I don't know if I should direct this here or to admin, but it seems most of you read both lists.
Anyway, when I first got CLHP started, I asked if you guys would be willing to install it on the common-lisp.net server. I didn't want to install it until the Apache module was working, and that time has come.
So I'd like to know if we can install it, so I can actually use it for the clhp home page?
On Thu, Nov 13, 2003 at 09:03:36AM -0500, Anthony Ventimiglia wrote:
So I'd like to know if we can install it, so I can actually use it for the clhp home page?
Ok, wearing my policy making hat, there are several separate questions here:
1) Do we wish to offer individual services (in terms of what is provided) to projects?
2) If so, how is this arbitrated? Who gets what, and who decides?
3) How does this relate to services to the lisp community (like the pastebot), when they directly tie in with an existing project?
[ etc... ]
My answers:
1) To a degree, yes.
2) Case by case basis.
- Sometimes providing a service X for a single project makes it trivially available to all the others as well. This is a Good Thing.
- Sometimes a project may need service unimportant or impractical for others. Here we must balance the needs of the single project and the administrative overhead. If the overhead is small, and the service is important for the project, sure.
- In other cases, not really.
3) Common-lisp.net may offer such services to the community. If they tie naturally in with an existing project, then "letting the project provide the service" makes perfect sense.
That said, wearing my voting hat:
I don't know.
How important is it to you? Is this "vital", or "kind of neat"?
How much overhead for Erik (on the long run)? If we do this, will it generate good infrastructure that can be utilized in the future, or will it be a wart on the system?
Cheers,
-- Nikodemus
Nikodemus Siivola writes:
On Thu, Nov 13, 2003 at 09:03:36AM -0500, Anthony Ventimiglia wrote:
That said, wearing my voting hat:
I don't know.
How important is it to you? Is this "vital", or "kind of neat"?
How much overhead for Erik (on the long run)? If we do this, will it generate good infrastructure that can be utilized in the future, or will it be a wart on the system?
I agree with both your questions and answers. And to sway your opinion, I'll offer some information.
As far as the project goes, I consider it vital to have the home page actually running on a clhp back-end.
As far as the overhead for Erik, the installation is pretty straightforward, and I don't really forsee any major server overhead at this point. In my previous post I gave some of the potential problems, which are no more of a risk than running CGI scripts. So it's a matter of trusting users who write clhp pages, and as I said before, I don't see why any of us would have malicious intents regarding the common-lisp.net server. If you guys are still uncomfortable with that, we can restrict the clhp-handler to only work in selected directories.
As far as the infrastructure, I think that it would be a benifit to us all to have it available on the server. Sure we could use PHP to provide the same functionality, but I'd assume that most of us would rather use Lisp than PHP.
Anthony Ventimiglia anthony@ventimiglia.org writes:
Nikodemus Siivola writes:
On Thu, Nov 13, 2003 at 09:03:36AM -0500, Anthony Ventimiglia wrote:
That said, wearing my voting hat:
I don't know.
How important is it to you? Is this "vital", or "kind of neat"?
How much overhead for Erik (on the long run)? If we do this, will it generate good infrastructure that can be utilized in the future, or will it be a wart on the system?
I agree with both your questions and answers.
Me too.... :)
In my previous post I gave some of the potential problems, which are no more of a risk than running CGI scripts. So it's a matter of trusting users who write clhp pages, and as I said before, I don't see why any of us would have malicious intents regarding the common-lisp.net server. If you guys are still uncomfortable with that, we can restrict the clhp-handler to only work in selected directories.
The problem is not necessarily one of us doing malicious things intentionally (I guess that goes without saying :-) ) but unintentionally by screwing up (c.f. Murphy's Law). So yes, it has to be sandboxed.
As far as the infrastructure, I think that it would be a benifit to us all to have it available on the server. Sure we could use PHP to provide the same functionality, but I'd assume that most of us would rather use Lisp than PHP.
Yes, that is true.
Regards, Mario.