**
Hello,

We are going to install ARS v6.3 on an HP-UX 11.i system with Oracle 9i. Is there any recommendations about Mid-Tier platform installation?. We are thinking to use :

* Microsoft W2000 Server
* Microsoft IIS
* Sun Java SDK 1.4.2

Thanks in advance. Regards,

Pilar.


This posting was submitted via the Web interface

**
You would be better off using Windows Server 2003 than 2000 since IIS 6.0 is much more secure than IIS 5.0. We have JDK 1.4.203 on our mid-tier 6.3 servers (on Win2K3) and 1.4.206 on our ARS 6.3 servers but have had problems with several other 1.4.2 distributions. I can only recommend 03 or 06. You would most likely be using ServletExec 5.0 instead of 4.1 since it installs with mid-tier 6.3, which is also an improvement.
Christopher Strauss, MSLS
Remedy Database Administrator
University of North Texas Computing Center
http://remedy.unt.edu/




From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Pilar Llorente Beltran
Sent: Tuesday, June 28, 2005 10:24 AM
To: arslist@ARSLIST.ORG
Subject: Installing Remedy Mid-Tier on a new platform


**
Hello,

We are going to install ARS v6.3 on an HP-UX 11.i system with Oracle 9i. Is there any recommendations about Mid-Tier platform installation?. We are thinking to use :

* Microsoft W2000 Server
* Microsoft IIS
* Sun Java SDK 1.4.2

Thanks in advance. Regards,

Pilar.


This posting was submitted via the Web interface
This posting was submitted via the Web interface

I am always highly sceptical of remarks concerning problems with Java 0x
releases, given they usually contain the minimum of bug fixes that are
mostly irrelevant to the majority if users.

Sadly, I can not find a list of bugs fixed between 02 and 03 to hand,
given Sun's website is pretty awful to navigate.

However, I'd always recommend the latest major Java version while suspecting
that "problems seen with various minor releases" are caused by something
else. The Midtier uses only the most basic components of the Java language
and I'm very doubtful to whether those components would be buggy!

To put this into context, the XML Gateway uses a wider range of APIs than
the Midtier 6.3 does, and we are comfortable with any 1.4.2 VM :)


John

Java System Solutions : http://www.javasystemsolutions.com


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

Real world experience and speculation are two different things though.
Take it for what it's worth. Chris has always been very thorough in
testing the behaviors and misbehaviors between java releases. Most of
the problems he experiences though are related to the compatability of
the server side java and the platform/JRE on the client side, not to
mention the number of browsers Chris has to deal with.

Axton

On 6/29/05, John Baker wrote:
> I am always highly sceptical of remarks concerning problems with Java 0x
> releases, given they usually contain the minimum of bug fixes that are
> mostly irrelevant to the majority if users.
>
> Sadly, I can not find a list of bugs fixed between 02 and 03 to hand,
> given Sun's website is pretty awful to navigate.
>
> However, I'd always recommend the latest major Java version while suspecting
> that "problems seen with various minor releases" are caused by something
> else. The Midtier uses only the most basic components of the Java language
> and I'm very doubtful to whether those components would be buggy!
>
> To put this into context, the XML Gateway uses a wider range of APIs than
> the Midtier 6.3 does, and we are comfortable with any 1.4.2 VM :)
>
>
> John
>
> Java System Solutions : http://www.javasystemsolutions.com
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
> (Support: mailto:support@arslist.org)
>


--
Axton Grams
AxtonGrams<åt>gmailcom


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

**

John:

Unfortunately, I have to disagree with your opinion. I've found Java programs/code that are very particular which version of Java you use. For instance, 1.4.206 is just fine, whereas 1.4.208 causes no end of problems. This is because SUN places additional code which is backported from either the 1.5 or even 1.6 efforts into the 1.4 tree. A good example of this is the ServletExec JSP Engine that ships with Mid-Tier 6.3 which will not run with Java 1.4.208 but will with Java 1.4.206. This fact is documented on the ServletExec FAQ web page. However, I will agree that upgrading or changing from one subversion (0x) to another of Java should not 'break' programs, but it can and in this case does.

James McKenzie
Remedy Engineer
C-E LCMC HQISEC-ESED
L-3 GSI


-----Original Message-----
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of John Baker
Sent: Wednesday, June 29, 2005 12:48 AM
To: arslist@ARSLIST.ORG
Subject: Re: Installing Remedy Mid-Tier on a new platform


I am always highly sceptical of remarks concerning problems with Java 0x releases, given they usually contain the minimum of bug fixes that are mostly irrelevant to the majority if users.

Sadly, I can not find a list of bugs fixed between 02 and 03 to hand, given Sun's website is pretty awful to navigate.

However, I'd always recommend the latest major Java version while suspecting that "problems seen with various minor releases" are caused by something else. The Midtier uses only the most basic components of the Java language and I'm very doubtful to whether those components would be buggy!

To put this into context, the XML Gateway uses a wider range of APIs than the Midtier 6.3 does, and we are comfortable with any 1.4.2 VM :)


John

Java System Solutions : http://www.javasystemsolutions.com


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

This posting was submitted via the Web interface

Axton: I've been compiling Java apps every day since the days of 1.0.2, back
when my Sparc 2 could only just manage to run 'Hello World' .. :)

James: That's very nice, but until I see the Eclipse project telling me I
can't run Eclipse on the majority of minor 1.4.2 releases then I'm afraid I
have to blame NewAtlanta/Remedy.

In general, and I'm sure there are exceptions, there is nothing wrong with
the latest minor revision. If coders wish to write non-platform independent
or low quality code rely on deprecated APIs, then that's their fault. I may
have posted this AR API example before, but let's examine it again:

public void releaseInstance(Object o)
{ o = null; }

This is present in the AR API and developers are encouraged to use it. The
fact that someone actually wrote that tells me that the the person who wrote
the Java AR API; something which I would consider quite important; could not
write Java. Would you not agree that such a cowboy developer should not be
allowd near a keyboard?

It's precisely this level of incompetence that results in poor applications
that don't conform to standards while making use of deprecated/broken APIs.
The XML Gateway uses the AR API, and I certainly haven't noticed any
problems on different minor revisions. What makes the XML Gateway's use of
AR so different to the MT? Both create and modify entries; both retrieve
workflow information? The XML Gw makes more use of the J2EE components than
the MT does, so why isn't it flaky on 6 out of 8 minor 1.4.2 revisions?

I am curious. How have previous posters managed to conclude that the MT6.3
isn't making use of a bunch of deprecated API calls, which could easily
result in unpredictable results across different VMs? I'd hope it made none,
but what about ServletExec?

As for ServletExec, I see no reason why anyone would even install it. Tomcat
is the accepted 'de facto' engine, which is precisely why it's so popular.
Given it is so popular, perhaps we should ask the question: Why do Remedy
use something that the majority shun?

But there again, I often wonder why people use IIS over Apache. Given the
overwhelming majority use Apache, you'd expect the people using IIS to
question it's existence :-)


John

Java System Solutions : http://www.javasystemsolutions.com


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

John,
as someone who is not as knowledgeable in java as you...would you
mind explaining what
public void releaseInstance(Object o)
{ o = null; }
Does and why it's so bad?


-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:arslist@ARSLIST.ORG] On Behalf Of John Baker
Sent: Thursday, June 30, 2005 2:30 AM
To: arslist@ARSLIST.ORG
Subject: Re: Installing Remedy Mid-Tier on a new platform

Axton: I've been compiling Java apps every day since the days of 1.0.2, back
when my Sparc 2 could only just manage to run 'Hello World' .. :)

James: That's very nice, but until I see the Eclipse project telling me I
can't run Eclipse on the majority of minor 1.4.2 releases then I'm afraid I
have to blame NewAtlanta/Remedy.

In general, and I'm sure there are exceptions, there is nothing wrong with
the latest minor revision. If coders wish to write non-platform independent
or low quality code rely on deprecated APIs, then that's their fault. I may
have posted this AR API example before, but let's examine it again:

public void releaseInstance(Object o)
{ o = null; }

This is present in the AR API and developers are encouraged to use it. The
fact that someone actually wrote that tells me that the the person who wrote
the Java AR API; something which I would consider quite important; could not
write Java. Would you not agree that such a cowboy developer should not be
allowd near a keyboard?

It's precisely this level of incompetence that results in poor applications
that don't conform to standards while making use of deprecated/broken APIs.
The XML Gateway uses the AR API, and I certainly haven't noticed any
problems on different minor revisions. What makes the XML Gateway's use of
AR so different to the MT? Both create and modify entries; both retrieve
workflow information? The XML Gw makes more use of the J2EE components than
the MT does, so why isn't it flaky on 6 out of 8 minor 1.4.2 revisions?

I am curious. How have previous posters managed to conclude that the MT6.3
isn't making use of a bunch of deprecated API calls, which could easily
result in unpredictable results across different VMs? I'd hope it made none,
but what about ServletExec?

As for ServletExec, I see no reason why anyone would even install it. Tomcat
is the accepted 'de facto' engine, which is precisely why it's so popular.
Given it is so popular, perhaps we should ask the question: Why do Remedy
use something that the majority shun?

But there again, I often wonder why people use IIS over Apache. Given the
overwhelming majority use Apache, you'd expect the people using IIS to
question it's existence :-)


John

Java System Solutions : http://www.javasystemsolutions.com



UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

In response to "...If coders wish to write non-platform independent or
low quality code rely on deprecated APIs, then that's their fault. I
may have posted this AR API example before, but let's examine it
again:

public void releaseInstance(Object o)
{ o = null; }

This is present in the AR API and developers are encouraged to use it.
The fact that someone actually wrote that tells me that the the person
who wrote the Java AR API; something which I would consider quite
important; could not write Java. Would you not agree that such a
cowboy developer should not be allowd near a keyboard?"

Well, this is precisely why we have something like the ARSList and I
am thankful to the people willing to share their findings to keep
someone else from shooting themselves in the foot.

If you like to learn things the hard way that's your business. If you
want to tap into the experiences of other users then read the
postings.

I won't argue with you that there is a lot of poor code out there and
the use of deprecated features of api's and databases (among others)
seems to be common practice. Take for instance the use of a LONG RAW
datatype in Oracle. This was deprecated with the release of Oracle 9,
but it is still used. This should have been converted to use a BLOB
or BFILE a long time ago.

Axton


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)

L,

That Java function cleared the local reference o to null. I assume the
author of that code meant to free the reference to the object, but didn't
seem to understand Java has a garbage collector and hence the function did
nothing. So for someone to write something so ridiculous indicates that
the person did not have the faintest clue how Java memory management
works.

Once upon a time, I had to explain this to someone in Remedy Engineering
with an actual sample of Java code!

It just isn't useful to have people with such a poor grasp of development
knowledge being let loose on a commercial product that others rely on.
Sadly, Remedy are just one of many large corporates guilty of having too
many "managers without technical skills" in positions that demand "people
with technical skills" :)


John

Java System Solutions : http://www.javasystemsolutions.com


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
(Support: mailto:support@arslist.org)