First, hello!

I read the article recently, and resonated with some of it. Here's 
some things that came to mind, in no particular order. We can 
split this into different threads if anyone is actually interested 
in discussing any of these.



1. The proposed experience almost sounds like the Remarkable 
e-paper tablet.

I have an rm2, and it shares ~80% of the design ideas: 
 - e-ink screen - durable, easy-to-use construction - focused 
 solely on reading and writing; no notifications - network only 
 for syncing files

It does differ in a few ways though: 
 - touch and pen instead of keyboard - transparently / always 
 online - wireless

Still, I think the RM tablets prove that there is a market for 
limited capability computing devices intended to focus on reading 
and writing.  A keyboard-based one that is basically an offline 
Emacs-lisp machine sounds intriguing to me. In fact, I originally 
got my rm2 intending to use it like that, but had trouble using a 
keyboard with it and also liked the current features more than 
expected. That said, handwritten notes are almost "write only"; 
can't really search and store them as well as plaintext.


2. WriteOnly sounds interesting

I don't know if you've been working on it any (I haven't seen 
other references, but I didn't exactly look carefully), but I do 
think that the "where to post" and "how to format" problems can 
bog down or distract from writing. A simple offline-first 
plaintext publishing system does sound nice.


  2a. Formatting

Picking a single "format" might be a challenge. I like to use 
org-mode in Emacs, and I'm not sure I'd want to re-write 
everything in markdown etc. Maybe something like Pandoc could 
solve that problem?


  2b. Discovery

What are some thoughts you've had on how the items would be listed 
and found? Is there a central federated list of recent posts? Does 
everyone have their own "blog"? If the latter, how would a reader 
find the content? Both hierarchical topics and tags are hard to do 
well, and bring back the "where to post" problem. Something like 
full text search is probably best, especially if all of the 
content fits in a few dozen GB fully replicated on each node.



3. Store-and-forward networking, like SMTP and Usenet

Usenet already is a store-and-forward system where every node 
copies all changes from its peers, and passes them on until 
everyone gets them. I'm curious what features it doesn't support 
that you might want, and whether it's necessary to build an 
entirely new system, or just extend that one (e.g. with metadata 
and clients that understand it).

I think we'd probably end up with a simpler signed binary format 
though, instead of the mess that is MIME and base64 encoding, 
possibly based on MessagePack.



Anyway, I hope at least some of these are interesting or thought 
provoking. I'm somewhat interested in helping or at least 
discussing the ideas.

Thanks,
--
Samuel H. Christie V

On 24 oct 19 04:35, Samuel Christie wrote:
>1. The proposed experience almost sounds like the Remarkable
>e-paper tablet.
>
>I have an rm2, and it shares ~80% of the design ideas:
> - e-ink screen - durable, easy-to-use construction - focused
> solely on reading and writing; no notifications - network only
> for syncing files
>
>It does differ in a few ways though:
> - touch and pen instead of keyboard - transparently / always
> online - wireless

I have a RM1 but barely use it because I do really need a keyboard. 
That’s the most important point but people don’t use keyboard anymore 
(people don’t write).
>
>
>2. WriteOnly sounds interesting
>
>I don't know if you've been working on it any (I haven't seen
>other references, but I didn't exactly look carefully), but I do
>think that the "where to post" and "how to format" problems can
>bog down or distract from writing. A simple offline-first
>plaintext publishing system does sound nice.

I haven’t released "writeOnly" but my own blog is something which is 
more and more similar to the principle.

See:
https://sr.ht/~lioploum/ploum.net/

But it is very personal with lot of hardcoded stuff. The work to make it 
generic would be non-trivial I guess.
>
>
>  2a. Formatting
>
>Picking a single "format" might be a challenge. I like to use
>org-mode in Emacs, and I'm not sure I'd want to re-write
>everything in markdown etc. Maybe something like Pandoc could
>solve that problem?
>
>
>  2b. Discovery
>
>What are some thoughts you've had on how the items would be listed
>and found? Is there a central federated list of recent posts? Does
>everyone have their own "blog"? If the latter, how would a reader
>find the content? Both hierarchical topics and tags are hard to do
>well, and bring back the "where to post" problem. Something like
>full text search is probably best, especially if all of the
>content fits in a few dozen GB fully replicated on each node.
>
>
>
>3. Store-and-forward networking, like SMTP and Usenet
>
>Usenet already is a store-and-forward system where every node
>copies all changes from its peers, and passes them on until
>everyone gets them. I'm curious what features it doesn't support
>that you might want, and whether it's necessary to build an
>entirely new system, or just extend that one (e.g. with metadata
>and clients that understand it).
>
>I think we'd probably end up with a simpler signed binary format
>though, instead of the mess that is MIME and base64 encoding,
>possibly based on MessagePack.

You may be really interested by Gwit:

https://sr.ht/~ivilata/gwit/
>
>
>
>Anyway, I hope at least some of these are interesting or thought
>provoking. I'm somewhat interested in helping or at least
>discussing the ideas.
>
>Thanks,
>--
>Samuel H. Christie V
>

-- 
Ploum - Lionel Dricot
Blog: https://www.ploum.net
Livres: https://ploum.net/livres.html

> I haven’t released "writeOnly" but my own blog is something which is 
> more and more similar to the principle.
>
> https://sr.ht/~lioploum/ploum.net

I don't quite follow this connection. Is any blog adhering to writeOnly if it publishes the source code for a simple "generator" that it uses to publish?

I write my posts with raw html in a text editor. Is that close to the principles? For example, you couldn't directly publish it to Gemini, but at the same time the conversion is something someone _could_ do relatively quickly. So I guess my question is, is any site that's created with minimal tools and publishing those tools adhering to these principles?

> ...is any site that's created with minimal tools and publishing those
> tools adhering to these principles?

That's a good question; I await Lionel's response. Also, I changed the
subject line for this branch of the conversation.

My take from the article was that WriteOnly had two purposes (though
things may have changed since then):


 1. Simplify writing and publishing

Primarily by making it simpler, so the author doesn't have to think
about the publishing process. Just write in a simple plaintext format,
and submit to the WriteOnly system, which takes care of everything else.
A bit vague, perhaps (hence some of my earliear questions about
discoverability and format), but I can see how the "plaintext source +
publish script" would be a first approximation. However, the current
example doesn't take any of the work out of *hosting* the posts yet.

I took this purpose to be the "value add" of WriteOnly, which helps it
achieve the second objective of motivating and enabling the
ForeverComputer platform.


 2. Work towards the ForeverComputer platform

It sounded like WriteOnly was the first step towards the proposed
typewriter-style computer, by enabling the desired workflows and
motivating its development even before the hardware exists. Later,
hardware ForeverComputers could be implemented with a streamlined focus
on supporting the WriteOnly workflow. To that end, it makes some sense
to iterate first in software (with more flexible, powerful computers)
before settling on a specific design.



Either way, I think WriteOnly should be more of a system than a pattern
to solve the hosting/publication problem and to build a community for
the ForeverComputer. The fact that many people *do* write blogs using
simple markup + publish scripts does that there is likely demand and
value for such a thing.

This brings to mind an article I read recently
(https://kristoff.it/blog/static-site-paradox/), which noted that
only experts use plaintext and self-hosting because it's unfortunately
hard, so average users are stuck in silos and commercial platforms.
Perhaps WriteOnly could help make self-publishing more accessible, and
restore more of the "classic web".


--
Samuel H. Christie V

On 24 oct 21 08:34, Kartik Agaram wrote:
>> I haven’t released "writeOnly" but my own blog is something which is
>> more and more similar to the principle.
>>
>> https://sr.ht/~lioploum/ploum.net
>
>I don't quite follow this connection. Is any blog adhering to writeOnly if it publishes the source code for a simple "generator" that it uses to publish?
>
>I write my posts with raw html in a text editor. Is that close to the principles? For example, you couldn't directly publish it to Gemini, but at the same time the conversion is something someone _could_ do relatively quickly. So I guess my question is, is any site that's created with minimal tools and publishing those tools adhering to these principles?
>

There’s no certification nor rule. We are inventing everything here. My 
own thinking about WriteOnly resulted in me writing a blog generator for 
myself. In a perfect world, this would then be generalized to allow 
anyone to use a similar generator but I’ve stopped halfway as it was 
good for me.

-- 
Ploum - Lionel Dricot
Blog: https://www.ploum.net
Livres: https://ploum.net/livres.html

On 24 oct 21 01:56, Samuel Christie wrote:
>This email failed anti-phishing checks when it was received by SimpleLogin, be careful with its content.
>More info on https://simplelogin.io/docs/getting-started/anti-phishing/
>
>------------------------------
>> ...is any site that's created with minimal tools and publishing those
>> tools adhering to these principles?
>
>That's a good question; I await Lionel's response. Also, I changed the
>subject line for this branch of the conversation.
>
>My take from the article was that WriteOnly had two purposes (though
>things may have changed since then):
>
>
> 1. Simplify writing and publishing
>
>Primarily by making it simpler, so the author doesn't have to think
>about the publishing process. Just write in a simple plaintext format,
>and submit to the WriteOnly system, which takes care of everything else.
>A bit vague, perhaps (hence some of my earliear questions about
>discoverability and format), but I can see how the "plaintext source +
>publish script" would be a first approximation. However, the current
>example doesn't take any of the work out of *hosting* the posts yet.
>
>I took this purpose to be the "value add" of WriteOnly, which helps it
>achieve the second objective of motivating and enabling the
>ForeverComputer platform.

You summarized it well. The hosting is currently provided by Sourcehut.

In a perfect world, WriteOnly would be a kind of black box with git as 
the backend and "publishing" would be akin to "git add *&&git commit&& 
git push".

But this is purely theoritical.
>
>
> 2. Work towards the ForeverComputer platform
>
>It sounded like WriteOnly was the first step towards the proposed
>typewriter-style computer, by enabling the desired workflows and
>motivating its development even before the hardware exists. Later,
>hardware ForeverComputers could be implemented with a streamlined focus
>on supporting the WriteOnly workflow. To that end, it makes some sense
>to iterate first in software (with more flexible, powerful computers)
>before settling on a specific design.

The workflow I envisionned morphed into Offpunk. I use Offpunk’s lists 
as a kind of Zettelkasten. 

With Offpunk, Vim and Git, I basically have a nerd version of WriteOnly. 

The perfect WriteOnly would be an integration of those (with also a 
support for emails, to replace Mutt). 

It is not yet perfectly clear in my mind how to do it. I’m also trying 
to code less to focus on my writer’s job.
>
>
>
>Either way, I think WriteOnly should be more of a system than a pattern
>to solve the hosting/publication problem and to build a community for
>the ForeverComputer. The fact that many people *do* write blogs using
>simple markup + publish scripts does that there is likely demand and
>value for such a thing.
>
>This brings to mind an article I read recently
>(https://kristoff.it/blog/static-site-paradox/), which noted that
>only experts use plaintext and self-hosting because it's unfortunately
>hard, so average users are stuck in silos and commercial platforms.
>Perhaps WriteOnly could help make self-publishing more accessible, and
>restore more of the "classic web".

Exactly. I think there’s something to be done but I lack the time and 
energy to focus on that right now. I will certainly come back to it 
later but if someone else want to work on it, that would be awesome.

-- 

Ploum - Lionel Dricot
Blog: https://www.ploum.net
Livres: https://ploum.net/livres.html