Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!cam-news-feed3.bbnplanet.com!news.bbnplanet.com!cam-news-hub1.bbnplanet.com!howland.erols.net!blackbush.xlink.net!unlisys!cs.tu-berlin.de!news.uni-hamburg.de!news.Hanse.DE!wavehh.hanse.de!cracauer
From: cracauer@wavehh.hanse.de (Martin Cracauer)
Subject: Re: DylanWorks from Harlequin
Message-ID: <1997Jan14.150144.2524@wavehh.hanse.de>
Reply-To: cracauer@wavehh.hanse.de
Organization: '(a (cons structive organization))
References: <32D562C6.2F67@taligent.com> 	<Pine.SGI.3.95.970111105103.15897A-100000@shellx.best.com> <ydrajq52ro.fsf@CLYDE.BOLTZ.CS.CMU.EDU>
Date: Tue, 14 Jan 97 15:01:44 GMT
Lines: 86

Scott Fahlman <sef@clyde.boltz.cs.cmu.edu> writes:

>I think it's clear that Dylan has missed the boat it could and shold
>have caught.  We now have to stand by and watch Java, a much worse
>language in many respects, takes over the niche that could have been
>Dylan's.  But there still should be a place in the world for a
>well-designed and more capable (better than Java) language of this
>general type and, if we can fix up a few warts, Dylan is the best
>candidate to fill this role.

Hm, I don't think Dylan could have taken the niche Java is in
now. 

The bytecode hype is just what Dylan avoids: having bad performance
from start. A few hours reading the JVM specification will lead to the
impression that JITs will *not* speed Java up to native-compiling
languages. Not to speak of general optimization problems. The process
of choosing the highest optimization level that runs the application
in question will be handed from the programmer to the user. How long
will it take when JVMs come up with a switch to set optimization
level? Then, what will users do: Always set to some low level (look at
what Macintosh users did with memory option in the usual case).

The integration-into-webbrowsers is something that wouldn't make sense
for a non-bytecode language. Additionally, although Dylan is designed
to seperate major parts from the development system, I doubt that this
fine-graded incremental transfer of individual class files used for
Java Web Applets is possible with Dylan (maybe someone can comment).

Dylan is a more complex language and I doubt it could have made it to
a "hype" language at all.

Then, the proposed implementations. I don't know what the CMU project
will do about marketing their product. But Harlequin is definitvly not
a vendor interestedt in $199 priced development environments. 

Besides these technical issues, I think the Java (and Web) hype did
more good than bad for other languages. Java forced people into
realizing that more than one language is needed to do computing. The
Web in all its glory lead to applications with much less proposed
livetime, but with even stronger requirements for developement
speed. IMO, People realize that its better to let some individual
implement a project fast and care less for maintainance in ten years
from now.

If I were to decide whether to let someone implement a Web service in
some unusual language, the strongest requirement would be a
development environment that has the best possible browsing
mechannisms and is prepared to show a new user how his/her code
runs. That way, I could make sure that I could find people to at least
modify the application after my wizard left. This environment question
is where java and Dylan are different. Java is bad bejond recognition
in this area, CMU announced very decent support for these tasks and
Harlequin will probably through in at least what they have in
Lispworks.

Then, the Web and Java hype led to better acceptance of less
platform-native look of applications. This is definitvly a plus for
everyone not using the "home" language of a platform and interestedt
in multi-platform development.

People are much more likely to accept other languages than C++ and C
now. At least that is what my impression is. Even FreeBSD, normally
locked into pure C (with a few bits perl for non-important things) now
has an important network service implemented in Modula-3. My customers
for Web application don't longer care at all for the language I
choose.

Last but not least it is now much important to feed ugly platforms
over the network, thus freeing the provider to choose whatever
platform and language implement the needed protocols.

So, it seems to be the market for advanced languages is bigger now
than a few years ago. Althoug a lot of people jumped on Java, my
impression is that the number of people floating around waiting to get
captured into the orbit of some language they like is as big as never
before (counted number of people, not market share).

Sorry for the long post.

Martin
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin_Cracauer@wavehh.hanse.de http://cracauer.cons.org  Fax.: +4940 5228536
"As far as I'm concerned,  if something is so complicated that you can't ex-
 plain it in 10 seconds, then it's probably not worth knowing anyway"- Calvin
