Re: What's the tab? // Mobile Lua // The future of Lua

[Wolfgang Pupp <wolfgang.pupp@...>]

Stefan,

I can see that you are trying to win people over for LuaOS and your
frustration at the apparent lack of interest, but you have to
understand: We can't just magically conceive your visions, plans and
ambitions; an awesome idea in your mind won't necessarily seem awesome
for anyone else.
You *need* to convince people, and you need something *tangible* for
that. Vague ideas simply are not enough. If it is absolutely
impossible to cobble some prototype together, your concept needs to be
all the more elaborate, your base assumptions untouchable. Examples:

> Mobile Lua is the next generation of Lua.
Needs a clear, hard definition of "Mobile Lua", and even then it's
doubtful that it'd become an untouchable assumption, let alone a fact
(too many sensible-at-the-time- assumptions were already proven
wrong).

> Code mobility is the untapped superpotential of Lua.
You need to back this up with something; how would "code mobility"
affect specific use-cases, what would it be able to achieve what stock
Lua can't?

> Mobile Lua is a fundamental cornerstone of the future Lua OS (an OS based on flexibly sandboxed mobile code).
This is a nice vision, but it's just that; you can't sell people on vaporware.

After all, if some Kim Lèpal, Ph.D. had shown up on this list 10 years
ago and insisted on "making Lua faster by rewriting it from scratch
with some assembly thrown in for speed", answered any doubts about
reduced portability with "custom assembly for every major and minor
platform" and then rambled on about directly interfacing with C, maybe
C++ ("no need to compile any bindings anymore") people *would* have
ignored her most certainly [1].

To sum up:
 - "Mobile Lua" needs either a proof of concept, or at least a clear
definition, a comparison with existing systems, a list of advantages
and some assessment of advantages vs. feasibility/effort.
 - Lua OS is impossible to sell, IMHO. Just consider the effort spent
on the Linux kernel, or Webkit. Even if using Lua would cut your
development time by a factor of ten, and "learning from past mistakes"
would shave off another 20%, it would *still* be a *monumental*
effort- one would need *very* convincing arguments (or boredom :P) to
start working on something like that.

----------
Unrelated:
Can we agree on a policy of "no preaching" on this mailing list? If
"arguing" is favorably presenting your arguments, then I'd consider
"preaching" anything beyond that (i.e. redundant repetition).
After all, there is no moral compulsion to *let* people convince you
of something (even if it were "the right thing") e.g.: If Rupert [1]
doesn't want to agree to no fancy "evolution theory", but you keep
insisting, *you* are the "bad guy".

And *please*, lets stay civil at all times- this list is already,
like, the nicest place on the internet; it'd be preposterous to
disrupt the peace for no gain.

--Wolfgang

[1] Any resemblance to real persons, living or dead, is purely coincidental.