Linux-Advocacy Digest #282, Volume #29 Sun, 24 Sep 00 00:13:13 EDT
Contents:
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Tim Hanson)
Re: How low can they go...? (T. Max Devlin)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Richard)
Re: The Government's Decision to Use Microsoft (lyttlec)
----------------------------------------------------------------------------
From: Tim Hanson <[EMAIL PROTECTED]>
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: Sun, 24 Sep 2000 03:23:41 GMT
2:1 wrote:
>
> Richard <[EMAIL PROTECTED]> wrote:
>
> > I am of the opinion that helping users (even the "incompetent"
> > users) and helping yourself is not an either-or proposition. By
> > helping users, you are helping yourself.
>
> But helping incompetent users does not necessarily help you, if you have
> to simplify the system, you might have to remove some of the power of
> it.
I don't quite agree with that. In the quest to make Linux more friendly
for non-geeks, no one is even hinting at removing any of its power. In
fact, it is moving handily in both directions along with fifty others.
There isn't an "either / or" choice here.
>
> >
> > Users have legitimate /universal/ grievances against the system.
> > The problem is that;
> > 1) users aren't able to verbalize their grievances, and
> > 2) programmers are used to ignoring, defending and even
> > accepting whenever the system /they/ created misbehaves.
> ^^^^^^^^^^
> That's to do with bugs. That is entirely different from the model uses
> for interaction.
>
> > How do you expect to get an objective opinion if you alienate the
> > only people who don't have an emotional investment in the current
> > state of software?
>
> The programmers, especially OSS ones, *are* the users. They have the
> emmotional attachment to the software. Why should they bend over
> backwards to change something that they created and they like,
> considering that noone is paying them to do it.
Many are paid to do so. Many of the software houses have programmers
coding open source because it benefits the business as well as the free
software community. Red Hat, Sun, Netscape (Mozilla), IBM, Helix,
Eazel, and SGI come to mind.
>
> > > The other problem is that it is very difficult to make a universally
> > > liked, foolproof program, since taastes vaary so much.
> >
> > It *isn't* subjective! Even beauty is universal among humans.
> > That's right, there are universal principles of beauty. Christopher
> > Alexander writes about these principles in On The Nature Of Order.
>
> Well, that's just damn wrong. Beauty is subjective. Take the example of
> that group of (black) people, where the women stretch their bottom lips
> until they are very big (I forget the name of those people). I,
> personally find that deeply unattractive, but those peopls seem to
> really like it. If that isn't a difference of opinion anout beauty,
> then what is? Even *birds* display different taste when it comes to
> artistic appreciation.
>
> There may be some things universally regarded as beautiful. Others are
> definitely not. Judging by discussion on this group, OS interfaces are
> not of the universally appreaeciated kind. Compare a known windows
> advocate to a command line junkie (me). That's a real, unresolvable
> difference in taaste. We just like different things.
>
> > > Seriously,
> > > though, if the (esp. free software) programmmers loathed users so
> much,
> > > they wouldn't give the users the programs.
> >
> > Because this is a side-effect of sharing software among fellow
> > programmers?
> > Because this is a way of promoting the education of new
> > programmers? (*)
> > Because they got it into their head that this is what they're
> > Supposed To Do but they still don't understand why?
> > Because deep down they hate users and inflicting horrible
> > software on them is a good way to express that hatred? (**)
>
> As a programmer, I don't share the code just to spiet users. If the
> users don't like the code, then they can use something else, or (group
> together and) pay someone else to make something that thay like. They
> can't expect the world to get better with no financial or labour
> contributions (complaining doesn't actually get the work done).
>
> > *: note how much more tolerant <programmers/academics>
> > are to uneducated and incompetent people who intend to join
> > the ranks of the <programmer/academic> class.
>
> It is usually because the other kind of user is unwilling to even try to
> learn, so the programmer/academic has to keep going over the same points
> over again. That irks some people.
>
> > **: as I pointed out elsewhere; if you are responsible for a
> > person's welfare and mental well-being then merely not
> > giving a damn about what happens to them is tantamount
> > to being deliberately nasty to them.
>
> Programmers are not responsible for the users mental well being at all.
> In fact, the OSS programmmers (esp. GPL ones) are not responsible for
> anything. No fitness for a particular purpose guarntees. Since the
> programmer is coding for himself, and merely allowing the user to take
> what advantage of his work that they can, it is in no way up to him to
> make sure that the user likes it.
>
> If someone gave you a box of hard drives and said 'these are a faulty
> batch, some might work, some might not, take them if you want them', you
> wouldn't complain if some didn't work or gave corrupted data.
>
> > Programs are plain data, and they /cannot/ operate on data; a
> > "non-running executable object" cannot operate on data. Only
> > processes can operate on data and so only processes are operators.
>
> I thought that you were trying to blur or eliminate the difference
> between processes and programs (ie persistence). Besides, a function
> definition is just a bit of data on a page describing how the function
> works. It is still an operator. A sequence of commands operates on data.
> It does not matter if they are resident in memory or not. That is an
> artificial abstraction necessitated by computer architecture and OSs.
>
> > > The problem with good OS design is that there is a balance to be
> struck
> > > between many different goals. Many of them are mutually exclusive.
> >
> > A lot fewer than most people believe. Programmers are just used
> > to rationalizing their failure to meet all the goals by claiming that
> > they are mutually exclusive. They never actually provide proof of
> > that claim and they don't go to a lot of effort to try to achieve all
> > their goals.
>
> Compatibility is very important, so is good design. Trying to make a
> brand knew OS compatible with an older, nastier, but fairly different
> version is bound to involve nasty hacks. Besides, that is a very
> negative point of view of yours. In the case of a commercial OS, money
> is put in one end and a (say) substandard product gets out the door.
> That is a failure. With free software, nothing goes in, something comes
> out, that is only success.
>
> > > 3Dsia also offers things like multiuser worlds. That is not the job
> of
> > > a filesystem. UNIX does provide multi-user interaction programs. It
> > > also does lan/wan visualisation. Again, the local fs doesn't need to
> > > provide that.
> >
> > Sure it does. As long as you understand that "local fs" means "any
> > fundamental component of the OS that provides a filesystem interface"
> > and you know that Plan 9 got away with turning almost all components
> > in the OS into filesystem servers.
>
> Plan 9 looks like a very interesting OS. I will look in to it when I
> have time.
>
> > Of course, they had to kludge it
> since
> > the Unix filesystem semantics are weak, ugly and pathetic.
> How do you know they kludged it in. It's inspired by unix, but it is in
> no way unix, so it has no probelems involving unix FS progbems.
>
> > In a good OS, the user storage component (the filesystem analogue)
> > would provide portals (powerful versions of mounts) to other storage
> > components on networked machines.
>
> What's wrong with mounts. How are these protals more powerful?
>
> > The only thing left for 3Dsia
> > to do would be to visualize what's already there; 3Dsia wouldn't need
> > to collect information from inconsistent interfaces (network, FS,
> etc).
> > Even avatars can be implemented by the shell if you have bidirectional
> > links.
>
> BASH makes symbolic links to directories look bidirectional. It's very
> useful.
>
> > > Yes it is. The UltraSpracs running Solaris in my department can be
> > > suspended. It saves the state, shuts down and powers off. When you
> wake
> > > it up, it even seems to remember where the mouse was: all running
> > > processes are restored in to the state they were left in. Personally
> I
> > > like this feature and wish Linux had it.
> >
> > There you go. I wonder how they do it though.
>
> It's probably not too complicated to save the state of eaach process and
> kernel table, as long as you have enopugh disk space.
>
> > It doesn't. But making a new vi process by executing the program
> > a second time doesn't make sense either (the code segment will
> > be the same in either case). And you have to realize that if you
> > turn vi into a server then it will have to be coded differently (and
> > I don't want to get into authentication of users and security since
> > that's just proof of the weakness of the Unix security scheme).
>
> Unix can authenticate users. Besides, you would either need threading or
> multiple processes. the advantage of multiple processes is that if a bug
> causes the process to overwrite something it shouldn't, the process
> dies. With threads the whole lot can die. Multiple processes are easier
> to write (and likley, therefore more bug free) and more robust. Threads
> use elss resources. I don't understand why one is `wrong' and the other
> is `right'.
>
> > > You can run secure server processed under unix.
> >
> > Not easily, let alone *trivially*; in Unix you have to reimplement
> > authentication for each secure process. Why not just have security
> > for /all/ processes? (and access control lists don't give you
> security)
>
> Why can't you just use the authentication library calls avaliable in
> linux?
>
> > > What about those who like the status quo. Why not maintain it? If I
> like
> > > the way a system works, why should I go out of my way to make it
> harder
> > > for me, but easier for someone else? That doesn't make sense to me.
> >
> > If you like the status quo then you have a twisted definition of what
> > "harder" means. The status quo is not easy for *anyone*. Those who
> > don't believe this just haven't been exposed to good alternatives.
>
> I don't necessarily care about anyone. If the status quo is easy for
> *me* to use, then why should I make it harder for *me* to use. I know
> it's not easy for anyone, I didn't say that. I find the status quo
> fine---I do not find that I am fighting against the OS to get my work
> done. I am willing to try alternatives, though when I have time.
>
> > > If the programmers get hurt, there will be no programs for users to
> use.
> > > I also don't think that there is a class war, noone I know tries to
> > > scare away users,
> >
> > No one has to /try/ since it's a fait accompli. Would you agree with
> the
> > sentiment if I phrased it "no one is too heartbroken when users run
> > away screaming into the night?" Maybe it's a little harder to
> understand
> > why programmers /should/ care about the users that run away screaming.
> > Possibly because programmers are so used to rejection that they've
> > grown insensitive to it, which causes them to write bad software which
> > causes users to reject their software, and so on in a vicious cycle.
>
> That's one way of looking at it. So are you saying that all the OSS
> programmers who wrote stuff for themselves to use should never have
> allowed anyone else to use it, just in case someone couldn't use it
> (nevermind those who could). The coders write for themselves, and their
> aquantainces. They don't just churn out code to spite people.
>
> > Note also that I drew a distinction between software developers and
> > programmers. I'm not concerned with the distinction between people
> > who write and don't write software, that's just technical. I am
> concerned
> > with the distinction between people who accept the limitations of
> > software and those who can't tolerate them.
>
> All softwaer will have limitations. If you can't tolerate them, then
> computing is the wrong field for you.
>
> > > I don't agree with the exclude part. Languages are only exclusive if
> > > people are not willing to learn. I can not expect to be able to
> > > communicate with a Chinaman, because I am not willing to learn
> Chinese
> > > (I don't have time amongst other things). I should not expect him to
> > > learn English just so I can communicate. Likewise if the users of a
> > > computer system aren't prepared to learn the language, then
> expecting
> > > other people to change their language for the user is a little
> arrogant.
> >
> > Ahhh, but you're only looking at the result of language differences,
> > not at the causes of language differentiation. Languages differentiate
> > for many legitimate reasons but one of the illegitimate reasons is to
> > support cliques and classes, to exclude people. The Jargon file
> mentions
> > people who use 'grep' instead of 'search' for looking through their
> bags!
>
> Only a serious lapse would cause someone to say `grep' to someone who
> wasn't a fellow hacker. I also don't see why programmers should be
> forces to use an inefficient language so that people who have no
> interest in understanding can understand.
>
> > > Programmers talk the language that is natural to them. They should
> be
> > > permitted to do that, just like you would let a chinaman talk
> Chinese.
> >
> > C++ and Unix are not natural. They are artificial constructs. And
> judging
> > by the number of people who find them difficult, burdensome and are
> > generally revolted by them, they certainly aren't a good fit for the
> human
> > mind.
>
> There are plenty of people who do like them and do understand them. Why
> shouldn't the people who find them a powerful communication medium be
> allowed to use them? Hell, me and most of my aquantainces at school
> couldn't get the hang of French or German, bu tnoone proposed that the
> language should be abolished because it suopported a minority clique of
> `superior' French speakers.
>
> > Sure they would. English makes no sense at all and hundreds of
> > millions of people use it. English is an amalgam of other languages
> > and that's why it's so inconsistent, arbitrary and inelegant when
> > compared to Japanese, French, German, Italian ....
>
> If you think English is inelegent, then you could do with reading more.
> Many books, plays, poems are written very elegently. In fact, the
> extermely broad vocabulary and range of things allows you to do many
> things elegently.
>
> > No one language fits all. And I agree completely.
> But you seem to think that the language that fits programmers and the
> like should be abolished for the sake of the users, because it doesn't
> fit them.
>
> > Programmers are stuck in a rut and they're unwilling to make
> > the effort of getting out of it. And then they (#) hypocritically
> > bash users for not making the effort to join the dark side of the
> > force. Programmers complain about "Back in my days ..." and
> > make users want to scream "Choke and die old timer."
>
> There are plenty of braod minded programmers out there. One thing
> programmers object to is incessent whining by users that everything
> should be `better' with out any help. The users od seem to complain
> endlessly about something that is free and that they choose to aquire.
>
> > Additionally, I despise the "good enough solution" effect where
> > a half assed solution to a problem will come to predominate. The
> > path of least resistance is not a legitimate reason to entrench the
> > status quo. Any effort to make a half-assed solution /better/ (eg,
> > improving Linux in any way) is anti-social and destructive in the
> > extreme (##). Unix's time has come and gone.
>
> Improving linux is in no way anti social or destructive. it hurts noone.
> You can't know that all these people who work on linux would be
> implementing your one-true-way system if they weren't working on linux.
> How on earth does contributing to linux hurt anyone? It in no way
> hinders the development of a persistent OS. The people who want to write
> that aer still free to do so if others choose to work on linux.
>
> > I guess it's deep seated resentment. Here I am taking the moral
> > and ethical high ground by choosing Smalltalk and working on
> > a revolutionary OS and other people work hard at deliberately
> > undermining me (or even worse, they *casually* undermine me).
> > How would you feel? The worse part of the whole thing is that
> > I understand why they're doing it, so I can't actually feel hatred
> > for them.
>
> Who is undermining you. How are they doing this? And other people
> working on a big project that they like and enjoy (linux) does not count
> other people as undermining you, unless you are very paranoid.
>
> -Ed
>
> --
> Konrad Zuse should recognised. He built the first | Edward Rosten
> binary digital computer (Z1, with floating point) the | Engineer
> first general purpose computer (the Z3) and the first | u98ejr@
> commercial one (Z4). He got there before Von Neumann too| eng.ox.ac.uk
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
--
"I have just read your lousy review buried in the back pages. You
sound like a frustrated old man who never made a success, an
eight-ulcer man on a four-ulcer job, and all four ulcers working. I
have never met you, but if I do you'll need a new nose and plenty of
beefsteak and perhaps a supporter below. Westbrook Pegler, a
guttersnipe, is a gentleman compared to you. You can take that as more
of an insult than as a reflection on your ancestry."
-- President Harry S Truman
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Sat, 23 Sep 2000 23:25:39 -0400
Reply-To: [EMAIL PROTECTED]
Said [EMAIL PROTECTED] () in comp.os.linux.advocacy;
>On Fri, 22 Sep 2000 18:50:51 GMT, The Ghost In The Machine
><[EMAIL PROTECTED]> wrote:
>>In comp.os.linux.advocacy, T. Max Devlin
>><[EMAIL PROTECTED]>
>> wrote
>>on Sun, 17 Sep 2000 01:52:44 -0400
>><[EMAIL PROTECTED]>:
>>>Said The Ghost In The Machine in comp.os.linux.advocacy;
>>> [...]
>>>>But you are correct; there are a number of obstacles in the establishment
>>>>of a viable operating system -- the one who gets there first usually
>>>>gets most of the business. Consider that DOS is a variant of CP/M,
>>>>which might be considered the first OS available on a modern PC
>>>>(specifically, the Osborne), as opposed to a console unit such
>>>>as, say, an IBM 4341 or VAX 11/780, which weren't all that portable. :-)
>>>
>>>So the one that gets there first....?
>>
>>... gets a good chunk of the business. This is not necessarily
>>the best for society, of course -- consider, for example, the
>>68000 microprocessor, a better design, at least from a programmer's
>>standpoint, than the 8086 and 8088. Yet the 8088 was the one adopted
>>by IBM. Why? Because it was first.
>
> No, it was available then. I might have been cheaper however.
>
> There's also this nonsense idea I occasionally hear about the
> 8088 being cheaper to use due to 8-bit perhipheral chips
> (nonsense because the 68K's are quite capable of using 8-bit
> peripheral chips).
I think that was a comment on the overall design, that 8 bit peripheral
chips, like the 8086, were cheaper.
>>And now we're more or less stuck with it. :-/
>
> This is more a side effect of bad MS engineering than
> anything else. Nearly every other commonly known OS
> has not suffered this problem (VMS, Unix, NextStep, MacOS).
MS doesn't engineer the CPU. But, then, the current chips aren't
anywhere near as limited as MS engineering, I guess you mean.
--
T. Max Devlin
*** The best way to convince another is
to state your case moderately and
accurately. - Benjamin Franklin ***
======USENET VIRUS=======COPY THE URL BELOW TO YOUR SIG==============
Sign the petition and keep Deja's archive alive!
http://www2.PetitionOnline.com/dejanews/petition.html
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: Richard <[EMAIL PROTECTED]>
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: Sun, 24 Sep 2000 03:28:58 GMT
Donovan Rebbechi wrote:
> Wow ! sYears, huh ? And what have you produced ? After years, the best
> you can do is troll advocacy groups ? What a dismal failure of a project.
> After all these years, all you've produced is some cheap class warfare
> rhetoric.
You judge without any understanding of the situation, the problem domain,
or even any kind of perspective. What the hell do you know about the life
cycles of research OSes? In case you didn't know, Linux is *NOT* any kind
of useful comparison since Torvalds had hordes to help him on the kernel
even before they got to the point of exploiting all of the FSF's previous
work, and they never had to do any kind of high level design work either!
When I proposed modifications to the shell, you said you didn't have the
technical skill to implement them. How long do you think it would take
you to get that skill? How long do you think it would take you to get the
skills necessary to build a complete operating system *from scratch*?
It can take years for someone to understand the Linux kernel, if they
even manage it. And you want me to **what**??
So what the hell are you condemning me of exactly? For thinking on a
scale of years and decades? Since when is short-sightedness a virtue?
Oh, that's right, we're talking about North America! Well then, since
long-term thinking is a sin now, the least I'll demand is that he who
casts the first stone be free from such sin ... and what was it you
said about your PhD?
Or wait a minute, maybe you're complaining that I don't spend every
minute of every day on my OS project. I never got the impression
that you favoured anti-social geeks before. Or is it that you want
to be able to scrutinize and pass judgement on my design? Not for
free! I want some kind of commitment before I go to the effort of
writing things up for someone's benefit! And after the amount of
thought, a flat zero, you put on the trivial subject of timestamps,
it will be a cold day in hell ....
So what is it that you are attacking and defending here? You do give
hints, so I'd say that you're defending your naive view of OS work;
the good guys wear white hats (Linux and other free software) and
there is no politics, only technical problems. Politics is another
thing you seem to be blissfully ignorant about (although how you
could be an academic and ignorant of politics is beyond me).
------------------------------
From: lyttlec <[EMAIL PROTECTED]>
Subject: Re: The Government's Decision to Use Microsoft
Date: Sun, 24 Sep 2000 03:39:56 GMT
lyttlec wrote:
>
> The Ghost In The Machine wrote:
> >
> > In comp.os.linux.advocacy, mark
> > <[EMAIL PROTECTED]>
> > wrote
> > on Fri, 22 Sep 2000 23:23:04 +0100
> > <[EMAIL PROTECTED]>:
> > >In article <[EMAIL PROTECTED]>,
> > >The Ghost In The Machine wrote:
> > >>In comp.os.linux.advocacy, Erik Funkenbusch
> > >><[EMAIL PROTECTED]>
> > >> wrote
> > >>on Fri, 8 Sep 2000 02:06:41 -0500
> > >><Z_%t5.136$[EMAIL PROTECTED]>:
> > >>>"Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
> > >>>news:[EMAIL PROTECTED]...
> > >>>> Wrong. In Linux, you can write a signal handler for any signal (other
> > >>>> than 9--SIGKILL), ***INCLUDING*** mathematic exceptions (which is what
> > >>>> is produced by a div_by_0 error.
> > >>>
> > >>>And with NT you can provide a Structured Exception Handler to handle any
> > >>>fault except NMI.
> > >>
> > >>Exactly. The failure lay in a poorly programmed application that failed,
> > >>not in the OS. The OS stayed up, but even the most rabid Linux Loony
> > > ^^^^^^^^^^^^^^^^
> > >My understanding was that it did not. That is an NT problem, since
> > >Linux does stay up in this situation, and doesn't require the programmer
> > >to write additional exeption handlers. This seems to be another hangover
> > >from the DOS days.
> >
> > I think you might has misunderstood me; my understanding is that
> > the app died, but the OS stayed up. (However, without the app,
> > the computer unit wouldn't be able to control its part of the ship.)
> >
> > I fail to see how Linux would be more intelligent in that situation,
> > although by default Linux doesn't kill things that divide by 0 (instead,
> > one gets a NaN, which is a fairly funny number), so maybe that was
> > the problem in a nutshell.
> >
> > I'll have to try dividing by 0 on both operating systems the next
> > chance I get. :-)
> >
> > [.sigsnip]
> >
> > --
> > [EMAIL PROTECTED] -- insert random NaN here
> I just did a program with lots of divide by zero for both Windows95 and
> Linux. It just printed out the values of 1/sin(x) where x would pass
> through 0. The first time under Linux, the application died but not the
> OS. A change got it to print NaN and continue. Under Windows95 (using MS
> VisualC++ 5.0) neither died, but I got a garbage number.
In the process of doing the above I wrote the following program :
// hello.cpp
// prints hello world
#include <iostream.h>
#include <string>
#include <math.h>
using namespace std;
inline void pr_message(string s = "Hello Russ!")
{cout << s << endl; }
int main()
{
pr_message();
for (int i = 0; i < 3; i++){
cout << 1/sin(i)<< endl;
}
}
It compiles and runs with g++, and Borland C++ Builder 4.52 and others,
but under VC++ 5.0 I get lots of errors saying such things as "std is
not a namespace" and that "<<" does not have a right hand operator of
type string.
Anyway, when I get all versions to compile and run :
g++ gives inf as the result of divide by zero ( it should if I read the
standard correctly)
Borland gives a "divide by zero" message and dies. The application dies,
but not the OS
VC++ prints 1.#INF.
What gives? I don't expect the Borland to be correct, as it is too old.
But does the gnu project and I read the standard wrong or is it MS?
------------------------------
** 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
******************************
