Re: [m-users.] users Digest, Vol 125, Issue 2

--===============1810390027714401927==
Content-Type: multipart/alternative;
	boundary="Apple-Mail=_D64AAD0B-E6F2-4F49-A52D-E42AE3BB09E6"


--Apple-Mail=_D64AAD0B-E6F2-4F49-A52D-E42AE3BB09E6
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

Hi Tomas,

Feel free to reach out to me directly if you=E2=80=99d prefer not to =
fill the list with AI discussion, but here are a few notes from my =
experience.

I used Claude to build a Python-style regular expression module with a C =
PCRE2 backend. I was pleasantly surprised by Claude=E2=80=99s clauses =
(sorry). The whole process took about three hours, including some light =
testing. Claude did wander down a few blind alleys and needed occasional =
redirection, but given my rusty C skills, it still saved me close to a =
week of work. For those interested, the (currently unpolished) =
repository is here: https://github.com/testock/mercury-re

A few tips:

1. When working with Claude and Mercury, make sure it has access to the =
compiler source. This lets you point it to relevant examples it can use =
to model its output.

2. Claude doesn=E2=80=99t always interpret compiler errors correctly, so =
you=E2=80=99ll need to keep an eye out for cases where it gets =
=E2=80=9Cconfused.=E2=80=9D

Overall, I found Claude quite helpful for the more tedious parts of =
writing Mercury-C wrappers, though I haven=E2=80=99t used it for more =
substantial Mercury development. I suspect AI assistants will prove =
particularly effective for Hindley-Milner languages like Mercury, where =
strong typing and compiler guarantees reduce the need for extensive code =
review. My use case doesn=E2=80=99t benefit from that as much, since =
interfacing with C is inherently less type-safe.

Best regards,
Ted



> On Oct 26, 2025, at 02:00, [email protected] wrote:
>=20
> Send users mailing list submissions to
> 	[email protected]
>=20
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://lists.mercurylang.org/listinfo/users
> or, via email, send a message with subject or body 'help' to
> 	[email protected]
>=20
> You can reach the person managing the list at
> 	[email protected]
>=20
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of users digest..."
>=20
>=20
> Today's Topics:
>=20
>   1. (cough) "AI" (Tomas By)
>=20
>=20
> ----------------------------------------------------------------------
>=20
> Message: 1
> Date: Sat, 25 Oct 2025 22:31:47 +0200
> From: Tomas By <[email protected]>
> To: Mercury Users <[email protected]>
> Subject: [m-users.] (cough) "AI"
> Message-ID: <[email protected]>
> Content-Type: text/plain; charset=3DUS-ASCII
>=20
> Hi all,
>=20
> Just read this on Slashdot:
>=20
> | I'm a programmer who started out hesitant about AI, and at first I
> | thought all that it could do was auto-complete better.
> | Then I tried Claude Code, and it really is like having your own
> | personal junior dev assisting you're every need. Like a junior, it
> | makes mistakes, but using the *massive* amount of good code that it
> | creates, and fixing what's left, is so much faster than writing it
> | all from scratch yourself.
> =
https://slashdot.org/story/25/10/25/0324244/meet-the-people-who-dare-to-sa=
y-no-to-ai
>=20
> and wonder if anybody has tried this with Mercury?
>=20
> I suspect that (1) this person uses C[*]/Java, and (2) the usefulness
> of this "AI" stuff for Mercury will be proportionally less in a
> similar magnitude as code length for same functionality, ie a factor
> ten or so.
>=20
> Anybody has any experiences?
>=20
> /Tomas
>=20
>=20
> ------------------------------
>=20
> Subject: Digest Footer
>=20
> _______________________________________________
> users mailing list
> [email protected]
> https://lists.mercurylang.org/listinfo/users
>=20
>=20
> ------------------------------
>=20
> End of users Digest, Vol 125, Issue 2
> *************************************


--Apple-Mail=_D64AAD0B-E6F2-4F49-A52D-E42AE3BB09E6
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=utf-8

<html><head><meta http-equiv=3D"content-type" content=3D"text/html; =
charset=3Dutf-8"></head><body style=3D"overflow-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;"><div><p =
data-start=3D"220" data-end=3D"232">Hi Tomas,</p><p data-start=3D"234" =
data-end=3D"375">Feel free to reach out to me directly if you=E2=80=99d =
prefer not to fill the list with AI discussion, but here are a few notes =
from my experience.</p><p data-start=3D"377" data-end=3D"863">I used =
Claude to build a Python-style regular expression module with a C PCRE2 =
backend. I was pleasantly surprised by Claude=E2=80=99s clauses (sorry). =
The whole process took about three hours, including some light testing. =
Claude did wander down a few blind alleys and needed occasional =
redirection, but given my rusty C skills, it still saved me close to a =
week of work. For those interested, the (currently unpolished) =
repository is here: <a =
href=3D"https://github.com/testock/mercury-re">https://github.com/testock/=
mercury-re</a></p><p data-start=3D"865" data-end=3D"882"><span =
data-start=3D"865" data-end=3D"880">A few tips:</span></p><p =
data-start=3D"865" data-end=3D"882">1. When working with Claude and =
Mercury, make sure it has access to the compiler source. This lets you =
point it to relevant examples it can use to model its output.</p><p =
data-start=3D"865" data-end=3D"882">2. Claude doesn=E2=80=99t always =
interpret compiler errors correctly, so you=E2=80=99ll need to keep an =
eye out for cases where it gets =E2=80=9Cconfused.=E2=80=9D</p><p =
data-start=3D"1183" data-end=3D"1637">Overall, I found Claude quite =
helpful for the more tedious parts of writing Mercury-C wrappers, though =
I haven=E2=80=99t used it for more substantial Mercury development. I =
suspect AI assistants will prove particularly effective for =
Hindley-Milner languages like Mercury, where strong typing and compiler =
guarantees reduce the need for extensive code review. My use case =
doesn=E2=80=99t benefit from that as much, since interfacing with C is =
inherently less type-safe.</p><p data-start=3D"1639" =
data-end=3D"1660">Best regards,<br data-start=3D"1652" =
data-end=3D"1655">Ted</p></div><div><br></div><div><div><br><blockquote =
type=3D"cite"><div>On Oct 26, 2025, at 02:00, =
[email protected] wrote:</div><br =
class=3D"Apple-interchange-newline"><div><div>Send users mailing list =
submissions to<br><span class=3D"Apple-tab-span" =
style=3D"white-space:pre">	=
</span>[email protected]<br><br>To subscribe or unsubscribe =
via the World Wide Web, visit<br><span class=3D"Apple-tab-span" =
style=3D"white-space:pre">	=
</span>https://lists.mercurylang.org/listinfo/users<br>or, via email, =
send a message with subject or body 'help' to<br><span =
class=3D"Apple-tab-span" style=3D"white-space:pre">	=
</span>[email protected]<br><br>You can reach the =
person managing the list at<br><span class=3D"Apple-tab-span" =
style=3D"white-space:pre">	=
</span>[email protected]<br><br>When replying, please =
edit your Subject line so it is more specific<br>than "Re: Contents of =
users digest..."<br><br><br>Today's Topics:<br><br> &nbsp;&nbsp;1. =
(cough) "AI" (Tomas =
By)<br><br><br>-----------------------------------------------------------=
-----------<br><br>Message: 1<br>Date: Sat, 25 Oct 2025 22:31:47 =
+0200<br>From: Tomas By &lt;[email protected]&gt;<br>To: Mercury Users =
&lt;[email protected]&gt;<br>Subject: [m-users.] (cough) =
"AI"<br>Message-ID: =
&lt;[email protected]&gt;<br>Content-Type: text/plain; =
charset=3DUS-ASCII<br><br>Hi all,<br><br>Just read this on =
Slashdot:<br><br>| I'm a programmer who started out hesitant about AI, =
and at first I<br>| thought all that it could do was auto-complete =
better.<br>| Then I tried Claude Code, and it really is like having your =
own<br>| personal junior dev assisting you're every need. Like a junior, =
it<br>| makes mistakes, but using the *massive* amount of good code that =
it<br>| creates, and fixing what's left, is so much faster than writing =
it<br>| all from scratch =
yourself.<br>https://slashdot.org/story/25/10/25/0324244/meet-the-people-w=
ho-dare-to-say-no-to-ai<br><br>and wonder if anybody has tried this with =
Mercury?<br><br>I suspect that (1) this person uses C[*]/Java, and (2) =
the usefulness<br>of this "AI" stuff for Mercury will be proportionally =
less in a<br>similar magnitude as code length for same functionality, ie =
a factor<br>ten or so.<br><br>Anybody has any =
experiences?<br><br>/Tomas<br><br><br>------------------------------<br><b=
r>Subject: Digest =
Footer<br><br>_______________________________________________<br>users =
mailing =
list<br>[email protected]<br>https://lists.mercurylang.org/listi=
nfo/users<br><br><br>------------------------------<br><br>End of users =
Digest, Vol 125, Issue =
2<br>*************************************<br></div></div></blockquote></d=
iv><br></div></body></html>=

--Apple-Mail=_D64AAD0B-E6F2-4F49-A52D-E42AE3BB09E6--

--===============1810390027714401927==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KdXNlcnMgbWFp
bGluZyBsaXN0CnVzZXJzQGxpc3RzLm1lcmN1cnlsYW5nLm9yZwpodHRwczovL2xpc3RzLm1lcmN1
cnlsYW5nLm9yZy9saXN0aW5mby91c2Vycwo=

--===============1810390027714401927==--