RE: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-06 Thread David Sommers
Sun and IBM haven't been playing well together.  I believe IBM is
currently working on an open-source version of the Java Runtime.  The
problem is - the Java spec itself runs thru the Java Committee and that
Committee is mostly chaired by Sun.  So... it'll be hard to be purely
open but at least there are non-Sun VM options.

/David.
 

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Cameron Schlehuber
Sent: Saturday, December 04, 2004 10:03 PM
To: [EMAIL PROTECTED]
Subject: RE: [Hardhats-members] GT.M inside (was Krung Thai Bank goes
live on GT.M)

The emphasis in the VA is on POJO's Plain Old Java Objects.
It looks like Hibernate will be used for SQL access.

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Nancy
E.
Anthracite
Sent: Saturday, December 04, 2004 4:14 PM
To: [EMAIL PROTECTED]
Subject: Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes
live
on GT.M)

I just got off the phone with my son who informed me that Sun's Java has
a 
restricted license  - look but don't touch - but there are truly open
source

versions of Java.  So is the VA making any distinction in what they are 
using?  Does it matter to us if we start writing GUI plugins for Pedi
that 
are written in Java in hopes they will eventually be integrated with
what
the 
VA is working on? 

Anybody know anything about Java Script licensing?  Does it matter?




---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members


---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members


open source Java Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-06 Thread Joseph Dal Molin
Brasil is heavily into Java and are at the core of an effort to create 
an open source Java I'm not exactly sure what that entails but for 
any that are interested I can connect you offline to one of key 
participants in that initiative...

Joseph
David Sommers wrote:
Sun and IBM haven't been playing well together.  I believe IBM is
currently working on an open-source version of the Java Runtime.  The
problem is - the Java spec itself runs thru the Java Committee and that
Committee is mostly chaired by Sun.  So... it'll be hard to be purely
open but at least there are non-Sun VM options.
/David.
 

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Cameron Schlehuber
Sent: Saturday, December 04, 2004 10:03 PM
To: [EMAIL PROTECTED]
Subject: RE: [Hardhats-members] GT.M inside (was Krung Thai Bank goes
live on GT.M)
The emphasis in the VA is on POJO's Plain Old Java Objects.
It looks like Hibernate will be used for SQL access.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Nancy
E.
Anthracite
Sent: Saturday, December 04, 2004 4:14 PM
To: [EMAIL PROTECTED]
Subject: Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes
live
on GT.M)
I just got off the phone with my son who informed me that Sun's Java has
a 
restricted license  - look but don't touch - but there are truly open
source

versions of Java.  So is the VA making any distinction in what they are 
using?  Does it matter to us if we start writing GUI plugins for Pedi
that 
are written in Java in hopes they will eventually be integrated with
what
the 
VA is working on? 

Anybody know anything about Java Script licensing?  Does it matter?

---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members

---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members

.

---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members


Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-04 Thread K.S. Bhaskar
Maury --

There are several cases of linking to consider with GT.M for x86
GNU/Linux.

The most common is the default, when a process executes a Do ^XYZ or a
ZLINK XYZ.  This is dynamically linked.

With GT.M the top level of a process does not have to be M code, but can
be in C (or another language).  Any calls from C to M are dynamically
linked, as are calls from M to C.  C to C calls can be dynamically
linked, or if two C modules are in the a shared library, they are
statically linked to each other.  Two C modules in different shared
libraries are dynamically linked.

Thus, for GT.M for x86 GNU/Linux, any code in M is always dynamically
linked, and code in C is dynamically linked with code in M, and for two
modules in C, it depends on how they are packaged.

With GT.M on AIX, HP-UX and Tru64 UNIX, where code in M can be put into
shared libraries, the same rule applies to M modules as to C modules on
x86 GNU/Linux, namely that two M modules in the same shared library are
considered statically linked to each other, and dynamically linked to
anything not in that shared library.

So, bottom line is that at least as far as GT.M is concerned, linking is
always dynamic unless someone explicitly chooses to make it static. 
This means that if someone makes an add on to VistA under the GPL, and
someone else makes another add on to VistA that is under a license
incompatible with the GPL, the fact that linking is dynamic means that
the GPL license of one module does not affect the GPL-incompatible
license of another module.

Your question was very incisive, and I guess the answer means that much
of our licensing discussion was unnecessary!  I am copying the
vista-open-source at yahoogroups.com, since much of the licensing
discussion happened over there.

Hope this explains things satisfactorily.  If not, please ask.

-- Bhaskar

On Fri, 2004-12-03 at 21:42, Maury Pepper wrote:
 Bhaskar,
 
 As long as you have the hood up, I'd like to ask about GT.M with respect to 
 linking.  As you know, the recent conversations about open source licensing 
 touched on the issue of static versus dynamic linking.  This seems to be the 
 line that separates contaminated from uncontaminated code.  (Excuse the 
 euphemisms.)  So, what I would like is your explanation of the linking that 
 exists between routines from two different packages running in a single 
 user's job space.  For example, if you are running FOIA VistA and an add-on 
 accounting package together, how might they be linked?  If there are multiple 
 answers, that's OK.  I think this is relevant -- because, based on my 
 understanding of what is going on, I think we may need to redefine certain 
 terms in an M context because I'm not sure that the industry standard terms 
 apply.
  
 -maury-
 
 
 - Original Message - 
 From: K.S. Bhaskar [EMAIL PROTECTED]
 To: [EMAIL PROTECTED]
 Sent: Friday, December 03, 2004 8:12 PM
 Subject: RE: [Hardhats-members] Krung Thai Bank goes live on GT.M
 
 
 Steve --
 
 There are some differences down in the detail level, mostly as a result
 of differences between hardware architectures, operating systems, as
 well as what development found funding and what did not.
 
 There are differences between GT.M on Alpha/VMS and GT.M on UNIX/Linux: 
 the underlying OS platforms are different enough that we have different
 manuals for them - although the manuals are generated from a common
 source with conditional text.  The UNIX implementations support
 identical implementations of the M language, but here are some examples
 of the differences:
 
   * Some platforms support a Direct IO flag which turns on the
 O_DIRECT setting for journal IO, which can either speed things
 up or slow things down depending on the IO subsystem.
 
   * GT.M on Sun SPARC Solaris supports Sun RPC call-ins.  The others
 don't.
 
   * The GT.M compiler on AIX, HP-UX and Tru64 UNIX creates object
 files in a format that can be incorporated into .so shared
 libraries.  The GT.M compiler on Linux and Solaris does not.
 
   * There may be differences in the ability to pass parameters in
 registers when calling between M and C on the different
 platforms, but I can't remember right now.
 
   * A GT.M process on AIX can have fewer database files open at one
 time than on other platforms (the limit is something like 9
 caused by fact that each shared memory segment uses a segment
 register).
 
 -- Bhaskar
 
 On Fri, 2004-12-03 at 18:44, Tomlinson, ,Steven B wrote:
  Aloha Bhaskar,
  Thanks for the clarification, I have been wondering what (if any)
  differences there were between the GT.M distributions.
  
  Steven B. Tomlinson
  [EMAIL PROTECTED]
  Pacific Telehealth and Technology Hui
  www.PacificHui.org
 
 ***
 This electronic mail transmission contains confidential and/or privileged 
 

Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-04 Thread Nancy E. Anthracite
I just got off the phone with my son who informed me that Sun's Java has a 
restricted license  - look but don't touch - but there are truly open source 
versions of Java.  So is the VA making any distinction in what they are 
using?  Does it matter to us if we start writing GUI plugins for Pedi that 
are written in Java in hopes they will eventually be integrated with what the 
VA is working on? 

Anybody know anything about Java Script licensing?  Does it matter?

On Saturday 04 December 2004 01:48 pm, K.S. Bhaskar wrote:
 Maury --

 There are several cases of linking to consider with GT.M for x86
 GNU/Linux.

 The most common is the default, when a process executes a Do ^XYZ or a
 ZLINK XYZ.  This is dynamically linked.

 With GT.M the top level of a process does not have to be M code, but can
 be in C (or another language).  Any calls from C to M are dynamically
 linked, as are calls from M to C.  C to C calls can be dynamically
 linked, or if two C modules are in the a shared library, they are
 statically linked to each other.  Two C modules in different shared
 libraries are dynamically linked.

 Thus, for GT.M for x86 GNU/Linux, any code in M is always dynamically
 linked, and code in C is dynamically linked with code in M, and for two
 modules in C, it depends on how they are packaged.

 With GT.M on AIX, HP-UX and Tru64 UNIX, where code in M can be put into
 shared libraries, the same rule applies to M modules as to C modules on
 x86 GNU/Linux, namely that two M modules in the same shared library are
 considered statically linked to each other, and dynamically linked to
 anything not in that shared library.

 So, bottom line is that at least as far as GT.M is concerned, linking is
 always dynamic unless someone explicitly chooses to make it static.
 This means that if someone makes an add on to VistA under the GPL, and
 someone else makes another add on to VistA that is under a license
 incompatible with the GPL, the fact that linking is dynamic means that
 the GPL license of one module does not affect the GPL-incompatible
 license of another module.

 Your question was very incisive, and I guess the answer means that much
 of our licensing discussion was unnecessary!  I am copying the
 vista-open-source at yahoogroups.com, since much of the licensing
 discussion happened over there.

 Hope this explains things satisfactorily.  If not, please ask.

 -- Bhaskar

 On Fri, 2004-12-03 at 21:42, Maury Pepper wrote:
  Bhaskar,
 
  As long as you have the hood up, I'd like to ask about GT.M with respect
  to linking.  As you know, the recent conversations about open source
  licensing touched on the issue of static versus dynamic linking.  This
  seems to be the line that separates contaminated from uncontaminated
  code.  (Excuse the euphemisms.)  So, what I would like is your
  explanation of the linking that exists between routines from two
  different packages running in a single user's job space.  For example, if
  you are running FOIA VistA and an add-on accounting package together, how
  might they be linked?  If there are multiple answers, that's OK.  I think
  this is relevant -- because, based on my understanding of what is going
  on, I think we may need to redefine certain terms in an M context because
  I'm not sure that the industry standard terms apply.
 
  -maury-
 
 
  - Original Message -
  From: K.S. Bhaskar [EMAIL PROTECTED]
  To: [EMAIL PROTECTED]
  Sent: Friday, December 03, 2004 8:12 PM
  Subject: RE: [Hardhats-members] Krung Thai Bank goes live on GT.M
 
 
  Steve --
 
  There are some differences down in the detail level, mostly as a result
  of differences between hardware architectures, operating systems, as
  well as what development found funding and what did not.
 
  There are differences between GT.M on Alpha/VMS and GT.M on UNIX/Linux:
  the underlying OS platforms are different enough that we have different
  manuals for them - although the manuals are generated from a common
  source with conditional text.  The UNIX implementations support
  identical implementations of the M language, but here are some examples
  of the differences:
 
* Some platforms support a Direct IO flag which turns on the
  O_DIRECT setting for journal IO, which can either speed things
  up or slow things down depending on the IO subsystem.
 
* GT.M on Sun SPARC Solaris supports Sun RPC call-ins.  The others
  don't.
 
* The GT.M compiler on AIX, HP-UX and Tru64 UNIX creates object
  files in a format that can be incorporated into .so shared
  libraries.  The GT.M compiler on Linux and Solaris does not.
 
* There may be differences in the ability to pass parameters in
  registers when calling between M and C on the different
  platforms, but I can't remember right now.
 
* A GT.M process on AIX can have fewer database files open at one
  time than on other platforms (the limit 

Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-04 Thread Kevin Toppenberg
Bhaskar,

This may be a bit off topic (re licensing), but it
does relate to the linking.

I recently divided my program up into modules, with a
bunch of debug code all grouped into one TMGDEBUG.m
file.  When I did this, I noticed that my execution
speed went way down.  It seemed that the module is
being loaded and dumped with every call.  Is this
true?

Kevin


--- K.S. Bhaskar [EMAIL PROTECTED] wrote:

 Maury --
 
 There are several cases of linking to consider with
 GT.M for x86
 GNU/Linux.
 
 The most common is the default, when a process
 executes a Do ^XYZ or a
 ZLINK XYZ.  This is dynamically linked.
 
 With GT.M the top level of a process does not have
 to be M code, but can
 be in C (or another language).  Any calls from C to
 M are dynamically
 linked, as are calls from M to C.  C to C calls can
 be dynamically
 linked, or if two C modules are in the a shared
 library, they are
 statically linked to each other.  Two C modules in
 different shared
 libraries are dynamically linked.
 
 Thus, for GT.M for x86 GNU/Linux, any code in M is
 always dynamically
 linked, and code in C is dynamically linked with
 code in M, and for two
 modules in C, it depends on how they are packaged.
 
 With GT.M on AIX, HP-UX and Tru64 UNIX, where code
 in M can be put into
 shared libraries, the same rule applies to M modules
 as to C modules on
 x86 GNU/Linux, namely that two M modules in the same
 shared library are
 considered statically linked to each other, and
 dynamically linked to
 anything not in that shared library.
 
 So, bottom line is that at least as far as GT.M is
 concerned, linking is
 always dynamic unless someone explicitly chooses to
 make it static. 
 This means that if someone makes an add on to VistA
 under the GPL, and
 someone else makes another add on to VistA that is
 under a license
 incompatible with the GPL, the fact that linking is
 dynamic means that
 the GPL license of one module does not affect the
 GPL-incompatible
 license of another module.
 
 Your question was very incisive, and I guess the
 answer means that much
 of our licensing discussion was unnecessary!  I am
 copying the
 vista-open-source at yahoogroups.com, since much of
 the licensing
 discussion happened over there.
 
 Hope this explains things satisfactorily.  If not,
 please ask.
 
 -- Bhaskar
 
 On Fri, 2004-12-03 at 21:42, Maury Pepper wrote:
  Bhaskar,
  
  As long as you have the hood up, I'd like to ask
 about GT.M with respect to linking.  As you know,
 the recent conversations about open source licensing
 touched on the issue of static versus dynamic
 linking.  This seems to be the line that separates
 contaminated from uncontaminated code.  (Excuse
 the euphemisms.)  So, what I would like is your
 explanation of the linking that exists between
 routines from two different packages running in a
 single user's job space.  For example, if you are
 running FOIA VistA and an add-on accounting package
 together, how might they be linked?  If there are
 multiple answers, that's OK.  I think this is
 relevant -- because, based on my understanding of
 what is going on, I think we may need to redefine
 certain terms in an M context because I'm not sure
 that the industry standard terms apply.
   
  -maury-
  
  
  - Original Message - 
  From: K.S. Bhaskar [EMAIL PROTECTED]
  To: [EMAIL PROTECTED]
  Sent: Friday, December 03, 2004 8:12 PM
  Subject: RE: [Hardhats-members] Krung Thai Bank
 goes live on GT.M
  
  
  Steve --
  
  There are some differences down in the detail
 level, mostly as a result
  of differences between hardware architectures,
 operating systems, as
  well as what development found funding and what
 did not.
  
  There are differences between GT.M on Alpha/VMS
 and GT.M on UNIX/Linux: 
  the underlying OS platforms are different enough
 that we have different
  manuals for them - although the manuals are
 generated from a common
  source with conditional text.  The UNIX
 implementations support
  identical implementations of the M language, but
 here are some examples
  of the differences:
  
* Some platforms support a Direct IO flag
 which turns on the
  O_DIRECT setting for journal IO, which can
 either speed things
  up or slow things down depending on the IO
 subsystem.
  
* GT.M on Sun SPARC Solaris supports Sun RPC
 call-ins.  The others
  don't.
  
* The GT.M compiler on AIX, HP-UX and Tru64
 UNIX creates object
  files in a format that can be incorporated
 into .so shared
  libraries.  The GT.M compiler on Linux and
 Solaris does not.
  
* There may be differences in the ability to
 pass parameters in
  registers when calling between M and C on
 the different
  platforms, but I can't remember right now.
  
* A GT.M process on AIX can have fewer
 database files open at one
  time than on other platforms (the limit is
 something like 9
  caused by fact that each 

Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-04 Thread K.S. Bhaskar
Kevin --

GT.M automatically links a module on the first call, but not thereafter,
unless you do an explicit relink with a ZLink.  The cause of the
performance hit lies elsewhere.

Perhaps GT.M's trace capability can help?  For example, try:

  View TRACE:1:^Trace($Job)
... whatever ...
  View TRACE:0

Followed by a dump of ^Trace($Job).  Of course tracing itself has an
overhead, but it might explain where your code is spending its efforts.

-- Bhaskar

On Sat, 2004-12-04 at 18:37, Kevin Toppenberg wrote:
 Bhaskar,
 
 This may be a bit off topic (re licensing), but it
 does relate to the linking.
 
 I recently divided my program up into modules, with a
 bunch of debug code all grouped into one TMGDEBUG.m
 file.  When I did this, I noticed that my execution
 speed went way down.  It seemed that the module is
 being loaded and dumped with every call.  Is this
 true?
 
 Kevin
 
 
 --- K.S. Bhaskar [EMAIL PROTECTED] wrote:
 
  Maury --
  
  There are several cases of linking to consider with
  GT.M for x86
  GNU/Linux.
  
  The most common is the default, when a process
  executes a Do ^XYZ or a
  ZLINK XYZ.  This is dynamically linked.
  
  With GT.M the top level of a process does not have
  to be M code, but can
  be in C (or another language).  Any calls from C to
  M are dynamically
  linked, as are calls from M to C.  C to C calls can
  be dynamically
  linked, or if two C modules are in the a shared
  library, they are
  statically linked to each other.  Two C modules in
  different shared
  libraries are dynamically linked.
  
  Thus, for GT.M for x86 GNU/Linux, any code in M is
  always dynamically
  linked, and code in C is dynamically linked with
  code in M, and for two
  modules in C, it depends on how they are packaged.
  
  With GT.M on AIX, HP-UX and Tru64 UNIX, where code
  in M can be put into
  shared libraries, the same rule applies to M modules
  as to C modules on
  x86 GNU/Linux, namely that two M modules in the same
  shared library are
  considered statically linked to each other, and
  dynamically linked to
  anything not in that shared library.
  
  So, bottom line is that at least as far as GT.M is
  concerned, linking is
  always dynamic unless someone explicitly chooses to
  make it static. 
  This means that if someone makes an add on to VistA
  under the GPL, and
  someone else makes another add on to VistA that is
  under a license
  incompatible with the GPL, the fact that linking is
  dynamic means that
  the GPL license of one module does not affect the
  GPL-incompatible
  license of another module.
  
  Your question was very incisive, and I guess the
  answer means that much
  of our licensing discussion was unnecessary!  I am
  copying the
  vista-open-source at yahoogroups.com, since much of
  the licensing
  discussion happened over there.
  
  Hope this explains things satisfactorily.  If not,
  please ask.
  
  -- Bhaskar
  
  On Fri, 2004-12-03 at 21:42, Maury Pepper wrote:
   Bhaskar,
   
   As long as you have the hood up, I'd like to ask
  about GT.M with respect to linking.  As you know,
  the recent conversations about open source licensing
  touched on the issue of static versus dynamic
  linking.  This seems to be the line that separates
  contaminated from uncontaminated code.  (Excuse
  the euphemisms.)  So, what I would like is your
  explanation of the linking that exists between
  routines from two different packages running in a
  single user's job space.  For example, if you are
  running FOIA VistA and an add-on accounting package
  together, how might they be linked?  If there are
  multiple answers, that's OK.  I think this is
  relevant -- because, based on my understanding of
  what is going on, I think we may need to redefine
  certain terms in an M context because I'm not sure
  that the industry standard terms apply.

   -maury-
   
   
   - Original Message - 
   From: K.S. Bhaskar [EMAIL PROTECTED]
   To: [EMAIL PROTECTED]
   Sent: Friday, December 03, 2004 8:12 PM
   Subject: RE: [Hardhats-members] Krung Thai Bank
  goes live on GT.M
   
   
   Steve --
   
   There are some differences down in the detail
  level, mostly as a result
   of differences between hardware architectures,
  operating systems, as
   well as what development found funding and what
  did not.
   
   There are differences between GT.M on Alpha/VMS
  and GT.M on UNIX/Linux: 
   the underlying OS platforms are different enough
  that we have different
   manuals for them - although the manuals are
  generated from a common
   source with conditional text.  The UNIX
  implementations support
   identical implementations of the M language, but
  here are some examples
   of the differences:
   
 * Some platforms support a Direct IO flag
  which turns on the
   O_DIRECT setting for journal IO, which can
  either speed things
   up or slow things down depending on the IO
  subsystem.
   
 * GT.M on Sun SPARC Solaris 

RE: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live on GT.M)

2004-12-04 Thread Cameron Schlehuber
The emphasis in the VA is on POJO's Plain Old Java Objects.
It looks like Hibernate will be used for SQL access.

-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Nancy E.
Anthracite
Sent: Saturday, December 04, 2004 4:14 PM
To: [EMAIL PROTECTED]
Subject: Re: [Hardhats-members] GT.M inside (was Krung Thai Bank goes live
on GT.M)

I just got off the phone with my son who informed me that Sun's Java has a 
restricted license  - look but don't touch - but there are truly open source

versions of Java.  So is the VA making any distinction in what they are 
using?  Does it matter to us if we start writing GUI plugins for Pedi that 
are written in Java in hopes they will eventually be integrated with what
the 
VA is working on? 

Anybody know anything about Java Script licensing?  Does it matter?




---
SF email is sponsored by - The IT Product Guide
Read honest  candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
___
Hardhats-members mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/hardhats-members