Just gotta know - does anyone use CLIM ... ???

I'm sorry, but this is a rant, of sorts...

I've posted question regarding CLIM in the past, including today.
They've always gone without reply, leading me to draw the following conclusions:

(1) They are incredibly stupid questions thus not deserving of a response. 

(2) No one has anything useful to say. This I doubt - the Lisp programmers I've 
know over the last 15 years are a bright lot.

(3) No one uses CLIM! I do. I'm living in the past???

Thanks,
Greg

Re: Just gotta know - does anyone use CLIM ... ???

No, I do not use CLIM, I use CAPI.  If you can post
some code illustrating your problem I will do the best I
can to puzzle it out.  From your post:

> I'm trying to create a MDI-type interface with CLIM. There's a single application frame
> as the top-level. Thru this top-level the user can create multiple instances another
application.
> The problem I'm having seems to be an inability to have the child frame integrated into
the
> top-level frame's read-command loop.

> Thoughts? Answers?

I am unsure of what you mean.  I think you have some a Lisp Listener in the parent
frame.  How is the child frame behaviour supposed to be affected by the top-level
repl?  What do you mean by integrated?

In CAPI the execution of a Lisp expression in a capi:listener-pane can execute
any arbitrary Lisp code, which includes changing the state of any displayed pane
within the application.  I think this would also apply to CLIM.

Wade


----- Original Message -----
From: <greg@adrenaline.com>
To: <lisp-hug@xanalys.com>
Sent: Monday, October 21, 2002 5:57 PM
Subject: Just gotta know - does anyone use CLIM ... ???


|
| I'm sorry, but this is a rant, of sorts...
|
| I've posted question regarding CLIM in the past, including today.
| They've always gone without reply, leading me to draw the following conclusions:
|
| (1) They are incredibly stupid questions thus not deserving of a response.
|
| (2) No one has anything useful to say. This I doubt - the Lisp programmers I've
| know over the last 15 years are a bright lot.
|
| (3) No one uses CLIM! I do. I'm living in the past???
|
| Thanks,
| Greg
|

Re: Just gotta know - does anyone use CLIM ... ???

* greg  wrote:

> (1) They are incredibly stupid questions thus not deserving of a response. 

> (2) No one has anything useful to say. This I doubt - the Lisp programmers I've 
> know over the last 15 years are a bright lot.

> (3) No one uses CLIM! I do. I'm living in the past???

My guess is that, for (3): people use CLIM, because I doubt the
vendors would support it if there weren't customers asking for it.  My
*guess* is that there are a small number of large customers with large
code bases that use CLIM, who are keeping it alive.  For (1), no I
don't think they're stupid.  For (2), well, having seen some of a CLIM
implementation (not the LispWorks one) and read the spec fairly
thoroughly, when people ask questions I just want to go and hide in a
corner somewhere: I think that there are really no good answers to
anything to do with CLIM.

Don't get me wrong, there are good ideas in there - very good ideas in
fact - like the whole notion of presentations, but I think they need
to be extracted and put into some framework that could more easily sit
on top of a modern GUI and was more clearly implementable.  Of course,
the effort required to do that, in a cross-implementation way, is
almost certainly beyond the financial means of the vendors (just as
getting CLIM to work right was probably beyond their means in the late
80s...).

I actually wish that Xanalys would tidy up and *document* CAPI, and
present it as a thing-that-could-be-implemented-by-anyone, as I'd much
rather see an Open CAPI + presentations than an Open CLIM (though the
people who are doing the free CLIM are doing heroic work, I think).
Even in its current state, CAPI provides a pretty good cross
Windows/Unix environment.  Of course I *completely* realise that doing
something like this is a huge amount of work with probably small if
any financial return - I don't want what I've said to be taken as
critical of anyone for not doing it!

--tim

Re: Just gotta know - does anyone use CLIM ... ???

Tim Bradshaw <tfb@cley.com> writes:

> * greg  wrote:
> 
> > (1) They are incredibly stupid questions thus not deserving of a response. 
> 
> > (2) No one has anything useful to say. This I doubt - the Lisp programmers I've 
> > know over the last 15 years are a bright lot.
> 
> > (3) No one uses CLIM! I do. I'm living in the past???
> 
> My guess is that, for (3): people use CLIM, because I doubt the
> vendors would support it if there weren't customers asking for it.  My
> *guess* is that there are a small number of large customers with large
> code bases that use CLIM, who are keeping it alive.
Than I would expect it to work quite better. I ran into a simular
problem to you. I wanted to understand CLIM and I implemented the
easiesst stuff I could think of under OpenGenera. Well there it has
worked. I than tried to port it to LispWorks CLIM and it did not
work. While looking through the CLIM Demos I found immediatly some
(bugs?) Just try the Address Example. If you wipe out the entry for
the telephone number in the upper left area. Than you can not enter
new values again. The right mouse click usually does nothing althoug I
guess it should do some things.



>  For (1), no I
> don't think they're stupid.  For (2), well, having seen some of a CLIM
> implementation (not the LispWorks one) and read the spec fairly
> thoroughly, when people ask questions I just want to go and hide in a
> corner somewhere: I think that there are really no good answers to
> anything to do with CLIM.
Well I can't tell for other, but it's nearly impossible to me to get
into CLIM. Simple questions on how to get push buttons into an
Appliation which does something when I click on it are digging through
masses of documentation where one can find burried some infos. Than
it's not said that I get something going. Fair enough you can find
elements of CLIM in CAPI well it gives you some idea on how things
might work. But it tedious.

> 
> Don't get me wrong, there are good ideas in there - very good ideas in
> fact - like the whole notion of presentations, 
Well I think the whole idea of presentations and accetping values is
brilliant. Anyway I do not get it properly I assume. 


>but I think they need
> to be extracted and put into some framework that could more easily sit
> on top of a modern GUI and was more clearly implementable.  Of course,
> the effort required to do that, in a cross-implementation way, is
> almost certainly beyond the financial means of the vendors (just as
> getting CLIM to work right was probably beyond their means in the late
> 80s...).
Well there is a undertaking to implement a free CLIM clone I wonder if
you might get some more informations from their side.

> 
> I actually wish that Xanalys would tidy up and *document* CAPI, and
> present it as a thing-that-could-be-implemented-by-anyone, as I'd much
> rather see an Open CAPI + presentations than an Open CLIM (though the
> people who are doing the free CLIM are doing heroic work, I think).
Well it's reverse engeneering and learning an own mindset. I guess the
authors doing that undertaking probably have known CLIM before.

There are anyhow some interesting questions. Why does all the vendors
provide their GUI stuff and have not adopted CLIM. Well it has worked
for them quite some time, so why does it seem that it's not well
supported anylonger. Other intersting questions why isn't there just
one book avaliable as a sort of tutorial. I even could buy a book
about Lisp Machine programming a year or so ago!

It would be really nice to hear from the Xanayls people why they do
not have followd the CLIM approach. 

Regards
Friedrich

Re: Just gotta know - does anyone use CLIM ... ???

Unable to parse email body. Email id is 636

Re: Just gotta know - does anyone use CLIM ... ???

* Friedrich Dominicus wrote:
> Than I would expect it to work quite better. I ran into a simular
> problem to you. I wanted to understand CLIM and I implemented the
> easiesst stuff I could think of under OpenGenera. Well there it has
> worked. I than tried to port it to LispWorks CLIM and it did not
> work. While looking through the CLIM Demos I found immediatly some
> (bugs?) Just try the Address Example. If you wipe out the entry for
> the telephone number in the upper left area. Than you can not enter
> new values again. The right mouse click usually does nothing althoug I
> guess it should do some things.

I think you'd expect it to work better *for the applications that it's
being supported for*.  Given finite resources they can only fix some
things...

> There are anyhow some interesting questions. Why does all the vendors
> provide their GUI stuff and have not adopted CLIM. Well it has worked
> for them quite some time, so why does it seem that it's not well
> supported anylonger. 

Well, for one thing all the vendors actually have adopted CLIM - I
think it works in all the commercial systems except Corman (if you
count that as commercial).  I also don't think it's worse supported
than it was - you're looking back to some fictitious golden age when
CLIM worked, when in reality it *never* worked very well, largely
because it never actually got finished (either in specification or
implementation). And I think there are just millions of reasons why
vendors might want to provide some different GUI framework.  CLIM,
after all, is really just an attempt to write something that works
like Genera but uses CLOS instead of Flavors, and runs on top of other
window systems (which really meant X11 at the time it was written).
Well, that's cool, and there were nice things in the Genera window
system, but it didn't work *anything like* the way people expect GUIs
to work now.  What end users actually want is applications which work
pretty much like all the other Windows/Motif/Gnome/etc applications
work. What they really *don't* want is something that looks like a
LispM - you may, I may, but end users don't.  So, lo and behold, the
vendors are producing things like CAPI which lets you write mildly
windowsy applications without running through enormous hoops.

> Other intersting questions why isn't there just
> one book avaliable as a sort of tutorial. I even could buy a book
> about Lisp Machine programming a year or so ago!

Well, the reason you could buy a book on LispM programming is that
when it was written there was money in writing books on LispM
programming, and it's cheap to keep books in print (especially if you
printed slightly too many, and you have a cheap warehouse).  For there
to be books on CLIM, it needs to be worth someone's while to write
them.  It probably takes a year to write a good book, and if you want
a competent user of the system to write such a book it's therefore
going to cost something like $50,000 to $100,000, since that's what
they'd be earning otherwise. I'd *love* to write a book on CL (though
not on CLIM), and I bet others would too, but since I'm not an
academic (and probably essentially no good CL people are academics who
have not already written books on it) I'm not about to do it, because
I need to eat, and writing a Lisp book isn't going to keep me in food.

--tim

Re: Just gotta know - does anyone use CLIM ... ???

* greg  wrote:

> The larger issue I believe to be a community issue. Does the Lisp
> community (however big it is?) want a portable UIMS? Is CLIM the
> UIMS they want? Do the vendors care? 

For me: yes definitely, no, definitely not, not a vendor.  For *me*
the issue is: how do you pay for such a thing?

--tim