What is the definitive language reference for PL/I 
Author Message
 What is the definitive language reference for PL/I

If one were interested in writing a compiler, what document
is specifies the language?  Where can it be obtained.

Thanks,

Joe

--

Joe McGuckin

ViaNet Communications
994 San Antonio Road
Palo Alto, CA  94303

Phone: 650-969-2203
Cell:  650-207-0372
Fax:   650-969-2124



Mon, 04 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I

Quote:

> If one were interested in writing a compiler, what document
> is specifies the language?  Where can it be obtained.

I'm sure people will correct me if I'm wrong, but I don't think there is
*one* definitive document.
The original specification from IBM in the "Vienna definition language"
(like BNF, but more) is not (AFAIK) available.  There are a couple of
ANSI standards, full PL/I and subset G, available in hardcopy for about
$100 or so.  The most widely used versions of PL/I are IBM's, which
contain a lot of extensions.  If you wanted a compatible PL/I, IBM's
_PL/I Lanuage Reference_ manuals for the various versions seem to
contain a complete definition, although oriented to the user and not the
compiler writer.  The new "workstation" compilers contain a ton of
additional language extensions and seem to be slated to become IBM's
standard version of the language on all platforms.


Tue, 05 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I


Quote:

>> If one were interested in writing a compiler, what document
>> is specifies the language?  Where can it be obtained.

>I'm sure people will correct me if I'm wrong, but I don't think there is
>*one* definitive document.
>The original specification from IBM in the "Vienna definition language"
>(like BNF, but more) is not (AFAIK) available.  There are a couple of
>ANSI standards, full PL/I and subset G, available in hardcopy for about
>$100 or so.  The most widely used versions of PL/I are IBM's, which
>contain a lot of extensions.  If you wanted a compatible PL/I, IBM's
>_PL/I Lanuage Reference_ manuals for the various versions seem to
>contain a complete definition, although oriented to the user and not the
>compiler writer.  The new "workstation" compilers contain a ton of
>additional language extensions and seem to be slated to become IBM's
>standard version of the language on all platforms.

With IBM's announced intent to support Linux, is there a reasonable
chance they may make a Linux version of one of the PL/1 compilers at a
reasonable cost?


Tue, 05 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
-----BEGIN PGP SIGNED MESSAGE-----


Quote:

>With IBM's announced intent to support Linux, is there a reasonable
>chance they may make a Linux version of one of the PL/1 compilers at a
>reasonable cost?

I really wouldn't count on it.

Also, what incentive is there for someone to program in
PL/1 if they must pay for the compiler, and *good* compilers
for other languages, such as C, C++, fortran, Ada-95, Pascal,
Objective C, Java, Perl*, SML, etc, are freely available?
If PL/1 is to survive, there *must* be a free compiler
available in order to spark interest.  Open source would
be best, as well.

Someone had mentioned a possible front end to GCC.  Why don't
some of the PL/1 experts here weigh in and lend a hand on that
project?  I believe that it could certainly be worthwhile.

        - Dan C.

(* Yes, there is a perl compiler available.)

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBN8hHKoshWrlwdTE9AQGuuQQAkVL1u/mSMW1tgFZpBo8ZhgOpsMDiD29i
MJFoGoQ/Twj6IdrFUeBVjjNMW1NydMUg7o0zbqTqEJDkxdsNVxhvbbsZpM46aAGn
nqscRwInOusGmTkO2011RUJTs1R9+Gr5PbrXQX8ERyoR/ehG2zoXBE/VWwDkM7GF
voXRyYcX6ZI=
=3dJH
-----END PGP SIGNATURE-----



Wed, 13 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I


Quote:
>-----BEGIN PGP SIGNED MESSAGE-----



>>With IBM's announced intent to support Linux, is there a reasonable
>>chance they may make a Linux version of one of the PL/1 compilers at a
>>reasonable cost?

>I really wouldn't count on it.

>Also, what incentive is there for someone to program in
>PL/1 if they must pay for the compiler, and *good* compilers
>for other languages, such as C, C++, FORTRAN, Ada-95, Pascal,
>Objective C, Java, Perl*, SML, etc, are freely available?
>If PL/1 is to survive, there *must* be a free compiler
>available in order to spark interest.  Open source would
>be best, as well.

>Someone had mentioned a possible front end to GCC.  Why don't
>some of the PL/1 experts here weigh in and lend a hand on that
>project?  I believe that it could certainly be worthwhile.

>    - Dan C.

>(* Yes, there is a perl compiler available.)

The problem with frontending GCC is GCC does not support a rich enough
language to allow a front end to work.  PL/1 has too many data types
that GCC can't support in a reasonable manner.  Decimal is completely
unsupported as are bit strings and PIC fields.  I believe that the GNU
people looked into this in the late early/mid 80's and decided not to
pursue it.

I'm not sure about ada, but C, C++, FORTRAN, Pascal, Ovjective C, Java,
Perl, do not have enough features to do the work that PL/1 can.

Most of the stuff I do takes 2-4 times longer in C/C++ then in PL/1 has
PL/1 supports the data types I need as builtin, in C/C++ I have to jump
through loops at best to simulate them making the program significantly
longer to write and harder to maintain.



Fri, 15 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
-----BEGIN PGP SIGNED MESSAGE-----


Quote:

>The problem with frontending GCC is GCC does not support a rich enough
>language to allow a front end to work.

``It all turns into assembly language, anyway...''

What do you mean, ``doesn't support a rich enough language?''
GCC is a tool for building a compiler; an author provides his
or her own front end to a low-level language which then gets
optimized and presented to a code generator to produce
assembly/machine code.  I'm talking about a real compiler here,
integerated with the GNU Compiler Collection; not a front-end
to C.  (Hint: that low-level language is NOT C).  Why GCC as
a front end?  Perhaps because it provides an interface where
a substantial amount of the work involved in producing a usable
compiler has already been done for you?  (Eg, lots of optimiz-
ation, and code generation, etc).

Quote:
>PL/1 has too many data types
>that GCC can't support in a reasonable manner.  Decimal is completely
>unsupported as are bit strings and PIC fields.  I believe that the GNU
>people looked into this in the late early/mid 80's and decided not to
>pursue it.

I'm not sure we're talking about the same thing; I'm talking
about a real compiler, a front end to the GCC (by the way,
GCC no longer refers to a *C* compiler, but a *collection* of
compilers for different languages, including some of those I
had named previously.  The acronym GCC stands for, ``the GNU
Compiler Collection''); are you talking about a front end to C?

Quote:
>I'm not sure about ada, but C, C++, FORTRAN, Pascal, Ovjective C, Java,
>Perl, do not have enough features to do the work that PL/1 can.

This is irrelevant.  We're talking a PL/1 frontend to the GCC
intermediate representation, optimization, and code generation
system, not the merits of one language versus another.

Quote:
>Most of the stuff I do takes 2-4 times longer in C/C++ then in PL/1 has
>PL/1 supports the data types I need as builtin, in C/C++ I have to jump
>through loops at best to simulate them making the program significantly
>longer to write and harder to maintain.

Again, irrelevant.  If you want to get PL/1 acceptance in
the mainstream world, then pitch in and produce a decent,
free, compiler for it that runs under FreeBSD and Linux.
I suggested a GCC front-end as a possible route to pursue,
and you seem to have taken exception to that.  If you don't
want to do anything but complain about how quirky C/C++
is, then do nothing, save for carrying on with posts like
your last.  But in that case, don't expect PL/1 to gain
lots of new users just because, ``it's better....''
The choice is yours.

        - Dan C.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBN82QOoshWrlwdTE9AQFQfQQAgmVsYikhGI9zpc7Whhs/uFhO64ptKt1/
9fIhnzER+XJa1CXyzs1xhX6bHfWqMetGqjFUh13bLijo9Bn4wgvjqZXE7O28a7Pa
c9fxaQ33EZir1ISQCOZ3kmarqgaC58OnghVIMgpXv1oGWL4zTVC2Y7LcIkZ/FFaT
fXsn3JXw1AA=
=BL1M
-----END PGP SIGNATURE-----



Sun, 17 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I

Quote:



> Also, what incentive is there for someone to program in
> PL/1 if they must pay for the compiler, and *good* compilers
> for other languages, such as C, C++, FORTRAN, Ada-95, Pascal,
> Objective C, Java, Perl*, SML, etc, are freely available?
> If PL/1 is to survive, there *must* be a free compiler
> available in order to spark interest.

There is.  It's Digital research PL/I for DOS.
It's free for non-commercial use.

Location is in the FAQ.

Quote:
>  Open source would
> be best, as well.

> Someone had mentioned a possible front end to GCC.  Why don't
> some of the PL/1 experts here weigh in and lend a hand on that
> project?  I believe that it could certainly be worthwhile.

>    - Dan C.



Mon, 18 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
-----BEGIN PGP SIGNED MESSAGE-----


Quote:

>There is.  It's Digital research PL/I for DOS.
>It's free for non-commercial use.

>Location is in the FAQ.

Errm, I could maybe get the PL/1 compiler for Multics
for free; it doesn't do me much good, though, since I
don't use Multics (at least, not directly).  Similarly,
I don't use DOS.  In other words, compilers for dead
platforms don't count, and to my knowledge, there don't
seem to be any free compilers for available for modern
platforms.

For those literalists who seem to frequent this news-
group, I a) don't really think that I can find a copy
of the Multics compiler, and b) think that even though
a compiler might exist for DOS, it's more or less useless
since few people USE DOS anymore.  (I feel confident in
saying that the DOS user-population is dwindling.)  If
people want mainstream acceptance for PL/1, a *free
compiler* must exist, but it MUST exist for *commonly
used platforms*.

Once again, I put forth a front-end for GCC as being
the logical choice, as not only is much of the work of
building a functional compiler already done for you, but
the backend is portable to quite a few different plat-
forms.  For those who feel that PL/1 supports a set of
types which is too rich for a front end to GCC, I gently
submit that issues such as data type representation at
the machine level are going to have to be dealt with
ANYWAY in the production of a reasonable compiler, and
anyone wanting to spread the use of PL/1 should be
prepared to solve those problems in the construction
of a solid, free compiler.

        - Dan C.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBN89J2oshWrlwdTE9AQHRRwP+J0r/FBobDQqzIecylJz7N3HVB9B+7AKA
mlnHEphQCbtD7hKRJBNcNW+JFjSWeN0ShCktgHQaqdJiKzrtUmFiF0no+gd/QrTV
5XVP89JZ9WH64u1Y/Ykat/CRLBj2C9hL8tMn6aZXrdPkldLolheB6SBIlcwGTPW2
Tsevne3lMv4=
=NTK2
-----END PGP SIGNATURE-----



Tue, 19 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I


Quote:
>-----BEGIN PGP SIGNED MESSAGE-----



>>The problem with frontending GCC is GCC does not support a rich enough
>>language to allow a front end to work.

>``It all turns into assembly language, anyway...''

>What do you mean, ``doesn't support a rich enough language?''
>GCC is a tool for building a compiler; an author provides his
>or her own front end to a low-level language which then gets
>optimized and presented to a code generator to produce
>assembly/machine code.  I'm talking about a real compiler here,
>integerated with the GNU Compiler Collection; not a front-end
>to C.  (Hint: that low-level language is NOT C).  Why GCC as
>a front end?  Perhaps because it provides an interface where
>a substantial amount of the work involved in producing a usable
>compiler has already been done for you?  (Eg, lots of optimiz-
>ation, and code generation, etc).

About 15 years ago the Gnu/FSF announced they were going to do a PL/1
front end, a few months later they announced they weren't.  I don't
remember the details, but my vague memory is they figured they'd have
to rewrite the entire gcc system to handle pl/1.  There was certainly
interest.

Quote:
>>PL/1 has too many data types
>>that GCC can't support in a reasonable manner.  Decimal is completely
>>unsupported as are bit strings and PIC fields.  I believe that the GNU
>>people looked into this in the late early/mid 80's and decided not to
>>pursue it.

>I'm not sure we're talking about the same thing; I'm talking
>about a real compiler, a front end to the GCC (by the way,
>GCC no longer refers to a *C* compiler, but a *collection* of
>compilers for different languages, including some of those I
>had named previously.  The acronym GCC stands for, ``the GNU
>Compiler Collection''); are you talking about a front end to C?

Nope, a true PL/1 compiler.  The GCC package may have evolved enough
since then to make it doable today, I'm not sure, I've never had the
time to really look into GCC.

Quote:
>>I'm not sure about ada, but C, C++, FORTRAN, Pascal, Ovjective C, Java,
>>Perl, do not have enough features to do the work that PL/1 can.

>This is irrelevant.  We're talking a PL/1 frontend to the GCC
>intermediate representation, optimization, and code generation
>system, not the merits of one language versus another.

The intermediate representation has to be able to represent the language.
If there is no intermediate representation for decimal, PL/1 strings
(both bit and character), and other PL/1 data types that have no real
equivalent in other languages, then the intermediat language has to
be rewritten to add those features which may lead to reqritting the
optimizer and code gen to handle opcodes not used by other languages.

Quote:
>>Most of the stuff I do takes 2-4 times longer in C/C++ then in PL/1 has
>>PL/1 supports the data types I need as builtin, in C/C++ I have to jump
>>through loops at best to simulate them making the program significantly
>>longer to write and harder to maintain.

>Again, irrelevant.  If you want to get PL/1 acceptance in
>the mainstream world, then pitch in and produce a decent,
>free, compiler for it that runs under FreeBSD and Linux.
>I suggested a GCC front-end as a possible route to pursue,
>and you seem to have taken exception to that.  If you don't
>want to do anything but complain about how quirky C/C++
>is, then do nothing, save for carrying on with posts like
>your last.  But in that case, don't expect PL/1 to gain
>lots of new users just because, ``it's better....''
>The choice is yours.

If I had the time and resources, I'd be writing a PL/1 compiler right now.
The current biggest non-work project I'm working on is a possible Multics
emulator.  There's been some on/off talk about either an emulator,
recoding, etc in alt.os.multics.


Thu, 21 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
-----BEGIN PGP SIGNED MESSAGE-----


Quote:

>About 15 years ago the Gnu/FSF announced they were going to do a PL/1
>front end, a few months later they announced they weren't.  I don't
>remember the details, but my vague memory is they figured they'd have
>to rewrite the entire gcc system to handle pl/1.  There was certainly
>interest.

A lot has gone on in gcc development in the last 15 years.

Quote:
>Nope, a true PL/1 compiler.  The GCC package may have evolved enough
>since then to make it doable today, I'm not sure, I've never had the
>time to really look into GCC.

The gcc2 rewrite was done to address fundamental problems in gcc1.
GCC can support some pretty wildly varying languages, I think it
could probably deal with PL/1.

Quote:
>>This is irrelevant.  We're talking a PL/1 frontend to the GCC
>>intermediate representation, optimization, and code generation
>>system, not the merits of one language versus another.

>The intermediate representation has to be able to represent the language.

In a very low-level form, yes.  The GCC intermediate language is
basically a step above assembler.

Quote:
>If there is no intermediate representation for decimal, PL/1 strings
>(both bit and character), and other PL/1 data types that have no real
>equivalent in other languages, then the intermediat language has to
>be rewritten to add those features which may lead to reqritting the
>optimizer and code gen to handle opcodes not used by other languages.

I see your point, but the beauty of the intermediate language
is that it allows you to write support for those facilities in
an architecture independent way.  Since it's sort of ``one step
above assembler,'' I think that it can be made to support PL/1's
native data types.  I mean, you're not going to find a SPARC
processor that directly grok's decimal, so if you want to run
PL/1 on a SPARC, you've got to figure out a way to represent
those data types.  That is, a representation for these things
is going to have to be thought through anyway,

Here's another way to look at it:  If you don't target the GCC
intermediate language, what WOULD you target?  Probably the
assembler or object code for a particular architecture, right?
Assuming that most architectures aren't going to directly repre-
sent some of the more esoteric members of the type set of PL/1,
you're going to have to figure out a representation for those
types on that architecture anyway.  When you want to move the
compiler, you're going to have to do the same thing all over
again (I can imagine that the MIPS processor doesn't do decimal,
either).  Why not instead apply that effort towards producing
GCC intermediate code, and then get something like 17 architect-
ures for free...  Besides, if the GCC intermediate representation
had to be modified to support a new language, and it could be
done in such a way so as to not break other languages, I couldn't
imagine that the GCC people wouldn't support that.

Quote:
>If I had the time and resources, I'd be writing a PL/1 compiler right now.
>The current biggest non-work project I'm working on is a possible Multics
>emulator.  There's been some on/off talk about either an emulator,
>recoding, etc in alt.os.multics.

Oh yeah?  Wow, that sounds really cool.  I have to go check
out the Multics newsgroup again...

        - Dan C.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBN9K2S4shWrlwdTE9AQF9XQP+OEvvm1Z9T7RZ1M1vqDjQ3DudTdQ1jadF
GBpZRcX6saUK/hk0yJnYExmk/Yop1E5xoX4Vh+qdISp2Qdd9MQA3W6Tg3hWV1QNb
HEFfkwuLM3eKjYI59rUomacGtkYr8q/pRA6RLVUjf4veXOoGvJAfvLA/Ik9zyVJr
/jk/BkQcZaI=
=YYM4
-----END PGP SIGNATURE-----



Thu, 21 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I


Quote:
>-----BEGIN PGP SIGNED MESSAGE-----



>>About 15 years ago the Gnu/FSF announced they were going to do a PL/1
>>front end, a few months later they announced they weren't.  I don't
>>remember the details, but my vague memory is they figured they'd have
>>to rewrite the entire gcc system to handle pl/1.  There was certainly
>>interest.

>A lot has gone on in gcc development in the last 15 years.

>>Nope, a true PL/1 compiler.  The GCC package may have evolved enough
>>since then to make it doable today, I'm not sure, I've never had the
>>time to really look into GCC.

>The gcc2 rewrite was done to address fundamental problems in gcc1.
>GCC can support some pretty wildly varying languages, I think it
>could probably deal with PL/1.

>>>This is irrelevant.  We're talking a PL/1 frontend to the GCC
>>>intermediate representation, optimization, and code generation
>>>system, not the merits of one language versus another.

>>The intermediate representation has to be able to represent the language.

>In a very low-level form, yes.  The GCC intermediate language is
>basically a step above assembler.

>>If there is no intermediate representation for decimal, PL/1 strings
>>(both bit and character), and other PL/1 data types that have no real
>>equivalent in other languages, then the intermediat language has to
>>be rewritten to add those features which may lead to reqritting the
>>optimizer and code gen to handle opcodes not used by other languages.

>I see your point, but the beauty of the intermediate language
>is that it allows you to write support for those facilities in
>an architecture independent way.  Since it's sort of ``one step
>above assembler,'' I think that it can be made to support PL/1's
>native data types.  I mean, you're not going to find a SPARC
>processor that directly grok's decimal, so if you want to run
>PL/1 on a SPARC, you've got to figure out a way to represent
>those data types.  That is, a representation for these things
>is going to have to be thought through anyway,

If they've made alot of improvements, maybe its worth looking into the GCC
system again.  I already agree with your other arguments about making it
available on many platforms, I'm just concerned that every thing needs
to be changed.

Sense you seem pretty famalier with this or at least alot more famalier
then I am, do you have a suggested starting point?  URL for the GCC
internals, etc?

[snip]

Quote:
>>If I had the time and resources, I'd be writing a PL/1 compiler right now.
>>The current biggest non-work project I'm working on is a possible Multics
>>emulator.  There's been some on/off talk about either an emulator,
>>recoding, etc in alt.os.multics.

>Oh yeah?  Wow, that sounds really cool.  I have to go check
>out the Multics newsgroup again...

Its sortof a low level softsell, each time I suggest it, I get less nay
sayers and more people who think it might be worth doing and/or doable.


Thu, 21 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
As long as the code generator supported emitting some sort of activation record,
you should be able to create a pl/i front end. Note that this is the same
requirement for PASCAL.

Everything else in the language not currently supported by the compiler
(e.g. a non C/C++/Fortran datatype) can initially be implemented via run
time libraries.

--

--

Joe McGuckin

ViaNet Communications
994 San Antonio Road
Palo Alto, CA  94303

Phone: 650-969-2203
Cell:  650-207-0372
Fax:   650-969-2124



Thu, 21 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
-----BEGIN PGP SIGNED MESSAGE-----


Quote:

> [snip]

>Sense you seem pretty famalier with this or at least alot more famalier
>then I am, do you have a suggested starting point?  URL for the GCC
>internals, etc?

Well, RTL (the intermediate language) is covered by documents in
the gcc distribution itself (assuming you can grab the latest GCC
tarball, check out: gcc-2.95.1/gcc/rtl.texi, you'll need TeX, of
course...).  I think that would be a logical place to start, since
it provides a starting point to assess the feasability of handling,
eg, PL/1 types.  Another starting point might be the GCC home page,
at: http://www.gnu.org/software/gcc/.

I'll see if I can dig around and find some more information.

Quote:
>>Oh yeah?  Wow, that sounds really cool.  I have to go check
>>out the Multics newsgroup again...

>Its sortof a low level softsell, each time I suggest it, I get less nay
>sayers and more people who think it might be worth doing and/or doable.

Cool!  I always thought that Multics was very hip.  Has anyone
poked at Bull recently to see if they'll release the actual code?
The Multician's site seems to indicate that there are only three
sites still running the software; it would seem that they have
very little commercially to gain by keeping the code locked away.
:-(

        - Dan C.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBN9VtG4shWrlwdTE9AQGbAAP+OoLL7accZp/XdRAgTVxGp8DxcqHNrFWS
4dmCgLUb9oYwdTMEPRXL6rSoM5VWv5EqpP9g/bAxJWzOZ3dMxspODv5nFFy/LhIK
aQlVNdJAZjvkhOHvfUytqqZxnJ6YBGZfFBBHBdb079Hhxzpb6rjMBsXgDd95TjOd
w2qkZ004O0Q=
=72Oi
-----END PGP SIGNATURE-----



Sat, 23 Feb 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I
Dont know if anyones interested, but I do have a 60%-70 complete PL/I
compiler that I half ported to NT a few years ago. I never got much of
a reaction to this work at the time, so I shelved the thing.

A compiler is one thing, but in this day and age one is expected to
have an editor, de{*filter*} etc, and although such things can readily be
writ{*filter*} for NT, its still a lot of additional work.

The definition that I used for PL/I is the ANSI X3.74-1987 document.
This is a formal definition of the syntax, grammar and semantics of a
large subset of the well known 'IBM PL/I' and should not be confused
with the Subset-G version, this manual is abstruse and daunting in
places.

The compiler is reasonably well written and has a hand coded recursive
descent parser, since PL/I lack of reserved words meant that many
parser generators were unable to handle it (well that was the case 8
years ago).

It generates  a DOS OBJ file, and this links and creates executable
code. However it needs to have an NT COFF obj file, and this is not a
quick fix !

The NT COFF obj file format is well documented and somewhat easier to
work with than the crude and awkward DOS version.

Its not a complete compiler therefore but does represent some 60%-70%
of such a project. The compiler is written (well) in 'C'.

Hugh

Quote:




>> Also, what incentive is there for someone to program in
>> PL/1 if they must pay for the compiler, and *good* compilers
>> for other languages, such as C, C++, FORTRAN, Ada-95, Pascal,
>> Objective C, Java, Perl*, SML, etc, are freely available?
>> If PL/1 is to survive, there *must* be a free compiler
>> available in order to spark interest.

>There is.  It's Digital research PL/I for DOS.
>It's free for non-commercial use.

>Location is in the FAQ.

>>  Open source would
>> be best, as well.

>> Someone had mentioned a possible front end to GCC.  Why don't
>> some of the PL/1 experts here weigh in and lend a hand on that
>> project?  I believe that it could certainly be worthwhile.

>>        - Dan C.



Wed, 06 Mar 2002 03:00:00 GMT  
 What is the definitive language reference for PL/I

Quote:

> The definition that I used for PL/I is the ANSI X3.74-1987 document.
> This is a formal definition of the syntax, grammar and semantics of a
> large subset of the well known 'IBM PL/I' and should not be confused
> with the Subset-G version, this manual is abstruse and daunting in
> places.

What do you mean "confused"? That document *is* the current definition
of Subset G. ANSI decided not to update the standard for the full
language. I'm sure that some of those who were involved on the SHARE
side can give a blow-by-blow description, but probably not in polite
language. The updated Subset G is still missing a lot of critical
functionality, so as far as I'm concerned the defining documents are the
old PL/I standard and the IBM compilers.

--

Shmuel (Seymour J.) Metz
Reply to host nsf (dot) gov, user smetz



Sat, 09 Mar 2002 03:00:00 GMT  
 
 [ 16 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Definitive reference for the LOGO Programming Language?

2. Definitive PL/I texts listing..

3. Derivation of PL/I (was Usenet group for PL/M language)

4. Derivation of PL/I (was Usenet group for PL/M language)

5. Needed Language Reference in various languages!

6. PL/M reference guide?

7. Good Introduction / Reference Book on PL/1

8. DR PL/I Reference Manual

9. A good PL/1 reference please.

10. pointer reference to based variable, and cobol to PL/I

11. PL/1-DOS References

12. PL/1 book references wanted

 

 
Powered by phpBB® Forum Software