Re: Modeler (modest) ambition

I haven=E2=80=99t done Java for a long while=E2=80=A6 and swing is probably n=
ot the best represented in LLM training data these days. But FWIW, I recentl=
y migrated some of our backend services from flask to fastapi with Claude, a=
nd while it wasn=E2=80=99t one-shot=E2=80=A6 it wasn=E2=80=99t far from it. I=
=E2=80=99ve also experimented with migrating some of our tornado-based servi=
ces to fastapi, and they worked well, as well despite the frameworks being m=
uch more distinct than flask and fastapi  A migration from swing to electron=
 would be more challenging I think, but migration to JavaFX might be possibl=
e, if you want to ditch swing. Not saying you should=E2=80=A6. Just saying=E2=
=80=A6 if you fire up a branch and give it a whirl, you might be surprised w=
ith how far it gets you.=20

Robert

Sent from my iPhone

> On Apr 26, 2026, at 6:59=E2=80=AFPM, Andrus Adamchik <[email protected]>=
 wrote:
>=20
> =EF=BB=BFI can tell one thing... While any serious Modeler work still requ=
ires effort and thinking (Claude or no Claude), fixing minor bugs that annoy=
ed us for years doesn't. They are just one-shotted from a single prompt. The=
 whole refactoring that I am currently doing only helps a human to understan=
d suggested fixes. But the agent can do them over the dirty code just as eff=
iciently :)
>=20
>=20
>> On Apr 24, 2026, at 2:08=E2=80=AFPM, John Huss <[email protected]> wrot=
e:
>>=20
>> Sounds great, I love these ideas!
>>=20
>>> On Mon, Apr 20, 2026 at 9:24=E2=80=AFAM Andrus Adamchik <aadamchik@gmail=
.com> wrote:
>>>=20
>>> So I feel like the advent of agentic coding opens up new possibilities f=
or
>>> CayenneModeler. While some want you to think you can simply prompt Claud=
e
>>> to "Rewrite CayenneModeler in JavaFX" (or Electron or JetBrains Compose
>>> Multiplatform), still with our limited resources, building and supportin=
g
>>> an entirely new thing is realistically out of reach.
>>>=20
>>> But I think we can take the current Swing app to a new level by augmenti=
ng
>>> our rusty Swing skills with AI. A few of my experiments adding this or t=
hat
>>> UI piece were mostly successful. A few idea on top of my mind:
>>>=20
>>> 1. A built-in MCP server for DB Import and CGen. This is to close the lo=
op
>>> on agentic coding, allowing to use both of these tools from an agent CLI=
.
>>> (The MCP idea was what prompted me to look into this to begin with)
>>> 2. Fix table editors UX (selections conflicting with cell editors)
>>> 3. Write unit tests
>>> 4. Modernizing L&F. Not sure how far we can get while staying in Swing,
>>> but worth a try. Looking at FlatLaf lib, which is a successor of JGoodie=
s
>>> that we already use
>>> 5. Implement dozens of small usability features (such as showing current=
ly
>>> selected Obj|DbEntity in tab view headers, etc.)
>>>=20
>>> My first pass (about 70% done) was not anything visual, but rather
>>> refactoring the existing messy code to rid it of various architectural
>>> experiments accumulated over the years (such as auto-bindings) and unify=
ing
>>> the MVC structure to be as close as possible to vanilla Swing. Ironicall=
y,
>>> Claude was not that helpful in this process. It was good old IDE
>>> refactoring, manually going through hundreds of files chasing dead code a=
nd
>>> inconsistencies.
>>>=20
>>> Anyways, just putting this on the radar.
>>>=20
>>> Andrus
>=20