me whether there is
an
> > Ada API specification available on the web.  When they learned Java,
> > they typically relied heavily on the online Java API, and as they
are
>
>This may be the problem.  The core of the language is quite
>independent of API's.  The type model, the control structures,
>subprograms and parameter passing,  genericity, concurrency, etc. do
>not need any API's to be understood.  If your students are looking
>for other interfaces when the core of the language provides
>everything they need, they are  looking in the wrong place.  This
>means that the motivation must be other than to write web
>applications, of course.
>
> > learning Ada they are a bit frustrated to not have a similar
> > reference.
> > There is a version for Ada 95 available on the AdaBrowse site, but I
> > don't know of anything similar for Ada 05.  The Ada Reference Manual
> > does provide similar information, but they want something closer to
> > the
> > Java API.
> >
>
>Ed Schonberg
>_______________________________________________
>GAP mailing list
>GAP at gnat.info
>/no-more-mailman.html
>To unsubscribe from this list, please contact the GAP GNAT Tracker 
>administrator
>within your organization.


John W. McCormick                mccormick at cs.uni.edu
Computer Science Department
University of Northern Iowa        voice (319) 273-6056
Cedar Falls, IA 50614-0507       fax (319) 273-7123
http://www.cs.uni.edu/~mccormic/


_______________________________________________
GAP mailing list
GAP at gnat.info
/no-more-mailman.html
To unsubscribe from this list, please contact the GAP GNAT Tracker
administrator
within your organization.


More information about the GAP mailing list