Fwd: An Open Letter

86 views
Skip to first unread message

mca

unread,
Sep 26, 2015, 4:07:48 PM9/26/15
to api-...@googlegroups.com, hyperme...@googlegroups.com, hyper...@librelist.com, Mel Conway
Thought I'd share this email from Mel Conway (remember Conway' Law?) regarding his current work.

I find the general idea very compelling and would love to see others weigh in on this either on twitter (using his hashtag) or even here on the list.

cheers.


---------- Forwarded message ----------
From: Mel Conway <conwa...@gmail.com>
Date: Fri, Sep 18, 2015 at 12:19 PM
Subject: An Open Letter
To:


To my friends, colleagues, and family:

I’ve just tweeted a link to a new paper:
“An Open Letter to the Tech Community: Humanize the Craft of Building Interactive Computer Applications”. 
I invite you to follow me on Twitter at @conways_law or http://twitter.com/conways_law
You can also find the paper at 

This five-page paper describes, in almost completely nontechnical terms, what I have learned during a half-century career of thinking mostly about one problem: simplifying the software-building process. It invites the tech community to think differently about applications so that application building can be taught in elementary schools. The paper proposes specific design principles for the new craft. 

You will read in the first paragraph:
“Over the span of human history the understanding of three important technologies has migrated from the domains of priests to the domains of elementary school teachers: arithmetic, writing, and the calendar. Now software technology must join this migration.”

Then:
“My search for a new approach has led me to thinking about building applications, not from what is particular to software, but from what is universal to humanity.” 

Let’s have a conversation at #HumanizeTheCraft . 

My best,

Mel
--
New mobile number: 978-712-4925

Stuart Charlton

unread,
Sep 26, 2015, 4:30:26 PM9/26/15
to hyperme...@googlegroups.com
Great link, Mike. 

My immediate thoughts drift to Smalltalk/Squeak.  Though perhaps in the modern world, a continuously manipulatable distributed system?

Stu

Sent from my iPhone
--
You received this message because you are subscribed to the Google Groups "Hypermedia Web" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hypermedia-we...@googlegroups.com.
To post to this group, send email to hyperme...@googlegroups.com.
Visit this group at http://groups.google.com/group/hypermedia-web.
For more options, visit https://groups.google.com/d/optout.

mca

unread,
Sep 26, 2015, 4:41:07 PM9/26/15
to hyperme...@googlegroups.com
yep -- i think you're contemplating some of the things i was thinking.

i've scanned some of the other material he has on this topic[1] and the big challenge i see is that his focus seems to be on local operations (e.g. at the machine level) -- at least that is my impression from the material. and that sounds very much like Kay's smalltalk/squeak work.

what i'd *like* to see is something that works for distributed (Fielding calls them 'Networked') systems. ones that "know" they are dealing with a large systems where parts are at a distance (time/space). i think it can be done and i think it takes some different techniques/trade-offs, etc.

i've seen Bret Victor (@worrydream) talk about this[2] but (again) at the *machine* level instead of the system level. 

bottomline -- i love that he's looking at ways to tighten the feedback loop. the larger the scale of the system the harder it is to grok and/or keep the "REPL" in your head. getting the feedback to some measure of "immediate" seems to be the right way to attack this.


Stuart Charlton

unread,
Sep 26, 2015, 5:04:40 PM9/26/15
to hyperme...@googlegroups.com
(I’m remembering our twitter debate some months back about distributed vs. networked systems ;)

re: tightening the feedback loop, there is some interesting work being done for making distributed development more like local development with Zipkin:

A lot of my previous noodling about behavior trees and hypermedia has led me to believe the way forward for a hypermedia-aware networked system programming environment is exactly a continuously interactive & manipulatable agent that maintains various “current states” vs “desired states" of various hypermedia state machines, along with a searchable library of interaction trees and data formats.    The enormity of the challenge has left me to focus more on the underpinnings of what this would require (continuous deployability & delivery, self healing, etc. basically the whole cloud native thing) than on the problem itself….

Stu

Alex Moore - Niemi

unread,
Nov 2, 2015, 6:55:17 PM11/2/15
to Hypermedia Web
Little tough to look through a twitter debate. Would you mind giving a summary of the difference? I definitely think of networked systems as distributed.

J Fernandes

unread,
Nov 15, 2015, 11:23:22 PM11/15/15
to api-...@googlegroups.com, hyperme...@googlegroups.com, hyper...@librelist.com, conwa...@gmail.com
+1 Thanks for sharing!

On 8 October 2015 at 16:48, MattM <matthewmc...@gmail.com> wrote:
Great link!  Thanks for sharing!
m@


On Wednesday, 7 October 2015 12:45:09 UTC-7, Jeremy R wrote:
It reminds me rather strongly of this keynote a little while ago... http://www.infoq.com/presentations/programming-distributed-cognition

--
You received this message because you are subscribed to the Google Groups "API Craft" group.
To unsubscribe from this group and stop receiving emails from it, send an email to api-craft+...@googlegroups.com.
Visit this group at http://groups.google.com/group/api-craft.
Reply all
Reply to author
Forward
0 new messages