Newsgroups: comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.kei.com!nntp.coast.net!col.hp.com!news.corp.hp.com!hparc12!news
From: Kenneth Green <kenneth_green@hpato.aus.hp.com>
Subject: Is Smalltalk a deadend ?
Sender: news@hparc12.aus.hp.com (News Daemon ID)
Message-ID: <30EB291E.792D@hpato.aus.hp.com>
Date: Thu, 4 Jan 1996 01:10:54 GMT
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Nntp-Posting-Host: hpautord.aus.hp.com
Mime-Version: 1.0
X-Mailer: Mozilla 2.0b3 (X11; I; HP-UX A.09.05 9000/715)
Organization: Hewlett-Packard Company
Lines: 97

If I continue to buy into Smalltalk, am I investing in a dead end ?
===================================================================

This is not intended as a religious slur against anyone's beloved language,
but a genuine dilemma which is raised by the recent explosion of interest
and activity surrounding Java.

As a recent convert to Smalltalk (PPD VW 2.0) I have been excited about its
well known benefits and by the clear productivity increases we have seen on
our trial projects. We are on the verge of beginning some larger
projects and I have been advocating Smalltalk over C++ for the host
application(s).

Over the last couple of weeks I have been looking into Java to see what all
the fuss is about. I have learned a lot and it seems to me that Java,
although immature, could present a serious challenge to Smalltalk.

Consider these points of comparison (which include my projection into what
I see as the likely the near future for Java):

Characteristic			Smalltalk		Java (in say 6 months)
---------------			------------------	----------------------
OO Approach			100%			90%
Rapid Development Cycle		Yes			Yes (but < ST)
Performance			OK			OK (?)
Comprehensive class lib		Very			Limited but growing
Maturity			High & stable		Low,unstable,improving
Portability			Proven & wide		Semi-proven & very wide
Multivendor support		Not interoperable	Yes, all interoprable
Comprehensive Dev Tools		V.Good			Free, beta, immature
CM tools			Expensive & custom	Works with existing
Runtime fees			Expensive		None
Active developers		1000s			10,000s
Active users			10,000s			100,000s to millions
---------------			------------------	----------------------


Overall, both offer a sound OO language with significant advantages over C++
and the rapid development and platform independence benefits of an
interpreted byte-coded language.

In six to twelve months (to allow for Java's evolution to release level):
  Both will have integrated development environments incorporating code
  browsers. editors, compilers, debuggers and profilers etc.

  Both will have extensive class libraries and a growing cohort of third party
  tool and library suppliers.

While Smalltalk is *much* more mature and stable, a huge amount of effort
is going into Java by both Sun and many third parties and we are likely
to see rapid evolution to a pretty useful level.

Java has the significant advantages of free development tools and no runtime
fees. It is likely that once professional quality development tools evolve they
will no longer be free, but runtime fees are vanishingly unlikely.

Another key advantage for Java is its universal standardisation. There is
little sign of convergence of the various Smalltalk flavours, despite progress
on the ANSI standard. There is little or no interoperability or portability
between Smalltalk environments. This problem has plagued Smalltalk's evolution
and limited its success in the marketplace. java is most unlikely to suffer
from this problem. 

The single standard offered by Java will mean that tools, utilities and
libraries can be created by developers and shared with (sold into) a huge
global user community. This is a self fulfilling success algorithm which
ensures that there is much innovation, easier ammortisation of development
costs, access to a large and growing pool of skilled and experienced staff
and price competition to the advantage of developers and end users.

Hence it seems to me that if Java is successful, and I think there is much
which suggests it will be, it will relegate Smalltalk to an ever shrinking
niche market. There will be hundreds of thousands of Java developers and tens
of millions of Java-equipped users.

There is no question that ST offers an excellent and very mature development
environment today, while Java is (no pun intended) very liquid. The picture
will be clearer in, say, 12 months. It would be riksy indeed to commit to Java
for important projects which have to be delivered in the near term. But what
would you choose if you had a slightly longer horizon ?

Thus I ask again, is investment in Smalltalk a step down a dead-end alley ?

Regards,

Kenneth
-- 
 _______________________________________________________________________________
 Kenneth Green				Hewlett Packard Australia Ltd
 Design Engineer			Australian Telecom Operation (ATO)
 					347 Burwood Highway
 Phone: +613 9210 5541			Burwood East   3130
 Fax  : +613 9210 5550			AUSTRALIA
 
 					Postal address:	
 Email: kenneth_green@aus.hp.com	P.O. Box 221, Blackburn 3130, Australia
 _______________________________________________________________________________
