java again really

[Les]

On Thu, 2008-01-03 at 11:06 -0700, Craig White wrote:
> On Thu, 2008-01-03 at 09:56 -0800, Knute Johnson wrote:
> > >On Wed, 2008-01-02 at 20:02 -0800, Knute Johnson wrote:
> > >> >Knute Johnson wrote:
> > >> >> To install and run Sun Java on F8 look at the simplified instructions 
> > >> >> on this site;
> > >> >>
> > >> >> http://rabbitbrush.frazmtn.com/sun-java-on-F8.html
> > >> >>
> > >> >> You don't need to remove anything before installing Sun Java.  JEdit 
> > >> >> should work with icedtea anyway.
> > >> >>
> > >> >>   
> > >> >    Some of the things the page says are wrong. But it will be easy now 
> > >> >with sim links to use the right java and jedit will work. By the way, I 
> > >> >tried it with icedtea and it did not work at ALL!
> > >> 
> > >> What is wrong with the document?
> > >----
> > >specifically, nothing, generally, you are really only plucking the low
> > >hanging fruit of a i386 install with emphasis on getting functional use
> > >of the sun java plugin running in Firefox.
> > >
> > >If someone has ambitions beyond that, your instructions are of little
> > >value. We have already covered this ground.
> > >
> > >Craig
> > 
> > We haven't covered any ground before Craig.
> ----
> https://www.redhat.com/archives/fedora-list/2007-December/msg03991.html
> ----
> > The document is not a step by step tutorial for the totally 
> > uninformed.  It is a minimal set of instructions to get Sun's Java 
> > installed on both i386 and x86_64 and to get them to work.
> ----
> to work yes, but primarily as a web browser plugin, which really only
> works for i386 anyway. The issue though is that other steps are
> necessary for using jre (or more likely, jdk) with other applications.
> ----
> >   I don't 
> > know what Karl's command line problem is but it does work just fine 
> > for me.  Part of his problem was that he removed icedtea and didn't 
> > have a clue about the alternatives program.
> ----
> indeed...we can only point out parts of Karl's problems and those are
> clearly some of them.
> 
> Craig
> 
I supported a Solaris network for several years, and have been in here
on these forums for about 2 years now, and this is the first mention I
have heard of "Alternatives".  I don't think that makes me stupid,
either.  I have not had problems getting things to work, and I do
understand how links work.  But I do not now understand the alternatives
command, nor why it is better.  I am preparing to move three systems to
F8, prior to moving my whole home network to F8, so if this is the best
tool for the job, I will need to research it, and look at what it
accomplishes, and if it opens up any security holes.  Each of these
tools has the potential to open your system to a new means of attack.
And thank you very much, I have had it with my systems being hacked,
wormed, virused, and abused.  That is why I have made every effort to
leave Microsoft completely behind, and if RedHat and Fedora begin to
open significant holes in this OS, then I will move on to the next OS
that will do what I need.

	I did a degree via University of Phoenix on IT and the alternatives
command was not mentioned there either.  
	
	Now none of that makes alternatives obscure, but it certainly casts
some doubt on the "common knowledge" terms being applied to it, I would
think.

	It is not common knowledge unless a large percentage, certainly more
than 60%, are aware of it in the normal course of business.  I do not
think alternatives meets that criteria, at least not by the folks here,
and these two groups are certainly not representative of the "common
user".  From what I have observed, most are active and at the leading
edge of the technology curve.  Thus if most of this group doesn't know
of it, it could hardly be common knowledge, don't you agree?  

	So then the discussion turns to whether it should be the common
practice to use it to set up the symbolic links for programs.  And how
this may impact other really common knowledge and common practices.
Virtually every application I have used sets the PATH environment
variable so that commands within that shell reference the desired
functions.  Most programming environments utilize one or more methods to
access the Libraries (although I am told, and have seen LD_LIBRARY_PATH
discouraged as an archaic practice, although I do not now remember the
current means of controlling which libraries are accessed for
programming), such as LD_LIBRARY_PATH or $APP_LIBRARIES.

	Standards are required for several reasons.  They establish rules for
software creators to follow to ensure compatibilities with existing code
and programs.  They make it easier for new users and developers to
achieve success.  And they make system maintenance across various
installations more robust and less likely for a new hire to muck things
up.  They reduce duplication of libraries, memory usage, disk usage and
so forth because common tools are properly installed and accessed.  They
also make fixing problems simpler, since what works to solve the issue
in one locale work in others as well.  Also things like patching and
updating software becomes simpler and more robust.  In general life is
better with standards.  But some developers chafe at not having "edit"
invoke their favorite editor, or "ed" being their choice to avoid typing
the full name.  Thus we get ed, which was originally a very simple line
editor, being replaced in some instances with some other editor.  Which
works fine until you actually need ed to solve a particular problem.
(note that I picked ed because it was a favorite of us old timers who
actually used line editors for some period, not because it is a typical
problem today, rather it is just an example.)

	So I think alternatives is a good idea, and it seems well implemented,
but it is new to me and I will have to study it for a bit to ensure I
know what it does and how it may break things for me, or what new
advantages it gives me.

	This study requires documentation, and not just the man page.

	Someone said I could enter stuff into bugzilla.  But all I posted was
an observation from reading through the man page one time.  I may submit
some later if I decide to use it and run into things that require a
bugzilla.  I do not think it necessary to submit a bugzilla about
including a reference in the man page.  That should be standard (there's
that S word again) practice and common knowledge (oops thare is that CK
term again) among system designers and coders. 

Regards,
Les H