[E-Lang] a slightly calmer Zooko attempts to clarify his meaning about prescriptivism (was: Re: Ppooko says "Dammit, stop being overly prescriptive.")

[Mark S. Miller]

At 02:37 PM Tuesday 4/3/01, zooko@zooko.com wrote:
>Okay I was pretty emotional when I wrote "Dammit, stop being overly
>prescriptive.", and I think I overshot the mark a bit.

No problem.  I'm glad you're passionate about this issue.  I endorse the 
direction you're pushing.


>In the discussion about an indexing method [...]
>"Can anyone think of a compelling argument that including this
>feature will cause *so* much damage [...]
>
>As far as I can tell, there is no argument that comes near to that conclusion
>with regard to indexing operators.
>[...]
>There *is* an argument at hand that *might* lead to that conclusion with regard
>to mutable collections -- the security argument.  I will try to address this
>issue in a separate letter.
>
>The argument about `.' as method-invocation syntax does not seem to lead toward
>that conclusion, to me.  I will try to address this issue in another separate
>letter.

I agree with the point.  There is an argument I consider strong against 
mutable collections that are polymorphic with the immutable ones: the 
security argument.  (Btw, this argument does not forbid mutables which 
aren't polymorphic with immutables, such as pure accumulators or write-only 
facets, but these wouldn't be familiar either, and so are irrelevant to your 
point.)

And there is an argument I currently consider strong against an invocation 
'.', though a usable command line completer written in E and working 
pleasantly for an E with the dot syntax would be a very powerful counter 
argument. ;)

None of the arguments made against indexing make a case that there's a 
compelling need to get rid of indexing.  And besides, early on Tyler agreed 
to add it to Initializer, so I'm not sure there's actually any real argument 
about this anyway.  Much of the recent discussion has just been an 
expression of our own curiosity with the question "How often would have I 
needed indexing, if I had a separate working iteration construct?"  It's 
worth examining such questions, and expressing surprise at the answers, as 
people have done, independent of what sides of what arguments might be 
supported by the answers to these questions.

But please maintain your vigilance on the initial familiarity issue, and 
keep pushing!  Thanks.


        Cheers,
        --MarkM