scsh-users
[Top] [All Lists]

tk

To: scsh@martigny.ai.mit.edu
Subject: tk
From: shivers@ai.mit.edu (Olin Shivers)
Date: 10 Jan 1996 19:33:04 -0500
Organization: Artificial Intelligence Lab, MIT
Reply-to: shivers@ai.mit.edu
   From: markf@ai.mit.EDU (Mark Friedman)
   Newsgroups: alt.lang.scheme.scsh
   Date: 26 Dec 1995 15:07:12 -0500

   To me it is bad engineering to create such un-reusable, un-abstracted
   software. I am willing to give Ousterhout the benefit of the doubt on
   the design side. Perhaps he in fact "designed for reuse" and just blew
   it in the implementation. I don't know what the design of Tcl/Tk
   intended but I know what the implementation does.

You are too kind. Building tcl into the tk interface is design, not
implementation. It's not like using the wrong internal data-structure,
it affects what the client sees. That's bad design.

I jumped Osterhout about this at lunch a few months ago. He just said,
basically, people use it, so it must be good.

There are certainly lessons the advanced language people could learn from
tcl. I pay attention, much as I am revolted by the language. You cannot
take away from John that he had a big hit with tcl. Probably bigger than any
of the deeper academic things he's ever done in his whole career. The
advanced language people (like me) that hold their noses and grimace could
stand to ask themselves: why have none of *my* languages *ever* transited
out into the real world and made a difference? That's a professional
embarassment. 

However, I will make the following prediction: tcl is the walking dead.
Visual Basic and Java are the wave.
        -Olin

<Prev in Thread] Current Thread [Next in Thread>
  • tk, Olin Shivers <=