Only one Ada vendor?

Only one Ada vendor?

[Sloan.Kohler]

Its Friday - thought I would try to stir things up...

Several people have stated that "Ada" refers to the latest approved
version (i.e., Ada => "Ada 2005").  Its been well over a year since
AdaCore announced Ada 2005 support started shipping a pretty complete
implementation.  As far as I know, no other vendor has yet announced
any intention to support the current standard.  Maybe they're working
on it in secrecy. Maybe they are waiting on customer demand.  Maybe
they have already lost most of their customers to AdaCore and simply
don't care anymore.

Whatever the case, I'm concerned that the number of viable Ada vendors
seems to be shrinking. The benefits of language standardization are
greatly diminished if only one vendor bothers to support the standard.

I have no relationship to AdaCore other than being a supported
customer who is very happy with their support and their business model
(i.e. they sell support, not license keys).  Their "build it and they
will come" approach to product development and attracting customers
seems to work well.

I'm aware that Aonix and DDC-I have had some recent Ada related
product announcements. That's a good thing but still no public
indication that they're working on Ada 2005 support. At least its
still possible to find Ada products featured on the Aonix, DDC-I, and
Green Hills web pages.

I'm upset/disappointed by what IBM has done to Apex since taking over
Rational several years ago. Apex was (& still is) an excellent, full
featured (though costly) Ada development environment.  Since IBM took
over it seems that Apex product development has completely stagnated
while they focus on collecting license maintenance fees from legacy
customers who are in too deep to switch compilers.  They also allowed
the same thing to happen with ClearCase - ClearCase users are
switching to Subversion for configuration management. Its not nearly
as capable as ClearCase but its good enough and can be deployed
without buying licenses.   If Apex and ClearCase were marketed with an
open-source, support focused business model (instead of a license fee
focused business model) I'm convinced they would have a lot more
success in the market.

So we need to have more than one Ada vendor to keep Ada viable but I
don't have hope that anyone else can effectively compete with AdaCore
unless they adopt a similar business model.

-- Sloan

Re: Only one Ada vendor?

[Randy Brukardt]

<Sloan.Kohler@gmail.com> wrote in message
news:1192806306.892546.73350@q5g2000prf.googlegroups.com...
> Its Friday - thought I would try to stir things up...
>
> Several people have stated that "Ada" refers to the latest approved
> version (i.e., Ada => "Ada 2005").  Its been well over a year since
> AdaCore announced Ada 2005 support started shipping a pretty complete
> implementation.  As far as I know, no other vendor has yet announced
> any intention to support the current standard.  Maybe they're working
> on it in secrecy. Maybe they are waiting on customer demand.  Maybe
> they have already lost most of their customers to AdaCore and simply
> don't care anymore.
>
> Whatever the case, I'm concerned that the number of viable Ada vendors
> seems to be shrinking. The benefits of language standardization are
> greatly diminished if only one vendor bothers to support the standard.

I agree. I do know that there are multiple vendors (at least three) working
on the Amendment features, because they have given me feedback on the ACATS
tests that are in development. All of three of the vendors have pointed out
enough test errors that I think it is unlikely that they did so by hand (my
own experience with creating/editing the tests without a viable compiler was
that finding errors was extremely difficult -- only a few errors were
detected that way).

The ACAA has confidentiality rules, so I can't say more. Personally, I think
what you are seeing is mostly an unwillingness to promise something whose
scope is not yet clear -- and may not have a strong demand from their
customers. That's certainly the case with Janus/Ada; we've been doing some
work on the new stuff, but it is a middle priority task.

                          Randy Brukardt, man of many hats. :-)

Re: Only one Ada vendor?

[Nasser Abbasi]

Hello;

<Sloan.Kohler@gmail.com> wrote in message 
news:1192806306.892546.73350@q5g2000prf.googlegroups.com...
> Its Friday - thought I would try to stir things up...
>
....
>Since IBM took over it seems that Apex product development has completely 
>stagnated while they focus on collecting license maintenance >fees from 
>legacy customers who are in too deep to switch compilers.

But one of the whole marks of Ada is that it is so much a standardized 
language that one could compile Ada code with any conforming Ada compiler 
and this should never cause a problem?

If this was say C or C++ code full of #ifdefs and each C compiler having its 
own set of a million different compiler switches I would understand, but Ada 
code?  So I am just curious why would it be hard to "switch" Ada compilers?

Nasser

Re: Only one Ada vendor?

[Simon Wright]

"Nasser Abbasi" <nma@12000.org> writes:

> If this was say C or C++ code full of #ifdefs and each C compiler
> having its own set of a million different compiler switches I would
> understand, but Ada code?  So I am just curious why would it be hard
> to "switch" Ada compilers?

Application code is very likely to depend on compiler-specific
extensions and support packages; eg GNAT's 'Unrestricted_Access and
GNAT.Sockets.

You might say 'so make your own socket binding', but our experience is
that that will support an imperfect subset of the underlying library;
for example, not providing TCP_NODELAY.

Another pressure is the need to avoid re-certification. We have strong
resistance to upgrading the version of GNAT we use in case it affects
application behaviour. And remember many of the projects that might
switch will be in long-term maintenance; if there's a complete set of
unit tests it would be less of a risk, but that can be a big IF.

--S

Re: Only one Ada vendor?

[Martin Krischik]

 Sloan.Kohler@gmail.com wrote:

> Its Friday - thought I would try to stir things up...
> 
> Several people have stated that "Ada" refers to the latest approved
> version (i.e., Ada => "Ada 2005").  Its been well over a year since
> AdaCore announced Ada 2005 support started shipping a pretty complete
> implementation.  As far as I know, no other vendor has yet announced
> any intention to support the current standard.  Maybe they're working
> on it in secrecy. Maybe they are waiting on customer demand.  Maybe
> they have already lost most of their customers to AdaCore and simply
> don't care anymore.

The funny thing is: The is only one C and C++ vendor which *fully* supports
C 99 and C++ 2003. Perhaps full standart compliance is out of fashion :-( .

Martin

-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: Only one Ada vendor?

[Gary Scott]

Martin Krischik wrote:
>  Sloan.Kohler@gmail.com wrote:
> 
> 
>>Its Friday - thought I would try to stir things up...
>>
>>Several people have stated that "Ada" refers to the latest approved
>>version (i.e., Ada => "Ada 2005").  Its been well over a year since
>>AdaCore announced Ada 2005 support started shipping a pretty complete
>>implementation.  As far as I know, no other vendor has yet announced
>>any intention to support the current standard.  Maybe they're working
>>on it in secrecy. Maybe they are waiting on customer demand.  Maybe
>>they have already lost most of their customers to AdaCore and simply
>>don't care anymore.
> 
> 
> The funny thing is: The is only one C and C++ vendor which *fully* supports
> C 99 and C++ 2003. Perhaps full standart compliance is out of fashion :-( .

Definitely not the case for Fortran.  Standard compliance is very high 
on the list of vendors.  It does take a while for the changes to be 
rolled in though, more so when there are fairly massive changes to the 
standard.

> 
> Martin
> 


-- 

Gary Scott
mailto:garylscott@sbcglobal dot net

Fortran Library:  http://www.fortranlib.com

Support the Original G95 Project:  http://www.g95.org
-OR-
Support the GNU GFortran Project:  http://gcc.gnu.org/fortran/index.html

If you want to do the impossible, don't hire an expert because he knows 
it can't be done.

-- Henry Ford

Re: Only one Ada vendor?

[Maciej Sobczak]

On 21 Pa , 19:23, Martin Krischik <krisc...@users.sourceforge.net>
wrote:

> The funny thing is: The is only one C and C++ vendor which *fully* supports
> C 99 and C++ 2003. Perhaps full standart compliance is out of fashion :-( .

It is not the issue of fashion, but rather demand - with no demand
there is obviously no pressure to spend resources to implement it.
Apparently there is no demand for *all* language features, which might
be a proof that some of them were defined over-eagerly. A lesson to
learn for language designers.

--
Maciej Sobczak * www.msobczak.com * www.inspirel.com

Re: Only one Ada vendor?

[Florian Weimer]

* Simon Wright:

> "Nasser Abbasi" <nma@12000.org> writes:
>
>> If this was say C or C++ code full of #ifdefs and each C compiler
>> having its own set of a million different compiler switches I would
>> understand, but Ada code?  So I am just curious why would it be hard
>> to "switch" Ada compilers?
>
> Application code is very likely to depend on compiler-specific
> extensions and support packages; eg GNAT's 'Unrestricted_Access and
> GNAT.Sockets.

And compiler-specific time behavior of the run-time library.  If you
move an application which uses Ada.Strings.Unbounded heavily to a
compiler whose run-time library supports only a single-character Append
operation whose run-time is proportional to the string length,
performance might be completely unacceptable.

Re: Only one Ada vendor?

[Robert A Duff]

Martin Krischik <krischik@users.sourceforge.net> writes:

> The funny thing is: The is only one C and C++ vendor which *fully* supports
> C 99 and C++ 2003.

Which vendor is that?

Which vendors "mostly" support C 99?

Just curious...

To be on-topic for comp.lang.ada, I'll say: the latest GNAT fully
supports Ada 2005 (modulo some bugs).  Other Ada vendors are adding Ada
2005 features incrementally, based on customer demand, which seems
pretty reasonable.

- Bob

Re: Only one Ada vendor?

[Georg Bauhaus]

On Sun, 2007-10-21 at 19:57 -0400, Robert A Duff wrote:
> Martin Krischik <krischik@users.sourceforge.net> writes:
> 
> > The funny thing is: The is only one C and C++ vendor which *fully* supports
> > C 99 and C++ 2003.
> 
> Which vendor is that?

http://www.comeaucomputing.com

The error messages look suspiciously similar to those
emitted by the Intel C++ compiler.

Re: Only one Ada vendor?

[Maciej Sobczak]

On 22 Pa , 14:36, Georg Bauhaus <rm.tsoh+bauh...@maps.futureapps.de>
wrote:

> http://www.comeaucomputing.com
>
> The error messages look suspiciously similar to those
> emitted by the Intel C++ compiler.

Both get their compiler front-ends from EDG (www.edg.com).

--
Maciej Sobczak * www.msobczak.com * www.inspirel.com

Re: Only one Ada vendor?

[llothar]

> without buying licenses.   If Apex and ClearCase were marketed with an
> open-source, support focused business model (instead of a license fee
> focused business model) I'm convinced they would have a lot more
> success in the market.

I don't believe. Way to many people think that all can be payed by
support
but this is simply not the truth. Especially for a development
environment
i can't  really see any successfull business model - and i don't see
to much of it (Elicpse is a total different beast) all others pay for
an IDE/compiler.

Re: Only one Ada vendor?

[Ludovic Brenta]

llothar <llothar@web.de> writes:
>> without buying licenses.   If Apex and ClearCase were marketed with an
>> open-source, support focused business model (instead of a license fee
>> focused business model) I'm convinced they would have a lot more
>> success in the market.
>
> I don't believe. Way to many people think that all can be payed by
> support but this is simply not the truth. Especially for a
> development environment i can't really see any successfull business
> model - and i don't see to much of it (Elicpse is a total different
> beast) all others pay for an IDE/compiler.

The last time I paid for an IDE and compiler was, I think, in 1993.
Since then I've only ever used Free Software.  I pay for it by
donating time to enhance said Free Software.

My current employer purchased licenses for Rational Apex and also paid
for support.  After a few years, they became dissatisfied with the
quality of the support and did not renew the maintenance contract.
Eventually they switched to GNAT Pro and now they only pay for
support.  The main driving factor for this switch were (1) support and
(2) combined cost of licenses and support.

-- 
Ludovic Brenta.

Re: Only one Ada vendor?

[adaworks]

<Sloan.Kohler@gmail.com> wrote in message 
news:1192806306.892546.73350@q5g2000prf.googlegroups.com...
>
> Whatever the case, I'm concerned that the number of viable Ada vendors
> seems to be shrinking. The benefits of language standardization are
> greatly diminished if only one vendor bothers to support the standard.
>
It is certainly not in the best interest of the language for there to be only
one compiler option.  However, the economics of Ada seem to govern
the support it receives from compiler publishers.

With the apparent decline in the number of Ada projects worldwide, and
the atttitude among some of the most influential DoD contractors that Ada
usage is deprecated in favor of C++ and Java, where is the incentive for Ada
compiler vendors and tool builders.

Ada continues to be a superior language design for most kinds of software, but
the lack of interfaces to common environments and operating systems, the lack
of tools, and the absence of a customer (e.g., the DoD) with an interest in the
language contributes to its decline as an option.

My role for the last few years has put me closer to the DoD decision-makers,
and I continue to encounter the widespread opinion that Ada is no longer of
interest.   There are, of course, more intelligent, better informed people in 
the
DoD and they understand the importance of Ada in weapon systems development.
But, increasingly, the language is not even in the repertoire of anyone who
graduated from a computer science program in the last fifteen or so years.

When Lockheed-Martin took the ridiculous decision to abandon Ada in favor
of C++ for some of our key weapon systems, the economic incentive for
building Ada compilers and tools declined substantially.    The people who
made this decision were generally pretty ignorant about Ada.  However,
the decision was made, to a large extent, on the perceived cost of using
the language -- training, compiler licensing, support contracts -- not just
a result of their monumental stupidity about the benefits of the language.

I attended a video-telecast briefing to DoD software professionals a few
years ago that was delivered by a then prominent Navy Admiral.  At one
point during his talk he came to, "And now let me say a few words about
Ada."   He went on to lament the experience of Ada and praise the fact
that, with the abandomment of the mandate, "we can now put that disaster
behind us..." His version of the Ada story included derision of the language
including citing a few "...amusing if they were not so serious ..." issues
with the language.   "We hired the best instructors we could for the
Academy [USNA] and even they couldn't get anyone to understand
the language ..."   My quotes are from memory, but pretty close.

By the time he got through with his assessment of Ada, anyone in the 
teleconference
who knew nothing of Ada would have been completely turned-off by it.

We someone need to get Ada on the scoreboard again.  We need some kind
of education process to correct the misinformation about it that is so
widespread.   There was a time when I thought this was a responsibility of
the Ada compiler publishers and tool developers.   Now, I realize that
those most of those compiler publishers (e.g., IBM-Rational) don't care
whether Ada continues to exist or not.   They will continue to make money
without Ada.

So, where does the educational process originate?   Who has the deep-pockets
necessary to make it happen.   The money wasted on meaningless advertising
after the advent of Ada 95 is a lesson learned.  No one wants to hear how
good it is.   In the minds of most software professionals, Java is just as good,
or good enough.   One major weapon system development is using so-called
"real-time" Java instead of Ada.   My objections are perceived as the ranting
of an Ada bigot.  They humor me, though, instead of deriding me.

Where is the flurry of articles and press releases in the computer press and
the general press about the new ISO standard Ada 2005?   No where I have
looked.   Who has written a good article about Ada 2005 for any DoD
publication?    Not very many.    Where are the books on Ada 2005? Only
one that I know of.   I had hoped to update Ada Distilled for 2005 standard
by now, but Ada is now a very small fraction of my time and I simply don't
write Ada code day-by-day as I did ten years ago.

Even so, I still get email about Ada Distilled.   Very little of it is from U.S.
or European correspondents, but from other places in the world.   I even
received an invitation to teach from it in Tehran (which I declined). 
Apparently,
they are using Ada for something or other in Iran.   I have no idea what
they are doing with Ada in Iran, but it was interesting to learn of their 
interest.

We need to get Ada visible again.   It is not useful that Ariane V was 
programmed
in Ada and that it keeps getting brought up as an example of a weakness in
the language.   We need to get information to the computer professionals
about the successes of the language.  We need to get high-profile projects
made public.  And we need more articles in the press about its successes,
and about Ada 2005.

I had a very good initiative started using JGNAT at NPS, where I am now
teaching.  The enthusiasm ran high, and a success with JGNAT could have
reversed a lot of the attitudes among the faculty.  AdaCore withdrew support
for JGNAT, and it turned out to be inadequate for the purpose we intended.
AdaCore might have been able to fix the problem, but chose, because of
economic considerations, to abandon it entirely.   I can report that the
incentive to use an Ada-based tool vanished very quickly among those
faculty members who were originally enthusiastic about JGNAT.

This is similar, in some respects to the DoD's decision to abandon Ada
right after the advent of Ada 95.   "Abandon?"    No.  That was not
the intent of Secretary Paige's memo.  But that was the interpretation
within the DoD.  At the time, I wrote in my JOOP Column that it was a
lot like "grabbing defeat from the jaws of victory," not an original phrase,
but appropriate, I thought.

I continue to believe that Ada is the correct choice for most of the software
we include in safety-critical and weapon systems design.   However, I am
a person of little influence.  I no longer have a voice in the press, and I am
constrained by what I can say in the military community.

If no one takes up the challenge to educate the public and the computing
community, and if we continue to simply rely on the more intelligent
customers making the choice of Ada instead of being more proactive, the
language is certainly doomed to extinction.   As of the present, I see no
one making  an effective case for Ada.   A few trade shows will not make
it happen.  We need a more pronounced effort.

Richard Riehle

Re: Only one Ada vendor?

[anon]

Note: The ISO consider Ada 2007 specification (Adopted Jan 2007) to 
      be an "Ada 95 Amendments 1" only. But It is still "Ada 95" as
      the parent specification. 

There are many reason that you can say "One Ada vendor" and that is 
GNAT. Any one can take a few movements come up with a lot of reasons 
and most of those will be true in one respect or another.

But to start with is that GNAT is to most not a true Ada compiler.  It is 
an "Experimental" and "Educational" Ada compiler only.  Adacore has not 
tried to make it a standard compiler. The reason people say this is that 
before the 2005 specification were adopted in Jan. 2007, Adacore had 
pre-released a compiler that had most if not all of the 2005 features in 
2004 and though 2006. This alone make the GNAT compiler "Experimental".  
Also, no standard Ada or other language compiler would allow obsolete 
features to exist in the current version. No matter what! They would have 
two compilers one with the old standard or outdate features and the second 
current standard only. An example of this is "pragma No_Run_Time" which 
is not a standard Ada pragma or a current feature listed in the GNAT 
manual. It is listed in the obsolete section of the GNAT manual, but 
it still work.  

Also business like standards and those who abide by those standards. 
Adacore with pre-releases of Ada specification does not meet the business 
needs for those standards. Then with Adacore also saying they are the 
leader in Ada, that scare off outside business from using Ada. The 
price tag of $14_000 per year for GNAT PRO hurts as well. Most Ada 
compilers are less than 10% of that with a 30 to 90 day free support 
included. With extended yearly support optional to buy if and when 
needed.  And for their own reasons businesses prefer to buy software 
instead on downloading for free. 

In both of the previous two reason Adacore has done more to "KILL" the 
Ada language than promote it.

The other companies like Janus or Green-Hills have not adopted the Ada 
2005.  On one of their product web pages it states that they would not 
follow GNAT and create a first Amendments to Ada 95 compiler, they will 
wait until a true and complete specification was adopted. In other 
words a specification that was not controlled by GNAT.

Then there the supply and demand reason.  Since there is little demand 
for Ada the companies will would prefer to use their resources on 
other languages that are in demand.

Then there is the anti-government reason.  To most people Ada is 
still tied to US government (direct link stopped in 1998). So, if 
you hate the US government you do not use Ada. In the world, this 
is a "BIG ONE"!


Some have said that other vendors are starting to make updates. Well 
I will believe it when they have the product available until then 
there is only one Ada vendor for Ada 2007.

As for Ada 95 its has drop a few but they are still around 40+ vendors.

I was not going to reply to this post but with a couple of people cutting 
me down again. I said what the hell!


In <1192806306.892546.73350@q5g2000prf.googlegroups.com>,  Sloan.Kohler@gmail.com writes:
>Its Friday - thought I would try to stir things up...
>
>Several people have stated that "Ada" refers to the latest approved
>version (i.e., Ada => "Ada 2005").  Its been well over a year since
>AdaCore announced Ada 2005 support started shipping a pretty complete
>implementation.  As far as I know, no other vendor has yet announced
>any intention to support the current standard.  Maybe they're working
>on it in secrecy. Maybe they are waiting on customer demand.  Maybe
>they have already lost most of their customers to AdaCore and simply
>don't care anymore.
>
>Whatever the case, I'm concerned that the number of viable Ada vendors
>seems to be shrinking. The benefits of language standardization are
>greatly diminished if only one vendor bothers to support the standard.
>
>I have no relationship to AdaCore other than being a supported
>customer who is very happy with their support and their business model
>(i.e. they sell support, not license keys).  Their "build it and they
>will come" approach to product development and attracting customers
>seems to work well.
>
>I'm aware that Aonix and DDC-I have had some recent Ada related
>product announcements. That's a good thing but still no public
>indication that they're working on Ada 2005 support. At least its
>still possible to find Ada products featured on the Aonix, DDC-I, and
>Green Hills web pages.
>
>I'm upset/disappointed by what IBM has done to Apex since taking over
>Rational several years ago. Apex was (& still is) an excellent, full
>featured (though costly) Ada development environment.  Since IBM took
>over it seems that Apex product development has completely stagnated
>while they focus on collecting license maintenance fees from legacy
>customers who are in too deep to switch compilers.  They also allowed
>the same thing to happen with ClearCase - ClearCase users are
>switching to Subversion for configuration management. Its not nearly
>as capable as ClearCase but its good enough and can be deployed
>without buying licenses.   If Apex and ClearCase were marketed with an
>open-source, support focused business model (instead of a license fee
>focused business model) I'm convinced they would have a lot more
>success in the market.
>
>So we need to have more than one Ada vendor to keep Ada viable but I
>don't have hope that anyone else can effectively compete with AdaCore
>unless they adopt a similar business model.
>
>-- Sloan
>

Re: Only one Ada vendor?

[Pascal Obry]

anon a �crit :
> Note: The ISO consider Ada 2007 specification (Adopted Jan 2007) to 
>       be an "Ada 95 Amendments 1" only. But It is still "Ada 95" as
>       the parent specification. 
> 
> There are many reason that you can say "One Ada vendor" and that is 
> GNAT. Any one can take a few movements come up with a lot of reasons 
> and most of those will be true in one respect or another.
> 
> But to start with is that GNAT is to most not a true Ada compiler.  It is 
> an "Experimental" and "Educational" Ada compiler only.  Adacore has not 
> tried to make it a standard compiler. 

Wrong!... Again :(

Please explain to us why AdaCore has validated an Ada compiler at some
point. This was GNAT Pro 4.x IIRC. GNAT is not an "Experimental" or
Educational compiler only. Have you looked at AdaCore customer list and
the projects using GNAT ????

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|              http://www.obry.net
--| "The best way to travel is by means of imagination"
--|
--| gpg --keyserver wwwkeys.pgp.net --recv-key C1082595

Re: Only one Ada vendor?

[Jerry van Dijk]

Colossus.Pike@worldnet.att.net (anon) writes:

> But to start with is that GNAT is to most not a true Ada compiler.  It is
> an "Experimental" and "Educational" Ada compiler only.  Adacore has not
> tried to make it a standard compiler.

Geez. It is not often that I feel compelled to post here nowadays. But this
sort of 'Character Deflamation' is really uncalled for.

We are lucky, considering todays marketplace, that there are several excellent
Ada Toolchains competing for our business. If and how companies invest in
upgrading their systems to the latest standards it up to them. The market will
decide who is right. But to work that market needs accurate and complete
information, not the FUD spreading above.

gr.
Jerry.

-- 
--  Jerry van Dijk
--  Leiden, Holland
--
--  The early bird may get the worm, but the second mouse gets the cheese!!

Re: Only one Ada vendor?

[Georg Bauhaus]

On Sun, 2007-10-28 at 12:13 +0100, Jerry van Dijk wrote:
> Colossus.Pike@worldnet.att.net (anon) writes:
> 
> > But to start with is that GNAT is to most not a true Ada compiler.  It is
> > an "Experimental" and "Educational" Ada compiler only.  Adacore has not
> > tried to make it a standard compiler.

>  But to work that market needs accurate and complete
> information, not the FUD spreading above.

Maybe there is a point in saying that some of the latest offerings
of AdaCore had included "experimental" support for the *new* features of
the Ada 2005 language, in addition to the fully supported Ada 95--much
like compilers for Eiffel or C++ include limited or changing support
for new features of the latest editions of the respective languages.

If some decision maker, by some necessity, has only limited knowledge of
what it means to have a compiler for some programming language,
it makes a difference to say "We fully support X, period! We
support the new Y as well, to the extent that customers...".
It is important to know the decision criteria, and to provide
suitable and still adequate descriptions to the decision makers.