Re: File Management

Re: File Management

[Heimo Claasen]

Indeed I do/did not really understand the effects of "make clean" and
"make distclean", Richard (this _is_ a newbie list, isn't it?).
Your explanation helps a bit, thanks.

Thus "make distclean" would do the job like "uninstall", if I
understand that right.
And it depends on the individual programs/source-packages which one
would work, right ?

Now, on this other aspect of weeding out things that had been
installed not by (what I would remember as "admin" or rather, root
user) some specific intention or attention with installing/compiling
one specific source "tar(gz)"-pack; but rather of what has been placed
there when a "distroé" more or less automatically had installed say,
"groups" of packages, e.g. "text tools".

Sure the package manager of _that_ distro ould be of help for
uninstalling too. That works simply enopugh with some "evident"
packages - e.g., if you use one mailer and like it, it's evident that
other mailer-packages could go.

That hoever, is different with the (sometimes quite large) "libxxxx"
packages, or some rather "inevident" named things.

Not being a programmer, nor a systems professional, it is not at all
evident, from the (sometimes arcane) short desciptions those package
managing tools offer, to conclude to the package's meaning and
usefulness or even necessity.

"ldd" there, could give a hint - though, oh! sure!, only if all those
"ldd" outputs then are cross-checked. Which is what I meant with "a
hassle", doing exactly that, manually.<bg>

But couldn't there be some kind of "sorting" prog/algorythm (please
take "sorting" as quite large a term) which could do just that ?

There should be, or at least it's thinkable to be do-able, a "listing"
of "non-shared", so to say "singular" files (or even packages), at which
you could then look at more in detail in order to decide if to delete
them or not.

// Heimo Claasen // <hammer at revobild dot net> // Brussels 2002-09-20
The WebPlace of ReRead - and much to read  ==>  http://www.revobild.net

-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: File Management

[Ray Olszewski]

I'm having a bit of trouble following your question, so let me first try to 
restate it, so we can see if I 'm getting it right.

Right now, the "ldd" program is able to examine a single executable and 
determine which shared libraries it depends on.

I believe you are suggesting that it would be desirable to have an 
application that works like this:

         1. Do an "ldd" examination of every executable on the system.
         2. From the output of (1), compile a union list of all shared 
libraries currently used on the system.
         3. Compare (2) to the actual collection of shared libraries on the 
system.
         4. (Optionally) remove any shared libraries that are not used by 
any executable present.

This is an interesting idea, and it is in principle doable. As far as I 
know, no one has actually done it; the most likely candidate for someone to 
be interested in doing this is one of the big Linux distributors (who might 
find it most useful to tie it in with its own package-management system, as 
a way to gain a competitive edge). Actually, any decent package-management 
system can do this much more easily, by extending a bit the way it handles 
dependencies (though there are "gotcha" risks with this approach if the 
system contains *any* applications installed outside the package manager); 
I think this is how Windows handles its analogous problem with managing DLLs.

Whichever way it were done, getting the details right would be a good bit 
of work. And in practice, though the idea is interesting, I doubt the 
actual utility of the application. The core set of .so libraries is pretty 
stable, and the few specialized ones tend to be associated with 
sufficiently specialized sorts of apps that someone who installs them would 
know what was there.

These comments only address your concern about "the (sometimes quite large) 
"libxxxx" packages". You also worry about "some rather "inevident" named 
things". For these, the straightforward way to learn what the command does 
is to consult its man page (the usual form of this suggestion is the 
familiar "RTFM", but I'm far too polite to give you advice in that form) or 
consult any of the many beginner's guides that exist in electronic and book 
form. Once you know what a command does, you can decide whether or not to 
leave it on the system.

In practice, distribution bloat has become something of a problem in Linux. 
Naive installs of every distro I've looked at recently (including Debian, 
alas) tend to lard up a system with a lot of unneeded stuff (not in the 
form of surplus .so libraries, though -- this is really an applications 
issue, not a shared-library problem). Today's bigger hard disks make it 
less costly to put up with the bloat, but a more interactive install system 
(Slackware used to have a nice one in the old days, around Slack 3.2 or so; 
any chance it still does?) that made it easier to fine-tune an install 
would be a help.

At , Heimo Claasen wrote:
>Indeed I do/did not really understand the effects of "make clean" and
>"make distclean", Richard (this _is_ a newbie list, isn't it?).
>Your explanation helps a bit, thanks.
>
>Thus "make distclean" would do the job like "uninstall", if I
>understand that right.
>And it depends on the individual programs/source-packages which one
>would work, right ?
>
>Now, on this other aspect of weeding out things that had been
>installed not by (what I would remember as "admin" or rather, root
>user) some specific intention or attention with installing/compiling
>one specific source "tar(gz)"-pack; but rather of what has been placed
>there when a "distroŽ" more or less automatically had installed say,
>"groups" of packages, e.g. "text tools".
>
>Sure the package manager of _that_ distro ould be of help for
>uninstalling too. That works simply enopugh with some "evident"
>packages - e.g., if you use one mailer and like it, it's evident that
>other mailer-packages could go.
>
>That hoever, is different with the (sometimes quite large) "libxxxx"
>packages, or some rather "inevident" named things.
>
>Not being a programmer, nor a systems professional, it is not at all
>evident, from the (sometimes arcane) short desciptions those package
>managing tools offer, to conclude to the package's meaning and
>usefulness or even necessity.
>
>"ldd" there, could give a hint - though, oh! sure!, only if all those
>"ldd" outputs then are cross-checked. Which is what I meant with "a
>hassle", doing exactly that, manually.<bg>
>
>But couldn't there be some kind of "sorting" prog/algorythm (please
>take "sorting" as quite large a term) which could do just that ?
>
>There should be, or at least it's thinkable to be do-able, a "listing"
>of "non-shared", so to say "singular" files (or even packages), at which
>you could then look at more in detail in order to decide if to delete
>them or not.



--
-------------------------------------------"Never tell me the odds!"--------
Ray Olszewski					-- Han Solo
Palo Alto, California, USA			  ray@comarre.com
-------------------------------------------------------------------------------

-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Newbie or user?

[Bill Pleasants]

With no successful use of Linux yet and after reading this list for
several weeks, an interpretation or view of Linux has formed.  Ray, you
have touched on some issues from a broad perspective so I will say this
here.  Maybe there is a more appropriate forum.  In a course I took
recently, I took on the proposition that all interpretations are valid
and none is the truth.  The truth is all that is and cannot be
comprehended by the mind.  Any interpretation is valid because it allows
us to think about an aspect of the truth. Everything known with the mind
is an interpretation, from nuclear physics to M$ sucks.  I have greater
or less freedom and power depending on what interpretations I choose or
invent.  So what I offer is not the truth and is intended to open
something up, to assist something that is evolving and wanting to come
forth.

I suspect that Heimo and others are doing what I have done.  I have used
my very limited grasp of Linux jargon (language used in a specialty) to
ask help in using Linux and communication is difficult.  I share the
vision expressed in various Linux discussions of Linux providing an
alternative to MS Windows for desk top computers with personal and small
business/organization applications.  This would mean that I would only
need to learn to use a GUI and particular applications via the GUI. 
With a mouse click or two I could see a neatly categorized and
searchable compilation of many available applications, what they do and
screen shots if helpful.  Another click would install one.  At any time,
another click would completely uninstall an application.  Most requests
for help would be answered in the form of "Click this and this and
this."  

Where are we now?  My interpretation of "newbie" is non programmer or
non developer (I don't know the jargon well enough here).  If I am going
to make use of the software freely and generously provided and the
information and advice that you pros freely and patiently give, I have
to learn programming to the level of file structure and nomenclature;
installation procedures and organization; editing, file management and
various other utilities. From my experience with Fortran programming I
suspect this would take weeks or months of eight hour days.  Thus the
cost to individual users is greater than for MS.

What is at the heart of the gap between vision and achievement?  I say
it is a matter of acknowledgment.  We human beings cannot live alone. 
We are alive and fulfilled to the extent that we participate with and
are recognized by others.  "RTFM."  Are we potential users put through
the newbie initiation so that we get a taste of the incredible
complexity you pros have mastered?  So that you know that we have some
appreciation of what you have provided?  After all, if Linux did provide
an alternative to MS Windows as I described, it would all just work. 
Most users would not know you or of your dedication it took to
accomplish that.

A secondary or consequent lack seems to be an extensive array of small,
routine program elements like drivers, install and uninstall sequences,
GUI links and things I know nothing of.  This lack is consequent in this
interpretation because these routine things are least appreciated and
don't even provide the intrinsic satisfaction of a creative
accomplishment.  I expect MS has a large budget to pay programmers to
"grind it out."  

Like John Lennon and I believe many Linux fans, I dream of an economy of
generous giving to supplant our economy of cautious trading.  This will
be more consistent with the love of others and joy of life that we left
in childhood.  And Linux developers have given much.  Until providers of
food, fuel, etc. give some of their time, an intermediate step might be
workable.  What if an opportunity were provided for users to make
donations to developers?  Public TV is a model of this.

Some ideas: Program elements could contain identification of the
developer in a standard format so that payments could be made like with
shareware.  Distributors could collect donations and distribute this
back. This might initially be focused on the routine stuff that is
lacking. Imagine this: every GUI window has a standard button which
links to a www on-line donation page.  A wide open channel for users to
express their heart felt appreciation.  I don't know the legalities of
this.  Maybe Linux would have to be registered as a religion.  Hail
Linus! :)

Sincerely,
Bill

-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: Newbie or user?

[Elias Athanasopoulos]

> comprehended by the mind.  Any interpretation is valid because it allows
> us to think about an aspect of the truth. Everything known with the mind
> is an interpretation, from nuclear physics to M$ sucks.  I have greater

I disagree. In Physics the truth is what you measure.

> Where are we now?  My interpretation of "newbie" is non programmer or
> non developer (I don't know the jargon well enough here).  If I am going
> to make use of the software freely and generously provided and the
> information and advice that you pros freely and patiently give, I have
> to learn programming to the level of file structure and nomenclature;
> installation procedures and organization; editing, file management and
> various other utilities. From my experience with Fortran programming I
> suspect this would take weeks or months of eight hour days.  Thus the
> cost to individual users is greater than for MS.

Newbie is someone that learns something (almost) completely new. If
you were to learn the Superstrings theory, then you would be a newbie
in Strings, but that doesn't mean that you have no clue about Physics.

Pls, don't compare Linux OSes and Unices with Windows. They are 
different things and thinking that one is developed for the substitution
of the other is completely nonsense.

> express their heart felt appreciation.  I don't know the legalities of
> this.  Maybe Linux would have to be registered as a religion.  Hail
> Linus! :)

Linux is a kernel, i.e. a tool.

Ciao, 
 Elias
 
-- 
http://gnewtellium.sourceforge.net			MP3 is not a crime.	
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs