Date: Fri, 29 Aug 1997 18:12:00 +0930 From: Greg Lehey <grog@lemis.com> To: Peter Korsten <peter@grendel.IAEhv.nl> Cc: freebsd-chat@FreeBSD.ORG Subject: Re: ATT Unix for Windows ! Message-ID: <19970829181200.30452@lemis.com> In-Reply-To: <19970829025323.52918@grendel.IAEhv.nl>; from Peter Korsten on Fri, Aug 29, 1997 at 02:53:23AM %2B0200 References: <19970825204932.12036@grendel.IAEhv.nl> <34020362.7DB1@fps.biblos.unal.edu.co> <19970825224258.55928@grendel.IAEhv.nl> <19970826083051.FR52594@uriah.heep.sax.de> <19970826235525.22143@grendel.IAEhv.nl> <19970827093336.NX00626@uriah.heep.sax.de> <19970828002532.43939@grendel.IAEhv.nl> <19970828083703.OY21311@uriah.heep.sax.de> <19970828174348.17499@lemis.com> <19970829025323.52918@grendel.IAEhv.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Aug 29, 1997 at 02:53:23AM +0200, Peter Korsten wrote:
> Greg Lehey shared with us:
>> On Thu, Aug 28, 1997 at 08:37:03AM +0200, J Wunsch wrote:
>>> As Peter Korsten wrote:
>>>
>>>> Er, that's actually what I did. :) I put the objects files in a
>>>> '.for' loop. I wanted something like '*.c' but it wouldn't work.
>>>
>>> SRCS!= echo *.c
>>>
>>> (But that's BSD make, alas.)
>>
>> There are similar constructs available for other makes.
>>
>>>> It's not really more complicated, it's more work. I have to make
>>>> a seperate Makefile (with the chance of errors) and edit that
>>>> when I add a file to my project.
>>>
>>> I consider auto-adding each new file fairly dangerous. I often drop
>>> files like `foo.c' in my workspace where i have been testing
>>> something.
>>
>> I would guess that Peter hasn't understood the power of make.
>
> Yes, Peter does understand the power of make. As with almost all
> Unix tools, you can do nearly anything you want with it.
I'm sorry. It looks as if you have taken personal offense to my
statements. That wasn't my intention.
No, the power of make is not that you can do almost anything you want
with it. Few will write such monstrosities as Roland McGrath's
original set of concatenated Makefiles for GNU libc. What I'm talking
about is the many-to-one mapping so typical of the make process.
> You don't auto-add each file you open. You may open a file, edit
> some, and insert it into your program. Or you insert a new file
> into your project.
Are you talking about your Microsoft IDE? Why do you need to do that?
> It's a bit easier than editing your Makefile.
Why edit your Makefile? Jörg shows how to eliminate that step.
> And speaking of versatility, you can also build your own application
> wizard.
What's an application wizard? Some kind of superguru?
> Just build the framework for your code with a couple of mouse
> clicks.
Why do I have to take my hands off the keyboard?
> Sure you can do this with make. But that just as well take time.
So far, you haven't shown anything which would require a modification
to a Makefile.
>>>> Why did you write your own keybaord mapping? Wasn't there a
>>>> suitable mapping available?
>>>
>>> There wasn't.
>>>
>>>> There certainly is a German keyboard
>>>> mapping for Windows.
>>>
>>> Sure. The German keyboard has been designed by a typist, not by a
>>> hacker. How else could they have laid out the {[]}'s in a way where
>>> you break your fingers? The consequence is that most hackers simply
>>> avoid German keyboards at all, and use US-ASCII ones. But they fail
>>> to write texts with German umlauts on them. My mapping allows for
>>> both.
>
> Well, that figures. Non-US keyboard layouts usually bother the ****
> out of me. Dutch keyboards seem to have existed, but they died very
> quickly.
Interesting. Do they use standard ASCII? I didn't know that.
>> Round about here we begin to see the divergence of your views. Peter
>> wants a standard, no matter how bad. Jörg wants a tool that he can
>> use. No trouble guessing which side I'm on :-)
>
> You misunderstand me, quite some. I don't want a standard at all
> costs. I just see that configuring Windows takes less time than
> configuring Unix.
I've never seen evidence for that. Every time I mess with Windows, I
end up cursing and swearing. But then, I know UNIX, and I have great
difficulty with the way Windows forces me to work.
> And, yes, Unix is more versatile. But do you also add extra functions
> to your television?
Yes, as it happens.
> Aren't you bored with those standard boxes everybody else has? Or do
> you customize your car?
I used to.
> Some people build their own stereos, some customize their car,
> some need to be able to totally customize their computer. But
> they're a minority.
Does that make us bad? Some people think. Some people come up with
brilliant ideas. But they're a minority.
>>>> Talking about desktops: I have a very personal desktop with NT,
>>>> that looks totally different from what everybody else uses at the
>>>> office. (For the insiders: color scheme Rainy Day, background Blue
>>>> Monday, automatically hiding taskbar on top, small icons, and
>>>> shortcuts to all drives on my desktop, together with Netscape and
>>>> the mandatory icons.)
>>
>> Wow. How many components of this desktop originated outside NT?
>
> None. Why?
Jörg's point was that you can choose from anything you like, you're
not bound to what one manufacturer thinks you should use. I was
wondering if you had any evidence of such flexibility under Microsoft.
>>>> Isn't it a bit strange that
>>>> the configuration of most X-applications is done in a text file?
>>>
>>> What else? Some piece of binary junk that can only be maintained by
>>> that very program itself? Store the layout in a bitmap?
>>> No, the only problem with this is that the authors of most X11
>>> software didn't think of adding a knob to allow you editing them
>>> without using a text editor.
>
> Correct.
Not quite. They thought of adding a knob. They just didn't add it.
The easiest way to add a knob is when you know what you want it to
look like. If you change your mind, you change the knob. You can't
do that with systems which hide the representation.
>> Is that a problem? They supplied a format which *can* easily be used
>> by other applications. Megaslop, by comparison, stores the
>> information in proprietary format binary files. Theoretically you
>> could save a few bytes by doing this (that was the argument I always
>> used for doing it this way, anyway :-), but in practice I'd guess that
>> Megaslop config files are larger even than the biggest X app-defaults
>> files.
>
> (It took me quite some time to find out that you actually mean
> 'Microsoft' when you write 'Megaslop'. IMO, this looks childish.)
You're welcome to your opinion.
> Yes, it's a problem. That's what I meant. I don't care about the
> way the data is stored. I've designed a user interface with custom-
> izable settings for the Amiga once and it also stored it's settings
> in a text file.
>
> What I do care about, is indeed the fact that Joerg mentioned: the
> inability to edit these settings with a preference editor.
Why can't you do that?
> If you have a graphical application, it's _stupid_ that you have to
> edit a text file to alter something in the appearance of that
> application.
So fix it. I don't personally have a problem with it. If somebody
built one of those stupid Microsoft card file config programs for X, I
would certainly not use it. I don't have time to mess around with
that kind of stuff.
> I don't know if you people have ever read a book about how a user
> interface should be constructed (I have),
So have I. Which one of the 20 conflicting books did you read?
> but you could learn a lot out of it.
What?
Once upon a time, before the advent of Microsoft Windows, there was a
kind of UI standard with which hitting the Enter key would take you to
the next field in a form. When you got to the end, it wouldn't wrap
around, it would submit the form. How do I do that with *any* GUI?
Of course, since we're talking GUIs, you still should have the option
of selecting an individual field with a mouse.
Most GUI standards place undue emphasis on the mouse. That's fine for
people who can't type, but for those who can, it's a pain (in more
than one sense. One day, a group of CTS sufferers will sue Microsoft
for damage caused by the completely avoidable double click, and will
win millions).
>> The disadvantages of the binary files are obvious when you think about
>> it: you need special programs to access them. I've never seen any
>> such file which can store comments. The typical Megaslop config
>> program shows you lots of "card file" menus, which effectively makes
>> it impossible to keep an overview. Never mind that computers have
>> replaced card files and the like--thanks to Megaslop, we still have
>> their restrictions.
>
> Like I said, I don't care about the file format. It's totally
> irrelevant. If your preference editor is good enough, you don't
> need to know about the file format.
If. I've never seen one which even comes close.
> Further, I fail to see what the logical grouping of settings with
> these "card files" has to do with losing the overview. There isn't
> an overview. These are distinct things. If you fill in your gateway,
> you don't need to be able to see what your DNS is.
My sympathies.
>> [Netscape configuration files]
>>
>> On the other hand, over half is comments, and you can use it to
>> configure Netscrape without any other documentation. Show me how to
>> do that with a Megaslop config file, even the text ones.
>
> Oh yes, mummy. I really really want to edit the file myself. No I
> don't want to use the program itself for it.
You're learning.
> You're reasoning from the Unix point of view. You do everything
> from a text file, so you want to be able to do that in Windows too.
> Well, you can't and you shouldn't be able to do so, because the
> average user (who is kept more in mind with Windows than with
> Unix) will certainly screw it up.
In May, I was teaching a class where all the "terminals" were PCs
running Windows 3.x and some broken terminal emulator. I spent half a
day trying to get one (and only one) of the things to connect to the
host. At the end I searched the system, found the config file (which
fortunately was in text format), and discovered that the IP address of
the host had been entered with a trailing ]. It seems that I should
have entered the IP address as [166.111.64.122], not 166.111.64.122.
The configuration editor saw that I had left off the trailing ], and
inserted it. No error message. No way to recognize the problem from
the config screen (which didn't display the trailing ]). I suppose I
should have reinstalled Windows. That seems to be the way to solve
this kind of problem.
You were saying?
>>>> X is not a real graphical user interface as Windows is. Many
>>>
>>> Define `real'. X11 is a windows environment, nothing else. The
>>> toolkits have to be provided separately.
>
> With 'real' I mean integrated. X is very modular and can do every-
> thing the GUI from Windows can too, with proper programming. You
> can choose your own window manager, alter the whole look-and-feel,
> change the way the mouse behaves.
>
> But what it lacks is a common base that describes how programs
> should work (and the professional Windows programs
Hah! Oxymoron.
> do work the same way), a file manager/Explorer like thing to move
> your files around, drag-and-drop, the fact that every decently
> written program behaves in the way you expect it to. Mac, Amiga,
> Atari, Archimedes, and NeXT all have this. X and it's windows
> managers don't.
If that's what you mean by a "decently written program", I'm glad that
X doesn't.
>>>> applications are tty-oriented applications with extra X-support.
>>
>> Some are. Others aren't. Don't forget that character-oriented
>> applications offer you a flexibility that no GUI can possibly offer.
>> One of the biggest mistakes in the Megaslop environment is that this
>> flexibility has been discarded because it's more difficult to learn.
>
> It isn't a mistake. This flexibility is a direct descendant of the
> limits that a text-only environment imposes on you. If you wouldn't
> have the flexibility, you would have nothing, compared to a GUI.
>
> The idea behind a GUI is to make it simple and consistent. Why
> should one need extensive training to be able to operate a computer?
> Why can't many people still program their VCR? Because the user
> interface sucks.
>
> Sure, GUI's are hard to program and not as much fun as writing
> a high-performance file system or device driver. But it's the
> interface to the user and that's why they're so important.
I don't know whether you've noticed--it looks as if you haven't--but a
lot of Microsoft style GUI stuff is available for UNIX. Why don't
people use it? Why don't I use it? Because it's simple. Sure, it's
easy to learn, but it's too restrictive.
>> Yup. There's no good way to represent errors graphically.
>
> Nonsense.
So why don't they do it?
>>> I've seen many win users re-installing their systems quite a number
>>> of times. I couldn't imagine why i should re-install one of my
>>> systems. The machine at home even has migrated a number of disks
>>> already, without re-installation.
>>
>> This is a habit that a lot of people get in to, even with FreeBSD. I
>> don't know how to explain to people that it's a bad idea.
>
> It isn't too difficult to totally mess up you Windows system.
Why not? I thought these graphical config things were designed to
stop just that.
> So it might be the best idea to start from a know situation: the one
> just after installing.
That's a declaration of bankruptcy for the whole concept.
> But say I accidently removed /etc and then rebooted, what would be
> the best way to fix it? Re-installing could be an option.
It would be a *lousy* option. I can't think of a worse one. Here are
the correct answers:
1. Read in the backup
2. If you've been misbehaving and haven't made *any* backups, read it
in from the CD-ROM.
How can you possibly imagine that reinstallation can be of any help at
all under these circumstances?
> If I may conclude my points, all I can say that both systems have
> their pro's and con's. I wouldn't be on this list if I were just
> another Microsoft devotee.
>
> But you can nitpick as much as you want on MS, Windows and NT,
> but some things just work better with them. Security in NT has
> a better base (VMS) that in Unix, despite the security bugs that
> reguraly show up.
I don't know enough about this to offer my own views, but it's in
complete contradiction to everything I've heard so far.
> Another thing, not to be forgotten, is that very many people
> use Windows 95 at the moment. It's pretty easy to couple it
> with NT. So, for an intranet, NT is probably more suited than
> Unix is.
I wasn't just slamming NT. In fact, I think that NT is the second
best operating system that Microsoft ever marketed. It's just that
the rest is all even worse.
> NT is something like five to ten years old. I don't expect all
> the bugs to be out of it, also considering MS's strategy to
> release first and fix bugs later.
or never.
> Unix had 25 years to mature. MS has 20,000 programmers and a lot of
> cash.
Linux is 5 years old. It has probably also 20,000 programmers and no
cash. It seems a whole lot more stable than NT.
> The stepover from 95 to NT 4.0 is pretty smooth. So expect to see
> a lot of NT or it derivates in five years. The monopoly is pretty
> scaring, but what can you do abot it? Unix isn't friendly enough
> to appeal to a mass market. It also lacks options that have become
> pretty standard by now. Noone, except some excep- tions, wants to
> use a text interface anymore. What I noticed as the helpdesk at an
> Internet Service Provider, is that the average customer becomes
> less and less knowledgeable about computers, and this is a trend
> that's continuing.
Right. And Microsoft is actively helping.
Greg
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970829181200.30452>