[AccessD] Access Development Future

Mon Oct 26 12:42:05 CDT 2026

Um, well:

a.  I'm retiring<g>
b.  I don't do much of anything outside of Access, which was what I was
speaking of.

Nice to hear that AI has come a bit further down the road with other
languages / products.

Jim.

-----Original Message-----
From: AccessD On Behalf Of Gustav Brock via AccessD
Sent: Monday, October 26, 2026 12:10 PM
To: Access Developers discussion and problem solving
<accessd at databaseadvisors.com>
Cc: Gustav Brock <gustav at cactus.dk>
Subject: Re: [AccessD] Access Development Future

Hi Jim

> And with AI, unless it's a small snippet of code, it doesn't always do a
good job.
> I think AI has been way over-hyped for what it can do at the present time.

You will have to think again - and fast!

I asked Claude Code to analyze the Blazor app I just wrote - or rather: I
orchestrated it, and Claude Code wrote the code - not parts of it but ALL of
it.
I'm not kidding - ALL OF IT - I am totally baffled by the capabilities of
the tool. 

The app was built using the model-first method and the Entity Framework and
an SQL Server backend. It is even localized, so it operates in both Danish
and English (translated perfectly by Claude, of course) as the client has a
few non-domestic members. It also features full API access and full swagger
documentation for it as the client has other web sites that will have to
communicate with this.

Here is a summary:
----
DmMembers (the app) comes to about 41,000 lines of hand-written code, made
up of roughly 25,000 lines of Razor and 14,000 of C#. 
That's a medium-sized business application.

Razor (pages and components): 
    164 files, 25,371 code lines
C#:
    196 files, 13,783 code lines
CSS:
    8 files, 1,579 code lines
JavaScript:
    4 files, 188 code lines
Hand-written total:
    372 files, 40,921 code lines
Resx (texts, Danish + English):
    2 files, 1,697 keys

Left out of the total:
    EF migrations: about 124,000 lines (163 files from 81 migrations).
They're generated, and mostly snapshots.
    Bootstrap and other third-party files

Functional size:

Pages: 135
Shared components: 26
Database tables (plus the Identity tables): 48
Model classes: 58
Service classes: 84
API endpoints: 5
Commits: 96, since 3 June 2026

How to read it
- Razor outweighs C#. A lot of the logic sits in pages' @code blocks rather
than in services, which is typical of Blazor Server apps.
----

The real chocking item is the date above meaning that with 96 commits, the
app was built in less than five months.
And again: the total cost for me has been EUR 18 per month. That is the best
money I've ever spent on any kind of software.

Three years ago, I said that this would be impossible in our lifetime. 
I started with Copilot but it is very chatty and often guessing and mixes up
things trying to be clever. And suddenly it stated that I had run out of
credits.

That forced me to Claude Code, and that's another ballgame; it communicates
in brief and exact terms and in any way behaves like your pro programming
partner.

After this experience I can say for sure: AI is the biggest invention in
programming history, and programming will never be same again.

So, Jim et all, go for it.

/gustav

-----Oprindelig meddelelse-----
Fra: AccessD <accessd-bounces+gustav=cactus.dk at databaseadvisors.com> På
vegne af Jim Dettman via AccessD
Sendt: 26. oktober 2026 14:50
Til: 'Access Developers discussion and problem solving'
<accessd at databaseadvisors.com>
Cc: Jim Dettman <jimdettman at verizon.net>
Emne: Re: [AccessD] Access Development Future

David,

 I'm down to one client as I am in the process of retiring.

 Most that I had moved on to web based solutions and that certainly is the
way the world is going, or using SAAS, which is usually outside the realm of
what a typical Access developer can offer (1-3 person IT shop).   I think
there is still a niche in the SMB market as there is a lack of developers
learning Access, and many of the current ones are retiring, and as you said,
not everyone can use the latest tech nor do everything on-line or in the
cloud.  Microsoft also continues to add new features to Access: modern
charts, zooming, breaking the 22" design limit, etc, which has been
surprising given the size of the Access team (they are doing a fantastic job
considering the resources they have to work with).

 However, with all that said, I don't know how long that niche will last.
COM being depreciated in 2029 will be the biggest issue if Microsoft doesn't
change their mind.  There are solutions/work arounds that are being
developed (calling API's), but life will change drastically for Access apps,
and some may not survive ("well if it's going to break that much, we might
as well re-write it in something else").   Thing is though, there's still
nothing like Access on the market, and many still live on the desktop
despite what Microsoft would like to see.   OWA for example is so lacking in
features compared to the desktop, that it's rare I find anyone using it.
It is slowly getting better, but it still has a way to go to match the
desktop version.

 The other issue with Access based apps is Cyber Insurance.  That's becoming
almost a given with any company, and few insurance companies like seeing an
Access app because of the lack of security.   Many of the applications
developed in the past (i.e. a health care clinic), would never fly in
today's time with a cyber insurance policy in place (or the company would
pay a hefty premium). 

 AI for me is a non-starter.   I think it's great when generalizing things,
but even with that, you need to be extremely careful even with that.  I've
often found things stitched together with a "yes, that can be done" when in
fact it cannot.    As far as coding?   So far I'm not impressed.   I know
there are several Access developers working with AI, and the demo's I've
seen, like adding a field to a table and a form, takes 30 minutes by the
time they go through all the gyrations.  You often hear the the term
"mis-aligned" now, which means "oops, it did something it wasn't supposed to
do.     And with AI, unless it's a small snippet of code, it doesn't always
do a good job.    I think AI has been way over-hyped for what it can do at
the present time.

FWIW,
Jim.

-----Original Message-----
From: AccessD On Behalf Of David Emerson
Sent: Sunday, October 25, 2026 5:26 PM
To: accessd at databaseadvisors.com
Subject: [AccessD] Access Development Future

Greetings Listers,

All has been quiet on the Access front.

I am interested on how many of us are still supporting Access applications?
Has AI made a major difference to existing client databases or is there
still a niche for our work in smaller businesses that don't have the funds
to take advantage of the latest technologies?

Regards

David Emerson

Dalyn Software Ltd

Adelaide, Australia

--
AccessD mailing list
AccessD at databaseadvisors.com
https://databaseadvisors.com/mailman/listinfo/accessd
Website: http://www.databaseadvisors.com

--
AccessD mailing list
AccessD at databaseadvisors.com
https://databaseadvisors.com/mailman/listinfo/accessd
Website: http://www.databaseadvisors.com
-- 
AccessD mailing list
AccessD at databaseadvisors.com
https://databaseadvisors.com/mailman/listinfo/accessd
Website: http://www.databaseadvisors.com