Discussion of Working Group #1 on the Scheme reports ()

Aaron W. Hsu | 9 May 23:56

Pork and Politics

Does anyone else feel like this is turning into the U.S. Senate and a
series of unfortunate pork projects? 

	Aaron W. Hsu

headers

Neil Van Dyke | 10 May 01:13

Re: Pork and Politics

Aaron W. Hsu wrote at 05/09/2010 05:56 PM:
> Does anyone else feel like this is turning into the U.S. Senate and a series of unfortunate pork projects? 
>   
I don't know about that, but Scheme tends to attract the kind of users 
who develop their own Scheme interpreters and explore their own ideas 
about languages, and who are generally strong-minded and opinionated.  
These also tend to be people who, in the year 2010, believe that Gopher 
and Latin are preferred means of communication. :)  (Not to pick on 
anyone, and I count myself among the eclectically-inclined; those were 
just two things I noticed today and so were fresh in my mind.)

The eclectic individualist streaks of Scheme enthusiasts are good, in 
that we should get a diversity of ideas.  On the other hand, finding 
common ground and discipline among such diversity will be difficult.  I 
am wondering whether proposals that many people find unsatisfactory will 
effectively be railroaded through, whether proposals will be shot down 
late in the game because people previously didn't know what of the 
cacophony to focus on, whether there will be a process shakeup in the 
near future, or whether the process as it is will bring things together 
more or less smoothly.

I see the WGs as a speculative attempt to farm the broader Scheme 
community for input, in hopes of making a standard with more appeal than 
R6RS, while outsourcing the back-breaking labor and general dirtiness of 
that farming process.  Seems like a clever scheme to me, :) and we 
shouldn't be surprised by the reality of the process.

--

-- 
http://www.neilvandyke.org/

(Continue reading)

headers

Jeff Read | 10 May 01:16

Re: Pork and Politics

On Sun, May 9, 2010 at 5:56 PM, Aaron W. Hsu <arcfide <at> sacrideo.us> wrote:
> Does anyone else feel like this is turning into the U.S. Senate and a
> series of unfortunate pork projects?
>
>        Aaron W. Hsu
>
Was it my intent to write that triggered this sentiment? If so, then I'm sorry.

It does seem to take a lot of grinding to get a polished jewel...

--Jeff

headers

Aaron W. Hsu | 10 May 06:51

Re: Pork and Politics

On Sun, 2010-05-09 at 19:16 -0400, Jeff Read wrote:
> Was it my intent to write that triggered this sentiment? If so, then
> I'm sorry.

Oh no, this wasn't meant as an attack on anyone in particular, or even
an attack really. I just feel like there are so many people, myself
included, who have "pet" issues, that are each trying to have their way,
that I feel like we should be voting for HealthCare.  In actuality, I
understand why this is happening and agree that we need to be discussing
all these issues. We're not actually standardizing anything yet; we're
just discussing things. At this stage, it's fine to develop these ideas,
especially considering that attempts at foundational principles and
philosophizing about how to approach the task have proven thus far to be
pretty fruitless. 

Nonetheless, I couldn't help expressing the sentiment that it feels
similar. I know others probably don't see any similarity at all.

    Aaron W. Hsu

headers

David Rush | 10 May 12:40

Re: Pork and Politics

On 10 May 2010 05:51, Aaron W. Hsu <arcfide <at> sacrideo.us> wrote:

just discussing things. At this stage, it's fine to develop these ideas,
especially considering that attempts at foundational principles and
philosophizing about how to approach the task have proven thus far to be
pretty fruitless.

Yes. And i think there's a really deep problem at the root of it which has led me to debate whether WG1 Scheme makes any sense at all.

I expect this will annoy a lot of people, but I feel that Scheme has actually diverged pretty far from the lambda-calculus and has therefore become very difficult to talk in terms of first principles. Failing that foundation of first principles, we get drawn into a discussion of precedents and/or audience appeal. Attempting to reach a consensus in the allotted time frame under those conditions - especially when we are all trying to pursue the Clinger ideal seems problematic to say the least.

The *second* issue, is that the entire RnRS meme shifted in meaning somewhere between R4 & R5. It became a "standard" which was used to measure "compliance" rather than a "report" on what implementations have actually held as valuable common ground. This may be unavoidable as the amount of application software has grown, but it is a particular problem for a group that is trying to plan the future of a common language. Does this little committee exist to twist implementor's arms and get them to build things the way we want them?

And so I come back to the notion of testing proposals by implementation - and I think that we could do the most good by making the group into a little hothouse of language innovation. We start with an R4RS hosted meta-circular interpreter and have all the little features made available as patches to it. We write sample applications using the features. We crash features into each other and look at the by-products. And ii expect that a new consensus and understanding of what is *essentially* Scheme will arise.

But it will take a lot more time than is in the schedule.

david

GPG Public key at http://cyber-rush.org/drr/gpg-public-key.txt

headers

John Cowan | 10 May 22:33

Re: Pork and Politics

David Rush scripsit:

> I expect this will annoy a lot of people, but I feel that Scheme has
> actually diverged pretty far from the lambda-calculus and has therefore
> become very difficult to talk in terms of first principles. Failing that
> foundation of first principles, we get drawn into a discussion of precedents
> and/or audience appeal. Attempting to reach a consensus in the allotted time
> frame under those conditions - especially when we are all trying to pursue
> the Clinger ideal seems problematic to say the least.

As I've said before, there is nothing minimal, beautiful, or lambda-calculus-
related about Scheme's procedure library (section 6 of R5RS).  It's a ragbag
of stuff inherited from the Lisp tradition and renamed in the discussions
leading up to R2RS.

> The *second* issue, is that the entire RnRS meme shifted in meaning
> somewhere between R4 & R5. It became a "standard" which was used to measure
> "compliance" rather than a "report" on what implementations have actually
> held as valuable common ground. 

According to those reports, they represented the intersection of the
participants' *agreements*, not their pre-existing implementations.

> And so I come back to the notion of testing proposals by implementation -
> and I think that we could do the most good by making the group into a little
> hothouse of language innovation. We start with an R4RS hosted meta-circular
> interpreter and have all the little features made available as patches to
> it. We write sample applications using the features. We crash features into
> each other and look at the by-products. And ii expect that a new consensus
> and understanding of what is *essentially* Scheme will arise.

This WG is about what the charter says it's about. You want something
else, go do something else somewhere else.

--

-- 
At the end of the Metatarsal Age, the dinosaurs     John Cowan
abruptly vanished. The theory that a single         cowan <at> ccil.org
catastrophic event may have been responsible        http://www.ccil.org/~cowan
has been strengthened by the recent discovery of
a worldwide layer of whipped cream marking the
Creosote-Tutelary boundary.             --Science Made Stupid