[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [teskridg@NMSU.Edu: future MCL]



    Date: Thu, 7 May 1992 21:30+0200
    From: Christopher Fry <cfry@mit.edu>

    [...]

    > 
    > More globally, CLIM seems pretty clean and well designed to me.  Also,
    > the specs for CLIM 2 clearly define the protocols to the differents
    > layers of CLIM for easy programmer access and I like that.

    This is the best report on CLIM that I've seen yet. Permit me to probe a little deeper.
    You are speaking of CLIM1 running on MCL, right?

Yes.  CLIM2 is not out yet.

    Is all of the functionality that you describe in the Symbolics "CLIM Release 1.0"
    manual, March 1991? If there is significant additional documentation, please let me know 
    what.

Yes, the Symbolics doc is the one that comes with CLIM1 for MCL2.  I
don't know of other doc.  I agree with you that a simpler tutorial would
be nice (ala Sonya Keene for CLOS).  I was able to get much more
perspective from the CLIM courses I took at Symbolics GMBH (Germany)
with Markus Fischer.  Although a bit expensive, they were very good (few
people and a machine for every student, some course notes, lots of
exercices...).  

What's missing regarding docs is more descriptions of the general
concepts of CLIM, sections ala "to have this behaviour, do this", and a
more consistent vocabulary (I must admit I am not an expert in views,
regions, areas, windows, dialogs,panes...) and it seems everybody uses its own
terminology (mac <> unix <> clim...).  A glossary would be nice.

    Are you familiar with the MCL2 window system?

I am familiar with the Mac system, but have never really used the MCL2
interface system.  I think (but I might be wrong) that it requires quite
a bit of understanding of "Inside Mac" if you want to use more specific
things (like pop-up menus, pointer tracking, dragging,...) while all
this is very simple in CLIM.

    Other than a Mac-look what other features  are in MCL window system that are not easily 
    done in CLIM 1?

I am not sure I can answer this since I don't know the MCL side that
much.  However, I believe that the Mac specific things would be harder
in CLIM (like Quicktime movies, resources support,...).  But then again,
since these things are very specific, you could maybe use MCL's hooks to
the toolbox inside CLIM.  Also, CLIM2 is supposed to allow for machine
specific gadgets.  Maybe this would be the way. Scott McKay or Bill York
will have to say that.

    If you were NOT planning on porting your code to a non-mac, would you prefer
    CLIM 1 over MCL windows?  & why.

I think CLIM is much more "high level" than MCL's interface system.  It
has much more (see the features I mentionned), but it's also much heavyer
stuff (at least 1Mo above MCL I think).  Same difference between CLIM
and MCL's interfacing stuff than between Lisp and C (I hope the MCL
developers won't kill me for this comparison - they know I love MCL!)

    My appliction needs nested multiply nested views, each of which may be independently scrolled.
    Can CLIM 1 support this?

If I understand what "nested multiply nested views" are, yes you can do
that with CLIM (scrolling support is automatic with output recording and
you can also decide what to redisplay manually).

    Is CLIM1 on a Mac significantly slower than MCL windows?

I have made no performance tests, but it probably is.  There is more
overhead, but you get much more too.  However, CLIM is designed to give
programmers access to its low level primitives for performance
improvements (you need the CLIM2 specs for that).

    Reliability?

Good and improving.  See the mailing list clim-request@bbn.com.

    Have you used Quicktime with CLIM1? I'm thinking of showing a videoclip inside
    a pane in a window and am wondering if this is going to be much harder in
    CLIM than MCL windows.

CLIM doesn't provide any support for that.  You will probably use the
same primitives in both cases (see the code of Daniel Ranson for
QuickTime support under MCL).

Vincent