freenode/#clim - IRC Chatlog
Search
15:46:45
nyef
I don't know. I think that the problem is more social/political, on the interpersonal rather than international scale, than technical.
15:49:27
nyef
Possibly, but it's more... there's a scale thing, in terms of size-of-project, and there's also bits where just nobody is putting in the work.
15:50:32
nyef
In SBCL, for example, I am completely confident that nobody fully understands dynamic-extent.
15:51:35
nyef
So there's improvements that can in theory be made, but there's a decent investment required, and comparatively little reward because it mostly works, and at least isn't known to be broken.
15:52:30
nyef
The flipside of this is how many "easily installable lisp environment" projects have there been over the years?
15:53:46
nyef
But Quicklisp? It may have a truck number of ONE, if anything happens to Xach we're in trouble, but basically *everybody* uses it.
15:56:23
nyef
But how many UI toolkits do we have? Bindings for Qt, bindings for gtk, wxwidgets, other obscure APIs, Cocoa, WinAPI, just talking to X11 by way of CLX...
15:58:08
nyef
There's clearly pain there. Clearly people have tried to solve it, at least for themselves and shared the work.
15:58:31
beach
For one thing, I have absolutely zero desire to program in an environment where things may crash due to some wrong declaration.
15:58:58
nyef
I could commit to an FFI based UI toolkit. I *absolutely* could. But they all suck in some way or another.
15:59:36
nyef
At the most basic, they are often fragile in terms of simply not breaking when there's a new version of whatever library they bind to.
16:00:11
nyef
Next level up, the documentation is almost invariably "here's how the mapping to C works, go use the C documentation".
16:01:38
beach
I think I know how to do it so that it is acceptable. But it is very tedious to do it right, and time is always lacking.
16:02:49
beach
I personally think that a few selected people could get SICL and a very good development environment going in a few years.
16:03:04
nyef
On the technical level, there's basically no agreement on implementation technology. And the existing options largely suck. On the project level, it's almost invariably an afterthought, and cut due to time pressure. And then writing good documentation is a *skillset*.
16:04:31
beach
I am also already putting in a modest amount of money by paying Bike 500€ per month for 6 months. He determines the percentage that this is worth.
16:05:05
beach
This is good because drmeister is also paying him some, and Cleavir work can be done by working on Clasp.
16:05:50
beach
I paid jackdaniel to maintain McCLIM for a while, and he was so grateful that he now does it for free. :)
16:06:28
beach
nyef: So now we have three selected people, but not full time. I think Bike and jackdaniel are both good choices for their respective tasks.
16:06:35
jackdaniel
yeah, on April I'll have more time for it, I had a lot of things happening on March
16:06:40
nyef
There's also this attitude, with very vocal proponents, that in order to use Lisp, you must use Emacs and SLIME, and that's spectacularly unhelpful.
16:09:41
beach
nyef: I am also working very hard to make it more accessible to improve things further. What I do is not enough, I know. But I think the current infrastructure makes it way too hard for people to dive in and contribute.
16:10:22
nyef
Mmm. And it's not necessarily the technical infrastructure that's the barrier, but the social infrastructure.
16:11:45
nyef
Yes, both. But the technical infrastructure is a mere matter of software. The social infrastructure requires social engineering.
16:11:58
jackdaniel
if you don't know how to build software (no entry point), then you may get discouraged
16:12:13
nyef
beach: Absolutely, but it might be tomorrow afternoon/evening your time at the earliest.
16:15:43
nyef
Mmm. I'm not necessarily that experienced in working in a group either, but I try to give the social aspects some consideration.
16:18:19
nyef
A large aspect of social engineering seems to be making it easier to do the right thing than to do the wrong thing.