Re: [GLISS] differentiating pre/post/neutral commands

[Graham Percival]

On Mon, Sep 03, 2012 at 02:39:45PM +0200, David Kastrup wrote:
> Graham Percival <address@hidden> writes:
> 
> > Why not hold the preliminary discussions on a separate list (to
> > which parser experts are encouraged *not* to read), then only
> > bring a proposal to -devel when it's ready?
> 
> Syntax discussions will not likely lead anywhere sensible without
> involving people actually having worked with parsers.  Many proposals
> can be "shot down" with "and what interpretation should we assign to x?"
> kind of questions rather than "I am the master of the parser, and you
> better will believe me that I'll never have x while I live".

Really?  In terms of the \./ pre-post-neutral idea, I found
Werner, Xavier, Nicolas, Janek, Keith, and Valentin (private
email) more convincing than you and Han-Wen.  I feel that there's
discussion leading towards improvement (currently along the
direction of verbifying music functions), rather than "no, go away
until you've read the Dragon book or finished Coursea's compiler
design course".
https://www.coursera.org/course/compilers

> Shooting bad ideas down fast means that one can cover more ground in the
> same amount of time before people run out of steam.

However, the *way* in which bad ideas are shot down matter.  Some
people shoot down ideas with a nuclear bomb; other people shoot
down ideas by trapping it in a net, lowering it to the ground,
giving it a drink of hemlock, then stroking its head as it falls
asleep.  The problem with the nuclear bomb approach is that the
explosion vaporizes all steam in the surrounding area.

> I am afraid that this might rather demotivate people to work on
> proposals.  If they take a week to write something up and it gets shot
> down basically in a minute, that's not exactly going to make them plunge
> with renewed vigor into the next week of writing up a proposal.

If we can foster an attitude of open discussion (either here, or a
separate venue), then people won't need to spend a week to write
something up.  Instead, they can dash off a quick idea (such as
the pre-post-neutral idea), get some feedback, then refine it with
10-15 minutes of work (such as the verbifying), then get more
feedback, etc.  At the moment Keith has some very good ideas about
verbs-vs-nouns in music functions, so another 15-30 minutes of
work are called for in order to make a more concrete proposal.

- Graham