Web 2.0 is all about incremental content!

don’t think it could really do
the “ghost text” effect.

It wouldn’t do the ghost text, only assist
it. There was a misunderstanding how “ghost
texts” work. Maybe you were thinking, that
the “ghost text” is part of the first response.

But usually the “ghost text” is a second response:

waiting for completion candidates to be suggested

Well you don’t use it for your primary
typing completion which is preferably fast.
The first response might give context information,
for the second request which provides a
different type of completion.

But the first response is not responsible
for any timing towards the second request.
That anyway happens in the client. And it
doesn’t hurt if the first response is
from a stupid channel.

Web 2.0 is all about incremental content!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Prolog missed the Web 2.0 Bandwagon. Unlike
web 1.0 which is static content. Web 2.0 is
all about dynamic content, including building
content incrementally.

IntelliJ just created Mellum, its open source,
their ghost texts are code snippets. So its
more like recalling typing marcros, giving
them a good guess. Not completing

partial identifiers:

Why Did JetBrains Create Mellum?
https://www.youtube.com/watch?v=7TqkvVXKxFA

LLM optimized for code-related tasks. https://huggingface.co/JetBrains/Mellum-4b-base

Mild Shock schrieb:

Web 2.0 is all about incremental content!

don’t think it could really do
the “ghost text” effect.

It wouldn’t do the ghost text, only assist
it. There was a misunderstanding how “ghost
texts” work. Maybe you were thinking, that
the “ghost text” is part of the first response.

But usually the “ghost text” is a second response:

waiting for completion candidates to be suggested

Well you don’t use it for your primary
typing completion which is preferably fast.
The first response might give context information,
for the second request which provides a
different type of completion.

But the first response is not responsible
for any timing towards the second request.
That anyway happens in the client. And it
doesn’t hurt if the first response is
from a stupid channel.

Web 2.0 is all about incremental content!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Disclaimer: Didn’t double check whether
there were already some LLM posts on SWI-Prolog
discourse, that would approach a solution.

The LSP and the Relay could access the same
end-user code repository. Not all data would
need a round trip through the client to go

from first request to second request.

Mild Shock schrieb:

Prolog missed the Web 2.0 Bandwagon. Unlike
web 1.0 which is static content. Web 2.0 is
all about dynamic content, including building
content incrementally.

IntelliJ just created Mellum, its open source,
their ghost texts are code snippets. So its
more like recalling typing marcros, giving
them a good guess. Not completing

partial identifiers:

Why Did JetBrains Create Mellum?
https://www.youtube.com/watch?v=7TqkvVXKxFA

LLM optimized for code-related tasks. https://huggingface.co/JetBrains/Mellum-4b-base

Mild Shock schrieb:

Web 2.0 is all about incremental content!

don’t think it could really do
the “ghost text” effect.

It wouldn’t do the ghost text, only assist
it. There was a misunderstanding how “ghost
texts” work. Maybe you were thinking, that
the “ghost text” is part of the first response.

But usually the “ghost text” is a second response:

waiting for completion candidates to be suggested

Well you don’t use it for your primary
typing completion which is preferably fast.
The first response might give context information,
for the second request which provides a
different type of completion.

But the first response is not responsible
for any timing towards the second request.
That anyway happens in the client. And it
doesn’t hurt if the first response is
from a stupid channel.

Web 2.0 is all about incremental content!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

A flesh an bood cooperative multitasking Prolog system
is sometimes tricky to do. We were agonizing over the
last days how we could test our timers and tasks.

Our existing framework doesn't work, since it neither
waits for a timer callback to be fired and to complete,
nor for a task to complete. But its seems its just an
instance of a Promise again.

Turn the test case itself into a Promise, and wait for
it. In Prolo terms, the test case is a success when the
.then() port gets reached with SUCCESS, or its a failure
if the .then() port gets reached with FAILURE or if the

the .catch() port gets reached. Interesting framework
that does just that:, whereby the use assert, to turn
FAILURE into an exception:

Node.js v20.0.0 - The test runner is now stable. https://nodejs.org/api/test.html#describe-and-it-aliases

BTW: Quite inventive vocabulary...

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Again JavaScript shines since the keyword "async"
makes the difference. We have recently experienced
its benefit, since we could remove all new Promise()
calls in our code where we are juggling with tasks.

new Promise() is only needed for callbacks that
then call resolve() or reject(), but task can
just use await and try catch. Now without
the keyword its a traditional test case:

test('synchronous failing test', (t) => {
// This test fails because it throws an exception.
assert.strictEqual(1, 2);
});

With the keyword its a test case that
can test timers and tasks:

test('asynchronous passing test', async (t) => {
// This test passes because the Promise returned by the async
// function is settled and not rejected.
assert.strictEqual(1, 1);
});

Mild Shock schrieb:

A flesh an bood cooperative multitasking Prolog system
is sometimes tricky to do. We were agonizing over the
last days how we could test our timers and tasks.

Our existing framework doesn't work, since it neither
waits for a timer callback to be fired and to complete,
nor for a task to complete. But its seems its just an
instance of a Promise again.

Turn the test case itself into a Promise, and wait for
it. In Prolo terms, the test case is a success when the
.then() port gets reached with SUCCESS, or its a failure
if the .then() port gets reached with FAILURE or if the

the .catch() port gets reached. Interesting framework
that does just that:, whereby the use assert, to turn
FAILURE into an exception:

Node.js v20.0.0 - The test runner is now stable. https://nodejs.org/api/test.html#describe-and-it-aliases

BTW: Quite inventive vocabulary...

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi,

Now SWI-Prolog has amassed 1/4 Million of
student notebooks, the SWI-Prolog discourse
has become a cest pool of stupid teachers

asking stupid questions. Development and
innovation in Prolog has totally stalled.
All Prolog systems are based on completely

silly WAM or ZIP, and cannot run this trivial
constant caching test case in linear time:

data(1,[0,1,2,3,4,5,6,7,8,9]). data(2,[0,1,2,3,4,5,6,7,8,9,0,1,2,3,4,5,6,7,8,9]). data(3,[0,1,2,3,4,5,6,7,8,9,0,1,2,3,4,5,6,7,8,9,0,1,2,3,4,5,6,7,8,9]).

test(N) :- between(1,1000000,_), data(N, _), fail; true.

Here some results:

/* Trealla Prolog 2.74.10 */

?- between(1,3,N), time(test(N)), fail; true.
% Time elapsed 0.236s, 3000004 Inferences, 12.692 MLips
% Time elapsed 0.318s, 3000004 Inferences, 9.429 MLips
% Time elapsed 0.371s, 3000004 Inferences, 8.095 MLips

/* Scryer Prolog 0.9.4-411 */

?- between(1,3,N), time(test(N)), fail; true.
% CPU time: 0.793s, 7_000_100 inferences
% CPU time: 1.150s, 7_000_100 inferences
% CPU time: 1.481s, 7_000_100 inferences

Guess what formerly Jekejeke Prolog and Dogelog
Player show? They are not based on WAM or ZIP.
Its rather DAM, Dogelog Abtract Machine.

Bye

Mild Shock schrieb:

Web 2.0 is all about incremental content!

don’t think it could really do
the “ghost text” effect.

It wouldn’t do the ghost text, only assist
it. There was a misunderstanding how “ghost
texts” work. Maybe you were thinking, that
the “ghost text” is part of the first response.

But usually the “ghost text” is a second response:

waiting for completion candidates to be suggested

Well you don’t use it for your primary
typing completion which is preferably fast.
The first response might give context information,
for the second request which provides a
different type of completion.

But the first response is not responsible
for any timing towards the second request.
That anyway happens in the client. And it
doesn’t hurt if the first response is
from a stupid channel.

Web 2.0 is all about incremental content!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi,

You see more and more frameworks try to get
flesh and blood async. I tell you, its a neglected
subject so far. There is a for example a whole

world of async streams already integrated
into the browser and node.js !

Für asynchrone API-Kommunikation mit unterschiedlichen
Transportprotokollen ist AsyncAPI als Beschreibungsstandard
entstanden, der sich am OpenAPI-Konzept anlehnt https://de.wikipedia.org/wiki/OpenAPI

Woa! This looks swagger:

Bringing Asynchronous APIs to the Forefront at APIDays Singapore
event, themed "Where APIs Meet AI: Building Tomorrow's Intelligent
Ecosystems", providing an excellent platform to introduce AsyncAPI to
the vibrant Asia-Pacific community.​ https://www.asyncapi.com/blog/2025-singapore-conf-summary

Bye


Mild Shock schrieb:

Again JavaScript shines since the keyword "async"
makes the difference. We have recently experienced
its benefit, since we could remove all new Promise()
calls in our code where we are juggling with tasks.

new Promise() is only needed for callbacks that
then call resolve() or reject(), but task can
just use await and try catch. Now without
the keyword its a traditional test case:

test('synchronous failing test', (t) => {
  // This test fails because it throws an exception.
  assert.strictEqual(1, 2);
});

With the keyword its a test case that
can test timers and tasks:

test('asynchronous passing test', async (t) => {
  // This test passes because the Promise returned by the async
  // function is settled and not rejected.
  assert.strictEqual(1, 1);
});

Mild Shock schrieb:

A flesh an bood cooperative multitasking Prolog system
is sometimes tricky to do. We were agonizing over the
last days how we could test our timers and tasks.

Our existing framework doesn't work, since it neither
waits for a timer callback to be fired and to complete,
nor for a task to complete. But its seems its just an
instance of a Promise again.

Turn the test case itself into a Promise, and wait for
it. In Prolo terms, the test case is a success when the
.then() port gets reached with SUCCESS, or its a failure
if the .then() port gets reached with FAILURE or if the

the .catch() port gets reached. Interesting framework
that does just that:, whereby the use assert, to turn
FAILURE into an exception:

Node.js v20.0.0 - The test runner is now stable.
https://nodejs.org/api/test.html#describe-and-it-aliases

BTW: Quite inventive vocabulary...

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi,

Why only phrase_from_file/2 and not also
phrase_from_url/2. Its not that difficult to
do, you can do it with change_arg/2 and

nothing else! Lets see what we have so far:

Trealla Prolog:
Base on memory mapping chars. So basically
this could be judged as a further argument
in favor of chars versus codes. But its
not Web 2.0, works only for files.

https://github.com/trealla-prolog/trealla/blob/main/library/pio.pl

SWI-Prolog:
Based on turning a stream into a lazy list.
Requires attributed variables and repositionable
streams. The stream is opened with open/3 but
maybe could be opened with http_open/3 as well?

https://github.com/SWI-Prolog/swipl-devel/blob/master/library/pio.pl

To be continued...

Bye

Mild Shock schrieb:

Web 2.0 is all about incremental content!

don’t think it could really do
the “ghost text” effect.

It wouldn’t do the ghost text, only assist
it. There was a misunderstanding how “ghost
texts” work. Maybe you were thinking, that
the “ghost text” is part of the first response.

But usually the “ghost text” is a second response:

waiting for completion candidates to be suggested

Well you don’t use it for your primary
typing completion which is preferably fast.
The first response might give context information,
for the second request which provides a
different type of completion.

But the first response is not responsible
for any timing towards the second request.
That anyway happens in the client. And it
doesn’t hurt if the first response is
from a stupid channel.

Web 2.0 is all about incremental content!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)