[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: OT: About the list behaviour
- From: "André Carregal" <carregal@...>
- Date: Sat, 29 Jan 2005 17:12:27 -0200
Yes, this list is great!
Please allow me to wildly digress over some recent observations about the list behaviour... feel free to press the "delete mail" button now, there is nothing important ahead, and I mean it. It's just an essay on the list itself.
You have been warned...
Lua is becaming widespread enough to be used by some very different audiences and those use Lua in very different ways. The list is becaming itself very heterogeneous and debates usually involve a subset of the so called Lua-Fu classes:
- The creators (those three guys everyone keeps talking about)
- Core tweakers (experienced C/C++ developers that know how to tweak the Lua Core and everything else that can be expressed in a .[c[pp]?|h] file)
- Module developers
- Application developers
- Application power users (configurators, instance programmers, whatever)
Altough the classes are usually intermixed and one can be positioned in every combinatory variation of them, the "hot threads" in the list usually concern the three first classes and presents itself as the pattern:
* Core tweakers and Module developers struggle to define where issue X should be implemented as a module or as part of the core.
* Core tweakers have a tendency to reject any addition to the core itself and recomend the use of mudules to overcome the lack of X in the core.
* Module developers have a tendency to ask for additions on the core so their modules can simply use X from the start.
The creators usually standby watching the fireworks, since the outcome tends to be one of:
(1) Issue X is proved implementable as a module. The thread dies gracefully.
(2) Issue X is not implementable as a module, but turns out to be "hot" enough to make the tweakers have a closer look at it. Some tweaker comes with a patch or a meta-mechanism. Patches have a reasonable chance of being integrated in the Core, but meta-mechanisms are what the creators crave for and are usually considered for the next version. The thread dies gracefully.
(3) Issue X is not implementable as a module and is not "hot enough" for the tweakers threshold. The thread holds for a lot longer than expected until:
(3a) everyone involved gives up and the threads dies erratically.
(3b) someone is clever enough to rephrase Issue X as something that fits in (1) or (2).
There's a corollary that says that by construction (3) is recurrent on a ever growing group of heterogeneous developers.
Enough babbling, sorry for the time and bandwith spent reading this.