I'm pretty certain these are the reasons that smalltalk failed:
1) C was already quite established by the time smalltalk came out
publically (smalltalk actually predates C by 3 years, but, postdates
its public release by 9). It seems dumb, but evidently, your language
needs to be superficially similar, in syntax at least, to the
predominant language of the day. See java and C++, or even the
relatively slow uptake of Python even in the face of active
development (vs. nothing in the perl camp).
2) Marketing. Smalltalk positioned itself originally as a new-fangled
language of the tomorrow also designed for teaching, and chose
ideology over pragmatism. At the time, "Optimization is the root of
all evil" and other modern day mantras were laughed at. Calling
yourself a language for teaching got you scoffed at.
3) Now with the web, you can program in whatever obscure VM-heavy
language you like. As long as you can manage to install it on the
server -you control-, nobody cares. That's not how it worked back
then; you wrote software which ended up at the client. And thus, you
had to develop for the home language of your target platform, or the
shortcuts that are inevitably taken in the name of cross-platform, and
the difficulty of bending the peculiarities of the target OS's 'home'
language to your own language is never perfect. The result was simply
that you couldn't really make software that was nearly as nice for the
end user as in the home language.
The Windows home language is C++. The unix/linux/bsd home language is
C. The Mac home language is Objective C.
If you weren't writing in the home language, you were writing toy
projects. Not fair, perhaps, but that's how it worked.
4) Home languages were picked at least partly on the choices of the
platform developers. You develop kernels and the like in C or
assembler, or something close to it; small talk was never designed for
that sort of thing (and neither is java). Thus, the guys who end up
deciding on the home language for any given platform, tend to have a
soft spot for C and its ilk.
5) Don't kid yourself, VMs blew, and have blown, for years. Only since
the past 6 years or so can you safely ridicule the idiots who think a
VM is some sort of insurmountable performance problem. 10 years+ ago,
you'd be an idiot for thinking that VM performance was something that
would become irrelevant 'any day now'. The smalltalk design pretty
fundamentally boiled down to VMs and very smart compilers to be even
halfway fast; it features a lot of virtual method ('message' in
smalltalk jargon) resolution; the very thing java's VMs have struggled
on for years. Smalltalk didn't even have the highly specialized
optimization opportunities inherent in functional-style programming
going for it (which led to the Lisp machines - whose failure seems to
be principally because the company that built it was incompetent but
didn't go bankrupt due to massive DARPA funding).
Smalltalk didn't have a chance in hell. Then once the web rolled
around and people started figuring out you really don't need to write
CGIBIN crud in C to serve up dynamic webpages, PHP, Perl, and, yes,
java, finally got their enterprise acceptance. Smalltalk wasn't there.
If Smalltalk had had seaside way back then, it might have obliterated
java, perl, PHP, and killed python and ruby before they ever had a
chance. But it didn't, and now smalltalk is one of over a 100
frameworks you can go with; people simply can't find it anymore.
On Sep 30, 8:27 am, kirk <[EMAIL PROTECTED]> wrote:
> Michael Neale wrote:
> > Thanks for that Kirk - quite interesting.
>
> > What do you mean by:
> > "They did so
>
> >> because management failed to recognized that they needed to loosen the
> >> reigns and let people have access in the same what that others were
> >> starting to do at the time."
>
> So this may seem obvious now but it became clear that the future trend
> was that companies were going to make all of their stuff available over
> the web. We failed to convince management that they needed to follow the
> trend. But I think that is more of a side bar.
>
> Other reasons, Smalltalk was seen as highbrow and/or patronizing,
> something that didn't help. Funny thing is, I see some of the same
> traits within the Ruby community, something that I don't see in Scala.
> Ok, that could be taken as flame bait, it really isn't intended as such.
> Just an observation and not a statement about either technology.
>
> Kirk
>
> > A bit more?
>
> > (smalltalk fascinates me both as tech and as history - I am sure you
> > could talk for hours and people like me would still have questions).
>
> > On Sep 30, 6:57 am, kirk <[EMAIL PROTECTED]> wrote:
>
> >> I'm a little late on this thread but being an old Smalltalker I do have
> >> some insight into the problems that lead to it's falling out of grace. I
> >> think there were many. First the old VM technology was much slower than
> >> stuff written in C/C++. Secondly the language it's self is bizzare if
> >> you consider where the vast majority of programmers come from. I know
> >> that shops didn't particularly care for the problems that came with C++.
> >> However moving to something as foreign as Smalltalk just wasn't an
> >> option. Developer seats were expensive and once you picked an
> >> implementation you were locked in. When Java came along it was a natural
> >> bridge between C++ and Smalltalk. C/C++ people could still code the way
> >> they were used to coding. Development shops were also nervous about
> >> Smalltalk coming from a few very small companies. ParcPlace was
> >> vulnerable and that eventually did it in.. just about the time that Java
> >> adoption was starting to take off. So although one might blame Java for
> >> Smalltalks fall, I think it was coming anyways. Sure IBM jumping in on
> >> the bandwagon gave it some legitimacy however....
>
> >> I worked for GemStone for a number of years. IMHO GemStone failed in the
> >> EJB market (even though they had many years of application server
> >> experience in Smalltalk and some time in Java) primarly because they
> >> couldn't make the cultural sift from Smalltalk to Java. WebLogic kicked
> >> "our" asses not because they were better, they weren't. They did so
> >> because management failed to recognized that they needed to loosen the
> >> reigns and let people have access in the same what that others were
> >> starting to do at the time. Again, this is an over simplification.
>
> >> Regards,
> >> Kirk
>
> >> Hamlet D'Arcy wrote:
>
> >>> A guy named James Foster just presented last week at a group I belong
> >>> to. His talk called "The Seaside Heresy" was video recorded and posted
> >>> (it's a bit long):http://programminggems.wordpress.com/2008/09/22/video/
>
> >>> He's from Gemstone, but the talk is about Smalltalk and Seaside (the
> >>> web framework). It's very cool to see the edit-and-continue
> >>> capabilities of Smalltalk played out in a web framework. When testing
> >>> your webapp from the browser, an exception puts you in a debugger, at
> >>> which point you can edit the code (not just variable values!), pop the
> >>> stack frame and continue rendering in the browser where you left off.
> >>> Very cool.
>
> >>> This is an interesting post about Smalltalk too, called What's Good
> >>> about
> >>> Smalltalk:http://www.cincomsmalltalk.com/userblogs/knight/blogView?showComments...
>
> >>> On Sep 29, 6:01 am, "Mark Volkmann" <[EMAIL PROTECTED]> wrote:
>
> >>>> On Sun, Sep 28, 2008 at 10:24 PM, Mark Derricutt <[EMAIL PROTECTED]>
> >>>> wrote:
>
> >>>>> You can just run squeak in headless mode by added the -headless command
> >>>>> line
> >>>>> parameter.
>
> >>>> I've tried that, but haven't been able to get it to work. Can you
> >>>> email me an example command along with the content of an example .st
> >>>> file you pass to the command that works for you? You can send it to me
> >>>> off list at r.mark.volkmann at gmail dot com.
>
> >>>> Thanks!
>
> >>>>> If you want to connect to this server with a GUI, you can
> >>>>> install the RFB package (Remote Frame Buffer -
> >>>>>http://map.squeak.org/package/d4f692a8-c7fa-4d49-927f-74aba7e8fd83)
>
> >>>>> On Mon, Sep 29, 2008 at 4:08 PM, Mark Volkmann <[EMAIL PROTECTED]>
> >>>>> wrote:
>
> >>>>>> The biggest issue seems to be finding an easy way to run a Smalltalk
> >>>>>> application from outside the Squeak environment. It seems that the
> >>>>>> proponents of it feel it is acceptable to have users run applications
>
> >>>> --
> >>>> R. Mark Volkmann
> >>>> Object Computing, Inc.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "The
Java Posse" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/javaposse?hl=en
-~----------~----~----~----~------~----~------~--~---
