Nikodemus Siivola <nikodemus@random-state.net> writes:
With equal formality, I approve.
In the interest of keeping our users happy, do you (Nikodemus) think we should go ahead and create the Cello project given that it's basically the same as Cells (with regards to license etc.) and given the fact that Mario approved Cells? Erik.
On Fri, Nov 07, 2003 at 03:00:16PM -0500, Erik Enge wrote:
In the interest of keeping our users happy, do you (Nikodemus) think we should go ahead and create the Cello project given that it's basically the same as Cells (with regards to license etc.) and given the fact that Mario approved Cells?
I suggested that Kenny apply for Cello separately mainly for organizational clarity (and the hunch that there are plenty of people who may be interested in one, but not the other). ;) ...but give Mario a day or two to respond, unless Cello is in urgent need of CVS space? (Kenny: If that is the case (danger of forking due to off-cvs parallel development, etc), let Erik know so that he can create it. In that case will worry about the administrivia later...) There is actually room here for a devil's advocate argument: "Should Common-lisp.net host multiple projects with greatly overlapping objectives?" For what it's worth, I think yes. Common-lisp.net should be a fertile ground where most flowers can bloom. Once we have MORE CODE we can start sorting the wheat from the chaff, but even then Common-lisp.net isn't probably the place to do it. Cheers, -- Nikodemus
Nikodemus Siivola wrote:
On Fri, Nov 07, 2003 at 03:00:16PM -0500, Erik Enge wrote:
In the interest of keeping our users happy, do you (Nikodemus) think we should go ahead and create the Cello project given that it's basically the same as Cells (with regards to license etc.) and given the fact that Mario approved Cells?
I suggested that Kenny apply for Cello separately mainly for organizational clarity (and the hunch that there are plenty of people who may be interested in one, but not the other). ;)
...but give Mario a day or two to respond, unless Cello is in urgent need of CVS space?
Absolutely not. I have to switch from Freeglut 1.x to a shiny new FG2.0, which will wreak havoc since I forked FG1 pretty badly. This will also mean I need to grab, build, and learn a couple of (open) text libs, because I also added text handling to FG. So my work is cut out for me before I have anything to release.
(Kenny: If that is the case (danger of forking due to off-cvs parallel development, etc), let Erik know so that he can create it. In that case will worry about the administrivia later...)
There is actually room here for a devil's advocate argument:
"Should Common-lisp.net host multiple projects with greatly overlapping objectives?"
I have zero problem with doing all this under one project (I'd just have separate bundles/directories/whatever for Cells and Cello) and it might even be a lot more manageable that way until/if one or the other catches on. I can see myself over on a Cello list saying, "Oh, that's a Cells question, let's take it to cells-devel so folks there can see it". Ugh, sounds awful. :) One project, then? :) kenny -- http://tilton-technology.com Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film Your Project Here! http://alu.cliki.net/Industry%20Application
On Sat, Nov 08, 2003 at 01:14:04AM -0500, Kenny Tilton wrote:
"Should Common-lisp.net host multiple projects with greatly overlapping objectives?"
I have zero problem with doing all this under one project (I'd just have
Heh. ;) That was actually a reference to multiple GUI-projects, as we already host Mario's lgtk (GTK for CL) project.
separate bundles/directories/whatever for Cells and Cello) and it might even be a lot more manageable that way until/if one or the other catches on. I can see myself over on a Cello list saying, "Oh, that's a Cells
...I really think they're stunting each other already. Take me for example: I'm definitely interested in Cells, but could not currently care less for Cello. ;)
One project, then? :)
As an admin: "Your choise." As a lisper: I haven't looked at the codebase, so I have no idea of the boundaries between them -- so this is rather theoretical. ;) Cells is a "dataflow oriented extension" to Common Lisp, in a sence analogous to CLOS. I think that if the paradigm and ideas of Cells are universal enough, then Cells (or another implementation of that paradigm) has the potential (yes, I know this is unlikely) to become "another CLOS" on the long term. Cello on the other hand seems to be at least three totally different projects rolled into one: 1) An experimental CLIM-like system based on Cells instead of CLOS. 2) A backend for the above, built on top of OpenGL. 3) OpenGL bindings for Common Lisp. Now, *if* these perceptions are correct (I'm largely guessing here), then there really should be three projects if the CL community were to reap maximum benefit from all this: Cells, Cello, cl-opengl <half-serious> On a totally different topic: Erik, I'm getting the feeling that we're starting to have enough people on Common-lisp.net that it might make sense to start talk@common-lisp.net -- a mailing list equivalent of cll for people who certifiably are writing open CL stuff instead of just alking about it. </half-serious> Cheers, -- Nikodemus
Nikodemus Siivola wrote:
On Sat, Nov 08, 2003 at 01:14:04AM -0500, Kenny Tilton wrote:
"Should Common-lisp.net host multiple projects with greatly overlapping objectives?"
I have zero problem with doing all this under one project (I'd just have
Heh. ;) That was actually a reference to multiple GUI-projects, as we already host Mario's lgtk (GTK for CL) project.
Oh. OK, looking back then to:
For what it's worth, I think yes. Common-lisp.net should be a fertile ground where most flowers can bloom.
Once we have MORE CODE we can start sorting the wheat from the chaff, but even then Common-lisp.net isn't probably the place to do it.
<heh> I can tell you already that I will continue the open thing only if others vote for it with their support. It is too much trouble otherwise. So there should be a nice self-selection thing going on. If resources are scarce you might want to evict inactive projects (I'll let you know if I pull the plug, btw.) Over on SourceForge I think there are tons of projects that never got further than being created (mine included until recently). It is interesting to see selectivity (by the admins) even being considered. You'd have to update the mission statement to "open source boutique" or something.
separate bundles/directories/whatever for Cells and Cello) and it might even be a lot more manageable that way until/if one or the other catches on. I can see myself over on a Cello list saying, "Oh, that's a Cells
...I really think they're stunting each other already. Take me for example: I'm definitely interested in Cells, but could not currently care less for Cello. ;)
POI: "stunting each other"? You mean Cells is hurt by the Cello thing, in that one gets the idea they are inseparable? The names are almost inseparable, maybe that's the problem. I know I have encountered confusion in others on that score.
One project, then? :)
As an admin: "Your choise."
As a lisper:
I haven't looked at the codebase, so I have no idea of the boundaries between them -- so this is rather theoretical. ;)
Cells is a "dataflow oriented extension" to Common Lisp, in a sence analogous to CLOS. I think that if the paradigm and ideas of Cells are universal enough, then Cells (or another implementation of that paradigm) has the potential (yes, I know this is unlikely) to become "another CLOS" on the long term.
Aside: I think it is very likely the dataflow paradigm will prevail, and I /know/ Cells won't because a fundamental rewrite is planned (and needed), and in the end brighter folks than me will have to find the Right Way. Dataflow is the Lisp of paradigms, first implemented in 1962, then Steele tried again with his PhD thesis cum language in the 80s, now Garnet and RETE and COSI and Cells and a couple people that approached me at ILC2003 with news of their own similar hacks. Back on topic: right, this has nothing to do with GUI, per se. GUIs are just the killer app for Cells.
Cello on the other hand seems to be at least three totally different projects rolled into one:
1) An experimental CLIM-like system based on Cells instead of CLOS.
? Cells extend CLOS (and defstructs are on the way). Cello is all CLOS all the time.
2) A backend for the above, built on top of OpenGL.
3) OpenGL bindings for Common Lisp.
Now, *if* these perceptions are correct (I'm largely guessing here), then there really should be three projects if the CL community were to reap maximum benefit from all this:
Cells, Cello, cl-opengl
God, I'd love a cl-opengl, tho I have enough bindings done to keep me happy for a while. But I ducked the scary ones. :)
.. people who certifiably are writing open CL stuff instead of just alking about it.
don't get me started. <serious> :) kenny -- http://tilton-technology.com Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film Your Project Here! http://alu.cliki.net/Industry%20Application
On Sat, Nov 08, 2003 at 10:32:55AM -0500, Kenny Tilton wrote:
I can tell you already that I will continue the open thing only if others vote for it with their support. It is too much trouble otherwise.
I'm not sure what you mean by support... if you mean "people rushing in to actively write code", I doubt you'll ever be overwhelmed. But as for getting users who may occasionally send bug-reports or submit their hacks, I think you're of to a good start: Looking at http://common-lisp.net/stats/ I see 98 hits on the Cells homepage already. ;)
If resources are scarce you might want to evict inactive projects (I'll let you know if I pull the plug, btw.) Over on SourceForge I think there are tons of projects that never got further than being created (mine included until recently).
Since disks are cheap, I hope we never have to face such a decision. I'd rather hope that Common-lisp.net CVS would in ten years time be analogous to the CMU AI repository now: some of the stuff may be old, but a lot of it is still damn good.
It is interesting to see selectivity (by the admins) even being considered. You'd have to update the mission statement to "open source boutique" or something.
Not unique really. Consider http://www.tigris.org. The problem I see with the Tigris approach is that it's really hard to predict what will succeed on the long term. That's why I like our current approach of keeping the barrier of entry to CL net fairly low. OTOH, I do hope that the requirement to send email without the guide provided by prefilled forms like sf.net and savannah provide serves to pre-filter the truly low end of the code-pool. (Yes, I deserve to be flamed for talking like this.)
POI: "stunting each other"? You mean Cells is hurt by the Cello thing, in that one gets the idea they are inseparable?
Stunting each other as in: * There are probably people who are eager to hack on CL GUI thing that is a) unique to CL, b) not CLIM. These people, however are not necessarily (at least initially) the least interested in Cells. * There are probably people who are eager to experiment with the dataflow paradigm and Cells, but don't give a damn about GUI's, and don't want to follow mailing lists that talk about them. Dunno. I'm probably just paranoid.
1) An experimental CLIM-like system based on Cells instead of CLOS.
? Cells extend CLOS (and defstructs are on the way). Cello is all CLOS all the time.
<handwaving>I ment Cells as a sort of superset of CLOS, or something like that. To paraphrase KMP: as CLIM uses the power of CLOS to "bend the space-time" I suspect Cello uses the power of Cells.</>
God, I'd love a cl-opengl, tho I have enough bindings done to keep me happy for a while. But I ducked the scary ones. :)
Heh. There seem to be plenty of cl-opengl bindings going around. I'm sort of hoping that one of them decides to use a public CVS (cl.net, sf, whatever) and a MIT/X/BSD style license. ;) Cheers, -- Nikodemus
Quoting Nikodemus Siivola <nikodemus@random-state.net>:
<half-serious> On a totally different topic: Erik, I'm getting the feeling that we're starting to have enough people on Common-lisp.net that it might make sense to start talk@common-lisp.net -- a mailing list equivalent of cll for people who certifiably are writing open CL stuff instead of just alking about it. </half-serious>
I thought that was what clump was supposed to do? Erik.
Erik Enge wrote:
Quoting Nikodemus Siivola <nikodemus@random-state.net>:
<half-serious> On a totally different topic: Erik, I'm getting the feeling that we're starting to have enough people on Common-lisp.net that it might make sense to start talk@common-lisp.net -- a mailing list equivalent of cll for people who certifiably are writing open CL stuff instead of just alking about it. </half-serious>
I thought that was what clump was supposed to do?
Erik.
That's generally the charter for clump, yes.
<half-serious> On a totally different topic: Erik, I'm getting the feeling that we're starting to have enough people on Common-lisp.net that it might make sense to start talk@common-lisp.net -- a mailing list equivalent of cll for people who certifiably are writing open CL stuff instead of just alking about it. </half-serious>
I thought of making that recommendation as well, since cll is not common-lisp specific, it might be nice to start an cl exclusive list. Maybe we won't have to waste time filtering through all the scheme vs lisp and python vs lisp threads that seem to be so dominant. -- (incf *yankees-world-series-losses*)
Quoting Kenny Tilton <ktilton@nyc.rr.com>:
One project, then? :)
You're probably right that "oh, sorry, that would be a Cells question"-situations would arrise, so I'm with you, one project until you've got enough userbase that there is a problem of having just one project. That aside, you say that having this opened up like this will only continue as long as people use/support it and that hadn't it been for that fact, having it open would've been to much work. For me, having my projects on common-lisp.net where they are backed up, on high-quality SCSI harddrives, where I have FTP/web/ViewCVS/CVS/AnonCVS/NNTP/Mailings support without having to deal with any of that myself is far more convenient than having my software on my home box where a) noone else can get to it so I become the bottleneck and b) backups aren't that great. I also use XEmacs with its CVS support so I wouldn't have notice whether my code was on my local box or not in any case. You really want to check out http://www.cs.utk.edu/~england/ssh.html to make sure you can do password-less operations on your CVS tree. common-lisp.net is supposed to be of real value to CL developers and the CL community. If you can identify issues which make it a hassle, I (and I'm sure the other guys) would be very greatful for any advice you can give us so we can avoid those in the future. Erik.
Kenny Tilton writes:
"Should Common-lisp.net host multiple projects with greatly overlapping objectives?"
I have zero problem with doing all this under one project (I'd just have separate bundles/directories/whatever for Cells and Cello) and it might even be a lot more manageable that way until/if one or the other catches on. I can see myself over on a Cello list saying, "Oh, that's a Cells question, let's take it to cells-devel so folks there can see it". Ugh, sounds awful. :)
One project, then? :)
I agree with this approach, since I may eventually end up making some sub-projects based on CLHP, if they are closely related, it might be better to keep them in a single project. -- (incf *yankees-world-series-losses*)
participants (5)
-
Anthony Ventimiglia -
Erik Enge -
Kenny Tilton -
Miles Egan -
Nikodemus Siivola