🚀 go-pugleaf

ede978e@monu6.cc.monash.edu.au (Robert D. Nicholson) writes:

>Is this a fair evaluation based on the criteria I have set.

>[...some interesting points deleted...]

>To sum up, the Nextstep development environment is well thought out and
non-trivial applications are less troublesome to create in this environment
than they are in most other \oo programming environments.  The drudgery of
                                                           ^^^^^^^^^^^^^^^
putting together the application's interface is simpified by using Interface
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Builder.  Application development under Nextstep is by no means a trivial
process.  Interface Builder does have some significant short commings.
Firstly, primitive graphical objects cannot be constructed in the
> tool, this greatly reduces Interface Builder's flexibility and
consequently, interface designers often find their creativity stiffled by
the tool.  Secondly, Testing the interface only gives a marginal impression
as to what the interface will actually feel like and lastly, sophisticated
interactive behaviour must be specified outside the tool in the underlying
programming language.
>---------------------

I have high-lighted a key phrase up there that points to the basic
reason for Interface Builder's existence -- to build interfaces.  It is
not, and I don't think has ever been represented as, a tool for general
code generation and testing.  (IMHO ad nauseum.)

I'm not sure what you're referring to when you say "primitive graphical
objects cannot be constructed."  Are you talking about the ability to
put boxes, lines, circles, etc. on components of your interface?  If so,
NeXT probably explicitly left this out for reasons of style.

I would debate the first part of your final summary sentence, "testing
the interface only gives a marginal impression as to what the interface
will actually feel like."  I think it gives a good impression of, and
ONLY of, the feel of the interface.  The second part of the sentence is
true, but is beyond the scope of what IB is intended for at this stage
of the game.

To put this in perspective, the UNIX user group I'm a member of had a
guest speaker this week, who talked about GUI development in an X
Window/OSF Motif environment.  She has a lot of experience in the area,
and never said a word about alternatives like NeXTSTEP, but the talk was
nevertheless about as strong an argument for the advantages of NeXTSTEP
as I've heard from a non-partisan speaker.  :~)

There were comments like GUI-builder tools worth using costing $12K -
$20K for one seat, a universal inability to make use of the tool to change
the interface components after custom code is added, the use of shared
libraries for cross-platform development (e.g. Windows, Mac) increasing
development time by a factor of two, the necessity to learn the Xlib and
Xt Intrinsic layers (which are bottom-level) because you frequently want
to do things that Motif (or Open Look or whatever) can't handle, GUI
development involving "major expense" in terms of learning curve and
effort (and hardware, but we probably accept that as a given), etc.
(These are her points, not mine.)

Obviously, X and its associated wares offer advantages to people or
nobody would ever use it.  But I sure am glad I'm using NeXTSTEP instead.
Despite the observations Robert has made, which are true, though mostly
beyond the intended scope of IB, we appear to have one of the very best
games in town.
-- 
Royce Howland, DKW Systems Corp.   |  "And since OS/2 2.0 is a 32-bit
       Everything is IMHO          |   operating system, programs are easier
royce@splunge.uucp  (NeXTMail OK)  |   to write and run faster, too."
or kakwa!atlantis!splunge!royce    |  ad for OS/2 2.0