From: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX - 2008-08-12 11:01:48
|
It's great to see the reactions of the past days. It's becoming clear I'm definitely not the only one with interest in ABCL. However, I didn't see anybody express the any concerns with the fact that my primary interest is with ABCL. Is there truely that little interest for J? The answer to the question above may give some direction on where to go next with the project. BTW: I changed the "long project name" from "J" to "ArmedBear / J" and made the description of the project more about ArmedBear. Somehow, ABCL feels more of an achievement of this project than J (although that's an achievement too). Thanks in advance for any reactions. Bye, Erik. |
From: Mark E. <ev...@pa...> - 2008-08-12 13:47:55
|
XXXXXXXXXXXXXX wrote: > It's great to see the reactions of the past days. It's becoming clear > I'm definitely not the only one with interest in ABCL. > > However, I didn't see anybody express the any concerns with the fact > that my primary interest is with ABCL. Is there truely that little > interest for J? Every time I have interest in J (i.e. I want to script Java Swing elements from ABCL), I end up rolling my own. But it is real nice to have J around to compare my implementation. I prefer to use the a little syntactic sugar like (#"registerCallback" object (lambda (arg) (display arg))) (this syntax was introduced by Alan Ruttenberg with [JSS][1]. I don't think this is any fault of J, but introducing a notion of "core" and "extension" to ABCL would be very interesting for me, primarily in response to my thinking of whether a Java Web Startable ABCL makes sense [abcld][2], and then how to meaningfully limit the initial download. [1]: http://svn.mumble.net:8080/svn/lsw/trunk/jss/ [2]: http://code.google.com/p/abcl-dynamic-install/ -- Mark Evenson <ev...@pa...> "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now." |
From: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX - 2008-08-12 14:31:44
|
Forgot to send the mail below to the ML. On Tue, Aug 12, 2008 at 4:30 PM, XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX wrote: > On Tue, Aug 12, 2008 at 3:47 PM, Mark Evenson <ev...@pa...> wrote: >> XXXXXXXXXXXXXX wrote: >>> >>> It's great to see the reactions of the past days. It's becoming clear >>> I'm definitely not the only one with interest in ABCL. >>> >>> However, I didn't see anybody express the any concerns with the fact >>> that my primary interest is with ABCL. Is there truely that little >>> interest for J? >> >> Every time I have interest in J (i.e. I want to script Java Swing elements >> from ABCL), > > That's something I'd like to do too, however, I don't really see what > J can add to it, because ... > >> I end up rolling my own. >> But it is real nice to have J around to compare my implementation. > > I don't really understand this remark: I've looked at J just now > (again) and it's 100% java (not 99.9, really 100%). What are you > referring to? Is what you're referring to in J or in ArmedBear? > >> I prefer to use the a little syntactic sugar >> like >> >> (#"registerCallback" object (lambda (arg) (display arg))) >> >> (this syntax was introduced by Alan Ruttenberg with [JSS][1]. > > That's nice syntax, especially given the dynamic dispatch. I was > working at doing something like it, but less advanced. It looks a bit > like the CCL (aka OpenMCL) Lisp-Cocoa bridge. > > BTW: binding to Swing elements is not something I'd like to call "J". > It's something that in the broader context of ABCL can be very useful. > What I consider "J" is roughly the src/armedbear/j directory in the > source tree: the editor (which looks inspired by emacs) with its > supporting files. Do we have different definitions of "J"? > >> I don't think >> this is any fault of J, but introducing a notion of "core" and "extension" >> to ABCL would be very interesting for me, primarily in response to my >> thinking of whether a Java Web Startable ABCL makes sense [abcld][2], and >> then how to meaningfully limit the initial download. > > This seems like a nice project. Do you have any special uses in mind? > > > > Bye, > > Erik. > |
From: Mark E. <ev...@pa...> - 2008-08-13 07:39:28
|
XXXXXXXXXXXXXX wrote: CCL (aka OpenMCL) Lisp-Cocoa bridge. >> >> BTW: binding to Swing elements is not something I'd like to call "J". >> It's something that in the broader context of ABCL can be very useful. >> What I consider "J" is roughly the src/armedbear/j directory in the >> source tree: the editor (which looks inspired by emacs) with its >> supporting files. Do we have different definitions of "J"? >> Hmm. Maybe I don't understand J well enough, but I thought the core of J was actually ABCL that uses "bindings" to Swing to actual affect graphical elements (like a text buffer). The buffer is usually manipulated in pure CL. Maybe this is only the architecture that I *thought* J had >>> I don't think >>> this is any fault of J, but introducing a notion of "core" and "extension" >>> to ABCL would be very interesting for me, primarily in response to my >>> thinking of whether a Java Web Startable ABCL makes sense [abcld][2], and >>> then how to meaningfully limit the initial download. >> This seems like a nice project. Do you have any special uses in mind? >> I started with: save-javavm-and-die but then I developed some other ones. If one really wishes minimal load times in a network context, one has to really develop some network load profiling tools/call-graph. The current monolithic org.armedbear.lisp package makes breaking up the load path in discrete jars (i.e. "org.armedbear.lisp.core.jar" followed by "org.armedbear.lisp.clos.jar") rather tough without transforming the source into "synthethic" classes. Of course real work intervened with ABCLD, and mainly it has devolved into a place where I think about how to refactor JSS into pure ABCL, meanining no need for 'bsh.jar' and 'jscheme.jar' in the classpath. Then the code would presumably get folded back into the main ABCL distribution, and then I would return to working on how to minimize load graph times. Mark -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now." |