Linux-Advocacy Digest #350, Volume #28 Fri, 11 Aug 00 09:13:06 EDT
Contents:
Re: Richard Stallman's Politics (was: Linux is awesome! ) long? (Isaac)
Re: Honest question about NT vs. Unix as Internet platform (mlw)
Re: Linux = Yet Another Unix (John Sanders)
Re: Anonymous Wintrolls and Authentic Linvocates - Re: R.E. (Roberto Alsina)
Re: C# is a copy of java (Donal K. Fellows)
Re: Anonymous Wintrolls and Authentic Linvocates - Re: R.E. Ballard says
Linux growth stagnating (Nathaniel Jay Lee)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Isaac)
Crossposted-To: gnu.misc.discuss,comp.software.licensing
Subject: Re: Richard Stallman's Politics (was: Linux is awesome! ) long?
Date: Fri, 11 Aug 2000 11:58:31 GMT
On 10 Aug 2000 23:00:30 -0700, Pat McCann <[EMAIL PROTECTED]> wrote:
>
>All parts of the work, of course. Like the parts that you are "deriving"
>from (eg, a library) and the parts that you are adding (the only thing
>that you get copyright on, BTW). The work as a whole does not receive
>copyright until the work as a whole is fixed in a tangible medium,
>necessarily from preexisting parts. If they don't preexist, you have
>nothing to fix.
>
So if a work references a library and the library is yet to be
written then really the work is not a work? I don't believe this
to be correct. I don't see anything in the statute that suggests
that an author is not entitled to copyright protection of an incomplete
program.
People who write large programs don't type them in all in one sitting.
Are you suggesting that it would be okay to make copies and distribute
their works if you got access to them before the works were finished?
Or viewed another way. The overwhelming majority of C programs call
functions in a run time library for printing, displaying, etc. I've
never heard anyone complain that the source code for such a program
was not a work because it didn't contain the code for libc.
>
>I probably omit too many details, like explaining "as I read the USC"
>with a quote from the USC-17-101: "A 'derivative work' is a work based
>upon one or more preexisting works...". Sorry if my composition skills
>are not swift enough to craft crystal-clear, bullet-proof arguments. It
>would take me a week per post. I think I made the point that all of a
>work's parts (divide it any way you please) must be preexistent before
>it can be fixed in a tangible medium, and if any these parts is a work
>in itself then it meets the definition from USC just quoted.
I had thought that was what you meant, but I was really hoping I was wrong.
I think this is an incorrect statement of the law. Any part can be fixed
in a tangible medium. That part can then be altered and if it ends up
in fixed form, it is again a work. I'm sure I can copyright individual
functions and modules of C programs which don't have a main function and
which cannot run. Those things are called libraries.
>
>I guess the point here what that we needn't get hung up about the
>meaning of "preexisting" in the definition of "derivative". If the
>use of an partial application with a library or a plugin is to be
>considered as a single whole work (not something I would do), both
>parts preexist at time of fixing the whole work and we need not say
>either component is a derivative of the other as I recall people
>worrying about. The whole WOULD BE a derivative of the parts,
>obviously (if they indeed form a whole which I reject).
>
I think I follow you here, but there are additional problems to
consider. As a concrete example let's consider a plug-in and
an application. I think you suggest that when the user gets around
to using the plug-in, the plug-in with application (under someone's
theory) would be a derivative work.
First, that does not appear to meet the statutory definition. What
you have is two complete works joined together. That looks more
like some kind of compilation than some derivative work. Compare
such a compilation to me taking some code written in C and converting
it in a mechanical fashion to another language. I can't claim any
authorship in the translation because my contribution is not creative.
It is instead a derivative work and not a joint one or a compilation.
Secondly, the necessary copying is done in memory as an essential step
to executing the code in the library. Such copying (and any essential
adapting) is specifically authorized in law whenever you legally own
a copy. You generally do not execute library code without linking
it to a program.
Thirdly, what you suggest leads to some unfortunate and IMO
obviously incorrect consequences. It seems to me that I can write a
plug-in for one program that might happen to work with second
program. Can the author of the second program exercise some right
to control the "derivative work" involving his program by restricting
my right to distribute my plug-in? Can I instead direct him to
prevent his users from creating derivative works using my plug-in?
If the answer is no in either case, can you explain how the law works
to make the answer no.
The FSF's answer is that the "derivative works" thing does not
apply in the two program situation. That does remove the
unfortunate consequences I describe, but I don't see any support
for their answer in the law. If I didn't know about the second
program I can't argue that I "based" my program on it rather than
on the first program.
I think the real way to resolve the problem is to properly define
"derivative works". Wouldn't it be fair to not allow someone
to encumber a 'preexisting work'? I think so, and that removes the
ability for me to write a Photoshop plug-in to rotate images
360 degrees and to then extort money from the makers of Photoshop.
Since my useless plug-in, which contains no portion or translation of
any copyrightable portion of Photoshop or Gimp is thus not a derivative
work I can write my plug-in without the author of either program
complaining.
In my view the word 'preexisting' is no mistake; the word is essential
to fairness.
>
>I assumed y'all were familiar with "joint work" from Section 101. The
>parenthetical point is not worth my time to explain my reasoning nor
>your time to comprehend it. Take the first, general, point. Remember
I wasn't familiar. I did go back and read it. You're right that you
should not have to explain it. I think that there are derivative works
that don't meet the definition of joint works, so I'd really need to see
all of the new definitions and rewrites of the rest of the statute before
I could agree with your proposed dropping of a term.
>
>As Section 101 defines terms, I agree you can't just throw one out. But
>I just think they've got a poor organization of concepts. Maybe it
>should replace "derivative" with "union" which is, like a collection,
>is a compilation, but one in which multiple parts are copyrighted by
>distinct parties. It shouldn't be necessary to write of modifications
>or the time-directed derivation of one thing from another; just that
>multiple things were compiled (possibly with some omissions) keeping
>copyrights separate on the things compiled and on the compilation
>itself. (This obviously requires a whole lot more thought.)
>
What if I take someone's code and fix a bunch of typographic errors and
a couple of one off bugs? There are no divisible parts for me to
claim authorship off, and my contribution probably does not have
sufficient creativity to merit any claim of authorship. I think we
have a derivative work that is neither a collection, compilation,
nor a joint work.
I can't argue that "derivative works" couldn't be removed from the
statute, but if it is to be done without consequence to the law, I'm
not convinced that the result will be easier to understand.
>
>I've seen that criticism used on people before, but I don't understand
>it. Deconstruct and rearrange? If that is something I do wrong by
>habit, I would like to learn what not to do. Please explain.
If we want to answer the question, "Does the law allow X", the most
convincing arguments take the law (statutes, legislative history and
caselaw) verbatim and argue that the law says yea or nay. If you
instead argue for a better or different way to express the law,
you'll need to be convincing about that too before your argument will
seem strong.
If you are instead arguing that the law should allow X, then it
would seem more appropriate to argue for the change in the law
and to then present the consequences of your rearrangement.
In this case I see that the law treats compilations and derivative
works differently. I am thus suspicious about an argument that
merges them and then reaches a conclusion. I agree with your
conclusion, just not your argument.
Isaac
------------------------------
From: mlw <[EMAIL PROTECTED]>
Subject: Re: Honest question about NT vs. Unix as Internet platform
Date: Fri, 11 Aug 2000 08:09:40 -0400
To be completely honest, either UNIX or Windows can probably get you
where you want to go.
A huge portion of any datacenter decision must be comfort level, right?
You are they guy carrying the beeper.
For me, I would never put NT in a data center. Too many weird things
happen. My biggest complaint about Windows and NT is the attitude that
"rebooting" fixes a problem. It, in fact, does no such thing. It merely
clears an error condition created by a serious bug. Rebooting assures
that the bug goes uncorrected.
You can see it the the reaction of people. When confronted which a hang,
crash, or odd behavior, a Windows or NT admin will instinctively reboot
and not think about it again (until it happens again). A UNIX admin will
get upset, and call it a bug the first time and try to figure out what
caused it before rebooting.
Also, the tool to administer a datacenter naturally work on UNIX.
Telnet, SSH, X, Rdist, etc. all are made to network. Remotely
administering a UNIX site is easy. Under NT one must try to configure
all these stupid systems, like PC/Anywhere or Terminal Server which in
their own right consume a lot of server resources, which the equivalents
under UNIX are very light weight. To run Terminal Server on a web server
as a way to remote admin is just plain stupid. Look at the requirements
in CPU and RAM just for terminal server!
I often upgrade or programs on client sites from my remote office. It is
as simple as:
tar czvf client.tgz client/
scp client.tgz [EMAIL PROTECTED]:/tmp
ssh [EMAIL PROTECTED]
cd /tmp
tar xpzvf client.tgz
./install
killall -HUP clientprog
(or use an rdist script for more complex installations)
Under NT it is not that easy. In all truthfulness, with the exception of
catastrophic failure i.e. when the server machine no longer responds to
IP packets, you never need to see it. It need not have a floppy, CDROM,
keyboard, monitor, or mouse. UNIX is made to work like this. NT is not.
You may not think this is important, but at 2:00am and the phone rings,
I would like to try to log in from home and fix it there, before I'd be
forced to truck it into work or the datacenter.
Mr Jama wrote:
>
> Greetings. This is not a troll. I am looking for serious help here.
>
> I am responsible for the technical architecture for the Internet site of a
> relatively major company (billions in income). It is my opinion that NT
> provides the most cost-effective solution for our requirements. First off, I
> think NT is as scalable as Unix. And even if it weren't, out user base is
> relatively small. We're not launching Amazon.com here. We have a small number
> of clients (thousands) that spend lots (millions).
>
> Anyways, I'm being told by our IT outsourcer that I don't know what I'm talking
> about, that Unix is far more secure than NT as an Internet platform, and that
> Unix is far more scalable. He's very adamant about this.
>
> Hey, I'm happy to be wrong. But I want some facts to prove it. He claims that
> Unix has about 2 security problems identified per month, whereas NT averages
> about 12. He also claims that NT takes 2+ weeks to post patches whereas Unix
> vendors typically take 2 days after a problem is identified.
>
> What do you guys think? Is NT really less secure than Unix for the Internet?
> What about browsers? Is this guy just a Unix weenie, or does he have a point?
>
> P.S> Unix platform in particular that we use is Solaris.
>
> Thanks,
> feel free to email responses.
> Jama
--
Mohawk Software
Windows 9x, Windows NT, UNIX, Linux. Applications, drivers, support.
Visit http://www.mohawksoft.com
I'm glad we disagree, it gives us a fantastic opportunity to be totally
honest.
------------------------------
From: John Sanders <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.alpha
Subject: Re: Linux = Yet Another Unix
Date: Fri, 11 Aug 2000 06:57:23 -0500
[EMAIL PROTECTED] wrote:
>
> Christopher Browne <[EMAIL PROTECTED]> wrote in message
> news:JQIk5.230497$[EMAIL PROTECTED]...
> > Centuries ago, Nostradamus foresaw a time when John Sanders would say:
> > >[EMAIL PROTECTED] wrote:
> > >>
> > >> John Sanders <[EMAIL PROTECTED]> wrote in message
> > >> news:[EMAIL PROTECTED]...
> > >>
>
> > By the way, the contentions of Xenix availability for Z-80 do not
> > agree with my recollections; the machines with Z-80s on which Xenix
> > got heavily deployed were the TRS-80 Model 16's and Tandy 6000's,
> > which used the Z-80 to control the "terminal," but then used a 68000
> > to run Xenix.
>
> TRS-80 Model II
The good old days. I worked on this machine in Tandy Towers in Ft.
Worth, Tx. I was hoping like hell that Moto would hurry up an release
the 68000 before Intel took another step to the 8088. But it didn't
happen. I left Tandy about this time. The plans were to add the 68000,
but the Model II was released with only the Z80. I think when the 68000
was finally designed in, it was boosted to the Model 16.
--
John W. Sanders
===============
"there" in or at a place.
"their" of or relating to them.
"they're" contraction of 'they are'.
------------------------------
From: Roberto Alsina <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Anonymous Wintrolls and Authentic Linvocates - Re: R.E.
Date: Fri, 11 Aug 2000 09:55:39 -0300
Nathaniel Jay Lee escribió:
>
> Roberto Alsina wrote:
> > If someone tried to do serious UI criticism, he could have said
> > "Use a vertical taskbar, buttons are easier to read"
> >
> > or
> >
> > "Use this gizmo, it's easier".
> >
> > or
> >
> > "Use a corner-hidden vert taskbar, it uses less space abd is as easy to
> > access".
> >
> > But no, the drivel is "the taskbar? Windows has a taskbar! Taskbar
> > eeeevil".
> >
> > That's lame.
> >
> > --
> > Roberto Alsina (KDE developer, MFCH)
>
> Mind pointing out where I said any of that?
Didn't say you did. It's just unfortunate that I hear it a lot ;-)
> I didn't say we can't adopt
> the good ideas that Windows has (hell, I use KDE, so I have no
> complaints on *interface inhancements*). The problem is the idea that
> Linux has to *be Windows*. If you don't know what I'm talking about,
> look at Corel. They do plug and play just like Windows. They reset
> configs without warning. They screw things up without warning. They
> change drivers every reboot (regardless of new hardware). It is one of
> the most Windows like Linux distros yet, and I heard a lot of the press
> 'praise' them for being so much 'like Windows'.
Well, it's their opinion. Recognizing hardware at boot is a good idea,
if it works. Kudzu does work, AFAICS. No idea if Corel is using
kudzu or not.
> I don't mind stealing good ideas from Windows (god knows Windows is
> stolen ideas from all over the place all put together in bad ways). I
> just see absolutely no reason to try to *clone* Windows. And screaming
> for it to be exactly like Windows (as so many seem to be) is just plain
> wrong.
Yup.
--
Roberto Alsina (KDE developer, MFCH)
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Subject: Re: C# is a copy of java
Date: 11 Aug 2000 12:52:41 GMT
In article <8muo59$hhu$[EMAIL PROTECTED]>,
<[EMAIL PROTECTED]> wrote:
> It sounds as though your solution is to use a variation of global
> variables in dynamic memory. That can be easily handled by a
> library of routines that control access to the global dynamic memory
> varaible which in turn call the standard dynamic memory library
> routines. This way only those allocations that need support of a
> memory manager will be suffer the overhead, the remainder will just
> use the standard library routines.
I don't quite understand how to apply that paragraph to my situation.
Sure, I can easily allocate blocks of words, but I fail to see how
they really help. Unless I do it all by implementing a memory manager
for some virtual memory model on top of this. (Which is extremely
nasty, and is really just punting on the whole issue!)
Plus, I need to save these states so that I can backtrack to them
during deadlock analysis. Typical programs can easily require several
thousand of these saved states. That takes a *lot* of memory, so much
so that it is more efficient to share subsections of it between saved
states if possible. But then you're into complex reference counting
schemes (or mark-sweep GC if you have a runtime that supports it.)
Memory, and architectures happy with over 4GB of real memory, are not
so cheap that saving a copy of every state that you need to traverse a
program's state graph depth-first is a practical suggestion.
Life would be easier if the number of virtual threads was constant.
An earlier version of the system that only supported a very small
subset of the current functionality was lean and mean, and could
operate with very little overhead indeed. The code's much more
general now; it can tackle much harder problems, but it is also much
harder to maintain. So much so that the development version isn't in
C any more but Java (OO, GC, bounds checking, and javadoc were the key
features. We've not really got the staff time for me to spend ages
tracking down strange memory corruption problems, and the relative
difference in code speed isn't too serious for our purposes...)
> If you are allocating many small block of memory have you looked
> into using a aggragate dynamic memory allocation algorithms to save
> on memory usage overhead?
I'm already using one. (Or is that several? I forget... :^)
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- Actually, come to think of it, I don't think your opponent, your audience,
or the metropolitan Tokyo area would be in much better shape.
-- Jeff Huo <[EMAIL PROTECTED]>
------------------------------
From: Nathaniel Jay Lee <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Anonymous Wintrolls and Authentic Linvocates - Re: R.E. Ballard
says Linux growth stagnating
Date: Fri, 11 Aug 2000 08:05:15 -0500
Roberto Alsina wrote:
> Nathaniel Jay Lee wrote:
> > I didn't say we can't adopt
> > the good ideas that Windows has (hell, I use KDE, so I have no
> > complaints on *interface inhancements*). The problem is the idea that
> > Linux has to *be Windows*. If you don't know what I'm talking about,
> > look at Corel. They do plug and play just like Windows. They reset
> > configs without warning. They screw things up without warning. They
> > change drivers every reboot (regardless of new hardware). It is one of
> > the most Windows like Linux distros yet, and I heard a lot of the press
> > 'praise' them for being so much 'like Windows'.
>
> Well, it's their opinion. Recognizing hardware at boot is a good idea,
> if it works. Kudzu does work, AFAICS. No idea if Corel is using
> kudzu or not.
If they are it's a hacked up version that doesn't work worth a ****! It
pretty much detects some different peice of hardware every time you boot
(at least it did this to me), and there is absolutely no way to turn it
off.
>
> > I don't mind stealing good ideas from Windows (god knows Windows is
> > stolen ideas from all over the place all put together in bad ways). I
> > just see absolutely no reason to try to *clone* Windows. And screaming
> > for it to be exactly like Windows (as so many seem to be) is just plain
> > wrong.
>
> Yup.
I'm glad you understand. I'm not one of those hard-ass "Oh, that looks
like Windows. It must really suck!" type of people. I don't care of
the common interfaces look like Windows. As long as I still have a
choice (and while I use KDE a lot at work, at home I'm more of an
Enlightenment/Windowmaker kind of guy). And as long as we don't destroy
the underlying system and all it's modularity, I don't care. And that's
why I'm asking for. Making it 'like Windows' would be the equivalent of
solid-welding each and every part together, even when a bolt or screw
would do just as nicely. I like the modularity of Linux (or any *nix
for that matter). You aren't forced to use the system *as is* but can
pretty much modify it any way you want. Default to easy, and let me
modify, that's fine. Default to easy, and tell me I'm an idiot when I
try to modify (like Windows) and there will be hell to pay.
--
[EMAIL PROTECTED]
Nathaniel Jay Lee
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
