| |
In lugnet.off-topic.debate, Todd Lehman writes:
> In lugnet.off-topic.debate, Jeremy H. Sproat writes:
> I don't care if MS is a money-grubbing empire as long as they make great
> products. The problem I have with it is that they're only making good (not
> great) products and that, in combination with their monopolistic nature,
> hurts the other guys out there who are making truly great products.
Yah, no argument there. I prefer the Windows 4 GUI over almost anything
else (1), and I can run a wide range of apps on my NT box. But, I know
the difference between reliability and popularity, and for
mission-critical apps where down time means money or customers lost, I'd
choose Linux over NT any day.
> And now this "The Unstoppable Windows NT" stuff on TV -- ha...we'll see.
Wow -- I haven't seen that one. That's about as arrogant as "Are you
W2K-Compliant?" There's one thing I'll always concede to MS, is that their
marketing sharks know what they're doing.
> > But fortunately, UNIX changed, went
> > open, evolved into something not only extremely useful but also incredibly • fun
> > for perhaps millions of computer hobbyists world-wide.
> If Microsoft someday produces a great operating system, I'll be willing to
> change my mind about them. I doubt they will, but I'll still be open-minded
> if they actually do it.
It won't happen, at least not from Microsoft. I'm betting that
third-party groups will pull it off. And after a landmark legal dispute,
MS will allow these third-party groups to market their more stable and
configurable Windows clones.
Yeah, and maybe the tooth fairy will start taking inflation into
consideration and leave $100 bills under our pillows. :-(
> > Someone's got to do it. It sucks, I know, but survival of the fittest will
> > win out. Hey -- we learned from VMS, we learned from Amiga, we'll learn • from
> > Windows.
> Heh heh. The gross thing is, MS may be the fittest company, but they don't
> produce the fittest products. They sell to the lowest-common denomintor,
> because that's where the money is. Can't really blame them too much for • that,
> I guess.
Well, I wasn't trying to play up the fitness of MS, but rather demonstrate
that this, too, shall pass. The fittest software will win out.
> > > > The rapid growth of the Internet would have never happened.
> > > Doesn't that assume that nothing else would have filled in the demand if
> > > those two players hadn't been there?
> > Nothing would have filled the demand. Without the immensly popular Wintel
> > platform saturating the market, client/server wouldn't have taken off,
> > mainframes would be the big thing, and we'd all be reading this through a
> > dumb terminal right now.
> Actually, without the immensely popular Wintel platform saturating the • market,
> something else better (perhaps from Motorola or Digital) could have moved in.
This would happen if cheap, networked, (semi-)user-friendly desktop
computing is an ecological niche, a vacuum to be filled. I feel that it
is an evolutionary niche -- a rare opportunity taken by a select group --
and if something were to take Wintel's place should Wintel have not come
around, it would have happened much later. Probably late enough in the
game that something completely different would have won out. Who knows
what? Peer-to-peer maybe, or perhaps distributed computing to the
extreme, but probably not client-server as we know it.
Maybe even the Internet of Johnny Mnemonic. X-P
> > The computer industry has always been slow. Look at what Babbage and Lady • Ada
> > accomplished what, a hundred some-odd years ago? Digital encoding a • hundred
> > years ago, but it took us until the '50s to figure out that it would work • just
> > as well in digital computers as well as a loom?
> Like Leonardo da Vinci's wooden gears, those were ideas that were ahead of
> their time technologically. OOP and GUIs weren't -- they were all totally
> possible technologicaly back then, albeit expensive.
But were Babbage and Ada really ahead of their time? If Babbage could
have done something so mind-bogglingly brilliant as to FINISH A SINGLE
PROJECT, practical digital computing would have come almost a hundred
years earlier. People were aware of what they were doing, but no one
could thnk of a practical use. Surprise, they had a practical use, for
the U.S. Census of 1890, which consisted of several million loom
punch-cards which needed to be processed. No one figured it out until
much later, though, so the census was tabulated by hand and we waited
until the 1940s to hold the next computer evolution.
It's the same with OOP. Cool, okay, but what were they going to *use* it
for? How would it pay the bills? OOP wouldn't, until decades later.
> I think OOP got lost because of the different mindset
> of the older generation of computer people and the newer younger hacker
> generation. Hard to make the case for OOP to unwashed masses of people
> springing up around the microcomputer and just wanting to hack (not that
> there's anything wrong with that).
Someone made the case for LISP to unwashed masses of people springing up
around the microcomputer. Same with LOGO, BASIC, etc. The difference is
that these languages were (at the time) more rapidly prototyped for the
tasks that needed to be done. There was a need for simpler languages.
> But Microsoft goes out of its way to make it harder to write abstraction
> layers on top of its own layers, or more accurately, they make seductive • moves
> to developers to entice them into using system-specific stuff to lock them
> into the MS platform.
This is just a marketing issue. I can use COM and ActiveX and Win32 all I
want, and it's useful for a Win32-only environment. Usually, however, I
find that, contrary to MS propaganda, something a little more generic
(read: potentially cross-platform) will fit the bill. Making my code
fully cross-platform tends to be easy. This is why I write Java code at
work, and why I write Perl code for kicks.
> > These things happen transparently on Win32.
> Rilly? That's good then. So there aren't any functions in the Win32 Perl
> libraries that return backslash for the directory separator? If you call
> the CommonDialog DLL to open a filespec, does that return forward slashes
> to you?
Apples and oranges, Todd. None of the CPAN libraries that I've used
return a file path that would be bad on generic UNIX. (There's the
discrepancy that Perl might be located in //yo-yo/dev/perl/bin instead of
/usr/bin , but you'll find this issue across many UNIX platforms anyway.)
This is a Good Thing, since CPAN is supposed to be comprehensive, and I
expect libraries from it to run across many flavors of OS's.
However, if I load a DLL and start calling Win32-specific code, then I'm
purposefully writing Win32-specific code and cross-platform goes out the
window (no pun intended :-). This too is a Good Thing, in that if someone
tries to run it on a UNIX box, it'll break right away instead of
introducing stealthy bugs into the system.
Compare this to purportedly cross-platform Perl code which has hidden
dependencies on UNIXisms, such as /dev/null or inetd. Bugs like this can
take forever to come out of the woodwork.
> > I can't speak for DOS or VMS or Mac, but I
> > would be genuinely surprised if they didn't exhibit similar behavior.
> OK, how about this. Let's say you're writing a tied hash object to implement
> a simple lookup via a disk file, where the hash key gets mapped into a file
> name. The key has to be encoded, of course, so that characters can be
> restricted to only those permitted in filenames. In Unix, you can use ":"
> in a filename. But in DOS, you can't...the colon means device. And on the
> Mac, the colon is a directory separator. So you could either restrict the
> set of allowable characters to some really low common denominator like
> alphanumerics, or you could exhaustively pick just those characters which
> are available in all of the OS's you support, or you could make the encoder
> based on the $^O variable. Isn't that disgusting?
I just hash the key using the lowest common denominator. It sucks, I know.
I've run across the need to hash filenames just recently, and I was using
MIME::Base64::encode_base64() for the actual hashing, until I realized
that the "/" character is part of base-64 encoding. Totally valid on VMS,
but totally incompatable across UNIX, Mac, and Win32. Now isn't that
terrible! I'll have to write my own base-64 encoder / decoder using
characters known to be good! The moral of the story is: well, there's no
moral. Hacking is an immoral activity. But I think I'm writing better
code because of these mistakes.
> > In fact, if they did introduce such atrocious incompatibilities, that would
> > be the fault of the Perl guys, and not the platform.
> What if they couldn't avoid it?
Then Perl isn't really cross-platform. And the platforms that break it
aren't completely at fault. It happens all the time. Sometimes Win32,
sometimes Mac, sometimes different flavors of UNIX. It'll get fixed, if
there's a need.
> > Downside = Perl in the next version of IIS, a most heinous piece of • software.
> > Upside = A better Win32 Perl, still for free.
> wretch :)
Yes, IIS is wretched! :-P
> > There is a Win32 Perl extension from ActiveState, which allows interaction
> > with OLE objects. In other words, I'll never write another line of VB code
> > again! That, in my view, has helped stike the death blow against VB.
> Stop, you're telling me too much. :-(
What, is that a good thing or a bad thing? I *loathe* VB, but I like the
hooks it gives me into the operating system and into different
applications. Now I can do the same with Perl. It isn't cross-platform,
sure, but neither was tcl/TK for the longest time. That didn't ever
diminish its usefulness.
(Note that I'm no big tcl fan, so it doesn't bother me to compare it to
VB.)
Cheers,
- jsproat
1. OS/2 has a much, much better GUI, IMO. Better user-level OS
interface, too. When MS was designing the Windows 4 look & feel, they
stole a lot from OS/2 but stopped short of using the really *good* stuff,
like file-system links and a better context menu.
|
|
Message has 2 Replies:
Message is in Reply To:
 | | Re: Perl rules!
|
| (...) I don't care if MS is a money-grubbing empire as long as they make great products. The problem I have with it is that they're only making good (not great) products and that, in combination with their monopolistic nature, hurts the other guys (...) (25 years ago, 17-Jul-99, to lugnet.off-topic.debate, lugnet.off-topic.geek)
|
433 Messages in This Thread:
(Inline display suppressed due to large size. Click Dots below to view.)
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
