[Jprogramming] self effacing names??

[Henry Rich]

I agree with you 100%.

In the beginning we had Ken Iverson, who invented the whole idea of 
array programming and had a lifetime of successful experience with 
language design.  Roger was his mantle-ee.  While they were active on 
the project we deferred to them.

JE development has now fallen to me, but I don't have free rein: before 
I make any language changes I get the agreement of all the rest of 
Jsoftware.  Sometimes I take the opinion of the community.  The wishlist 
contains several examples of my brilliant suggestions that were shot 
down, mostly by Chris Burke.  In retrospect he was right every time.  
Most new ideas turn out to have flaws.

We should formalize language development and bring in more users who can 
propose ideas and shoot them down.  I would be happy to learn from the 
practice of other languages.

Henry Rich

On 9/14/2021 12:31 PM, Michal Wallace wrote:
> I don't like this law. :)
>
> It's not a problem that I have to build a broad consensus before making a
> change to the language,
> but (respectfully) it's kind of a problem that *you* don't.
>
> I mean, if any of us wanted to put in the effort, we could fork J and make
> our own versions that work however we want.
> Back when J went open source, Eric pretty much explicitly asked us not to
> do this.
>
> Honestly, I could not be happier that you took it upon yourself to
> implement {{ .. }}
> and I'm sure whatever you have planned for name:: is going to be cool and
> useful, too.
>
> I guess my problem is that these changes seem to be appearing out of the
> blue
> *without* any formal process from the community.
>
> There should be a real system in place by which these decisions get made.
>
> If you look at python or java, they have the idea of enhancement proposals
> that go through a lengthy discussion
> and yes, consensus-building process,with people submitting comments and
> potential implementations, etc.
>
> And then there's some formal decision making process at the end.
>
> For python, used to be the case that Guido van Rossum ("Benevolent Dictator
> for Life") had the
> final say, and he almost blew up the python community when he made the
> final decision on
> assignment-as-an-expression, and last I heard had kind of at least
> temporarily gone into "exile"... (??)
>
> If you or Eric are to be the final decision maker, that's a completely fine
> way of doing things,
> but there should still be a process.
>
> Like... A written out, formal process, with an explicit statement of how
> the ideation, consensus building,
> and decision making processes are meant to happen... And above all, these
> must all be something
> more than "whoever-it-is-just-uses-their-best-judgement".
>
> It is obvious that you have good judgement. That's not the question...
> This is about the day in the future when the J community is 10 or 100 or
> 1000
> times bigger than it is today.
>
> I know nobody thinks that is going to ever happen, but sometimes that's how
> the internet works.
>
> When I learned python, python was pretty much exactly where J is now: a
> wacky fringe language
> that only a tiny group of people used:
>
> http://www.paulgraham.com/pypar.html  (Paul Graham on why you should hire
> python programmers in 2004
> even though basically 0 companies in the world use python. Note that he
> also mentions J in the same category.)
>
> Anyway, sorry if I'm ranting here.
>
> This is the kind of problem that every open source project faces as they
> start to approach
> the bend in the hockey stick of growth.
>
> I don't know how far in the future that bend is, or if it'll ever happen,
> but we should think through the
> possibility and try to make our development process robust enough to handle
> whatever growth comes our way.
>
>
> On Tue, Sep 14, 2021 at 11:44 AM Henry Rich <henryhrich at gmail.com> wrote:
>
>> Possession is nine points of the law. :)
>>
>> I have big plans for name:: .  More important, I am opposed to
>> assignment in tacits because they destroy the functional-programming
>> purity of the tacit language.  You will need to build a broad consensus
>> before making such a change.
>>
>> Henry Rich
>>
>>
>>
>> On 9/14/2021 10:39 AM, Michal Wallace wrote:
>>> what's up with this new `name::` syntax?
>>>
>>> I have no problem with the concept, but I have a different suggestion for
>>> what `name::` should mean: a verb that assigns the variable with that
>> name.
>>> That seems like a much more commonly wished-for thing (assignment in
>>> trains) :
>>>
>>>
>> https://code.jsoftware.com/wiki/System/Interpreter/Requests#Support_assignment_in_tacit_expressions
>>> Could you use `name:.` instead for self-effacing names? It's even got a
>>> mnemonic... The big colon followed by the diminutive period could
>> indicate
>>> the change in status.... :)
>>> ----------------------------------------------------------------------
>>> For information about J forums see http://www.jsoftware.com/forums.htm
>>
>> --
>> This email has been checked for viruses by AVG.
>> https://www.avg.com
>>
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
>>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm


-- 
This email has been checked for viruses by AVG.
https://www.avg.com