If you look at the generated HTML then it has a name=“something” attribute right? Maybe even an id=“someid” right?

So you can ‘click’ the button using

p_web.script('$("#someid").click()')

Or

p_web.script('$("[name=''somename'']).click()")

Also in yesterday wed webinar Roberto showed a new tool (written by ai) which works inside the dict editor. Really impressive.

One thing the prompts said was to get Claude to create and update a context.md file. At the end of each phase Claude creates a summary of what it did, what files it generated and what changes it made to the original plan and adds them to context.md.

It also said to exit Claude between phases, to avoid using memory, which just chewed up resources. And also to break large phases into smaller sections with individual prompts, rather than trying to complete the entire phase in one pass.

It all seems to be good advice?

My main ERP system (naviteer.com) has 22 years of my development time invested in it and other than having A.I. write small chunks of code, diagnose existing chunks of code when attempting to chase down a complex bug, it can only offer so much.

I don’t mind because it forces me to still write code, which I still enjoy after 42 years.

And I enjoy building with A.I. because I have always preferred thinking like a software architect and getting A.I. to do the grunt work is perfectly fine with me! It could still be much faster though and UI design skills can be crappy at times. It will only get better.

Greg

urayoan:

2. Exporting the Expanded Generated Code
   Clarion has the option to view/export the complete generated source code (the expanded .clw file without template references). From the IDE, you can use Compile → Generate Source or something similar. This output is much more readable for me because the template dependencies are already resolved.

Claude is correct here. As a Clarion developer you should have a better idea how to Generate and/or Build in Batch all the applications in a solution.

I may just put it out there in source format for a small fee and each company/developer can do what they want with it. All it requires is Visual Studio to build it.

I did add a Power User feature that toggles between two toolbars.

Basic query and report mode:

Power User:

The FULL mode option is just an INI file switch.

If off, then read-only queries else any script can be run.

Hotkeys:

Greg

https://claude.ai/public/artifacts/30a80937-a5a2-4720-9954-256cc8c572b0

I have a NT button and i want to execute the move() javascript function when the button is pressed.
I have tried multiple permutations and this is my current permutation for the button embed. However, it is not correct as the function does not execute.
This is the current iteration:
p_web.script(‘$(’‘[“move()”]’‘).call();’)
Can someone help me with the correct syntax to call this function from a NT button?
THanks,
Ron

This is almost like a story but hopefully can reflect a case use and a comparison point to others here

Translation
2. Exporting the Expanded Generated Code
Clarion has the option to view/export the complete generated source code (the expanded .clw file without template references). From the IDE, you can use Compile → Generate Source or something similar. This output is much more readable for me because the template dependencies are already resolved.

I’m not a Clarion Guru but I never heard of Expanded Generated Code in Clarion, so I asked:
Translation
It’s not for migration; I’d like to have the ability to review Clarion code and make it more robust with the help of AI. It’s a topic of discussion in the forums because, unlike Java, Swift, or Kotlin, for example, Clarion’s code isn’t stored in text files, which are more readable with easier-to-resolve references. In your previous reply, you mentioned MAP, PROCEDURE, etc. I don’t recall Clarion having a function to export the generated code in an expanded form. Could you explain what you mean by that?

To that, Claude answer was this
Translation
You’re right to correct me—I was imprecise. Let me be honest: What I said about “Generate Source” / “Export expanded” probably doesn’t exist as such in Clarion. I mixed up concepts from other environments. The .app files are the IDE’s binary project, and the .clw file is what Clarion generates internally when compiling, but there isn’t a clean, “fully expanded” export that I know for sure. If something like that exists in recent versions, I haven’t been able to confirm it.

So I asked Claude the following (look at the Added Memory)
Translation
From now on, I want you to be honest and not try to convince me with information you don’t know. If you don’t have all the information, just tell me what you don’t know and clearly let me know you need more information or context. You’re free to search and cite references, but you understand that if the reference you’re using isn’t reliable, please let me know. I need you to be precise in your answers, not make random assumptions. Likewise, if you need to correct me, do so. That’s why I use AI—because I know you have a better ability to remember things, even faster than I do to some extent. Keep that in mind, regardless of the language I use.

urayoan:

Another thing I’ve noticed is that almost all AI programs have memory, and that memory can interfere with new tasks. The same memory that should help remember patterns and the user’s workflow can replicate errors if they’re stored as memory. It’s a powerful configuration that, in the long run, can become a nightmare if it’s not reviewed periodically.

That is a very good point.

Memory can be helpful, but it can also become a quiet source of repeated mistakes if bad assumptions or bad patterns get stored in it. Then instead of helping, it starts biasing every new session in the wrong direction before the work even begins.

So yes, memory is useful, but only if it is treated as something that occasionally needs to be reviewed and cleaned out, not as something that should be trusted blindly.

Just imagine,AI got a memory from a bad instruction or bad code? Every new chat will become a mistake just to start.

For most clients, is easy to check what they have in memory and edit it to the user linking.

For example in Claude

In Gemini

ChatGPT

jslarve:

Another thing, is to make sure you know what state you’re in before you start changing code.

Yes, that is another really important part of using AI well.

You need to know what state you are in before you let it start “helping.” If you do not have a clean branch, a checkpoint, or some easy way to roll back, then you are not really experimenting. You are gambling.

And I agree completely about knowing when to stop. Sometimes the smartest move is not to keep patching a bad AI session, but to throw it away, tighten the prompt, and start fresh from a known good state.

urayoan:

I do the same to make sure and it works fairly good.

That makes a lot of sense, and I think that is very close to the most practical way to use AI right now.

Small tasks, clear boundaries, protected areas it is not allowed to touch, and a checkpoint before broader changes. That is not being overly cautious. That is just good development discipline.

I also like your point about telling it to stop and ask when it is unsure. That is a much better pattern than letting it confidently wander off into changes you did not ask for.

This works as expected EXCEPT when that procedure is closed and I re-select the log button, the Select embed doesn’t run a second time. (I have to select some other button/procedure then when I reselect the log button, it seems to have reset itself?)

I moved my code to the Accept embed point, and that works without having to select a different button/procedure. When I close the log procedure and re-press the log button, it opens the appropriate procedure that was “de-activated” using the Select embed.

HOWEVER, using the Accept embed point, bypasses the Browse “run only one” code. Pressing it three times opens 3 copies of the procedure.

My work around is to leave the button embed points empty and have the button call a “Source Procedure” with the same code I had in the embed points.

Any info/ideas on why the Accept embed bypasses the run only one code in the Browse Procedure?

1. Read-only, i.e. limiting the ability to change or delete data, is only one of the two main reasons why the powers that be don’t want ad hoc tools. The other one is confidentiality. They don’t want vast chunks of data being copied out into Excel workbooks. Also, a lot of places will solve the read-only problem by making periodic copies of the production data that their developers can test stuff on, and that power users can run their queries on.
2. Read-only has disadvantages. I’m probably on the powerful end of power users, but I create views and procedures in my database (where the weekly dump of production data is still read-only for me), and not being able to do that would be a big limitation.
3. Occasionally, when people are interested in trying to do their own stuff I will help them set up SQL Developer (the free Oracle version of something like SMS), but I wouldn’t consider spending money on them. If they actually got past the point of tinkering (I’ve never had anyone past that stage) they’d probably want to keep on with the more capable version they were already used to
4. The part of our government (OCIO) that actually looks after computer services likes having me around because otherwise they would have to deal with the many requests that come my way, but I’m a part of the furniture. Putting a tool on everyone’s desk that might just end up involving more time dealing with requests from wannabe power users may not be all that attractive.
5. There are tools out there (BI) that position themselves as being easy for end-users to learn. They normally involve setting up a sort of data warehouse scheme and don’t require the users to learn SQL. One of the systems I have been involved in set that up invested in that. I was the only person who ever created new reports. Now if they need new stuff they go to the OCIO expert to set them up for them. One of the things that pushed them in the BI direction was that the vendor of the original system was charging them in the order of 2-3K for ad hoc reports, but I’m not sure the BI investment has broken even yet.
6. If your target is people sort of unfamiliar with SQL, one thing you might need to focus on is error messages. The free Oracle version gives you very verbose messages about what went wrong and how you might fix it, which I find sort of annoying, but learners might not.
7. The people with the money have probably run into people who are beguiling them with the power of AI, and how it will give users data at their fingertips without the hassle of learning SQL. Convincing them that the way of the past is still probably the way of the near future may be a tough sell.
8. I do actually have a paid SMS-type product in addition to the free Oracle one. That’s because a few years ago the Minecraft exploit got them all worried about the security of the Java/Apache based free Oracle product. Oracle fixed their product within a few days, but three months later the OCIO insisted we stop using it and got a site license for the commercial product. They haven’t maintained it. I happily use both – they have different strengths.
   Overall, not saying that there is no one out there who would use your porduct, but I think it’s a very small niche.

”Do I have a clean branch so that I can back out of this experiment?”

If I find that the session is going so poorly that it’s taking more work to get where I want than it would be to start clean, I just start clean. Discard that sub-par work and start over. Sometimes it’s worth fixing, sometimes not. If not, at least you know how to tune your prompt with extra cursing next time.

One thing that help me with good results is, give IA small tasks and it will perform better. When I know a change will impact my whole project, I make sure I have a git check point or a backup, just if things go messy. Then i tell de IA (like a boss ) “do X task keeping the current functionality without edit Y code” or better yet, ask if not sure about the changes and let me decide. I’ve seen IA sometimes asking like a multiple choice to take a path with a change.

kenns:

I have been looking to alter the color and font of the message to make it more outstanding, more attention attracting.
I have been looking at PROP:COLOR and PROP:FONT, but cannot determine where to place the statements, if this is indeed the way to address this. Any suggestions?

You cannot color or change fonts with a Message, you have to use a Window:

Copy the Msg procedure incase you want it
Delete the Msg procedure
Add Msg procedure, but pick Window template
Pick an MDI Window type and make it CENTER
Note: an MDI window MUST be Started in a new thread or you will get an error.

Add a local String variable Loc:Wip and Window String control for it.
Loc:Wip='The '&Clip(ProcName) & ' is a Work In Progress.'

For the MoreMsg same, but could be a TEXT that is READONLY,FLAT and 3 lines tall so Height 30 to allow a long bit of text. If this was blank you could have some standard text so you don’t have to keep typing it.

IF MoreMsg THEN
   Loc:Msg = MoreMsg
ELSE 
   Loc:Msg = 'Please Check back soon...
END

Prototype and Parameters can be the same with both Type and Label.

(string PROCNAME, string MOREMSG)

Exception is Parameters on a function cannot have the return type, only the Prototype e.g. (),Long,Proc

I tried to get 2 AI agents to write it but that was “interesting” [lol]

I have it working and just want to tidy it up a bit.

To answer your question, I was going to suggest that you just edit the window in the procedure you are using, but I see it uses a MESSAGE statement to display your message, so no, you can’t edit the font in a MESSAGE. (Unless you have Capesoft’s message replacement template - which uses a window, not a message)

The AI does not get points just for sounding like it found the problem. The real question is whether, after looking at what it gave you, the explanation and the fix actually hold up.

That is where the human still has to be the programmer in the loop, and having another AI review the first one’s work can be a very sensible extra check.

With Flutter, Swift, or Kotlin, you can often tell the AI, “build me one of these”, because those are source-centric environments. The AI can work from the actual codebase, follow the structure, and see much more of the application’s big picture.

Clarion is different.

Unless you are building a hand-coded source program outside the IDE, a lot of the real structure is not sitting there in plain source form for the AI to reason over. It lives in the dictionary, templates, embeds, generated code, and procedure context. So right away, the AI is working with a much narrower window into what the application really is.

Then you add the second issue, which is that Clarion is a niche language. The models simply have not seen the same volume of examples, patterns, discussions, and historical code that they have for mainstream languages.

So I think AI in Clarion is usually less about “build me one of these” and more about “help me do this.”

That is where it starts to shine more: narrower scope, better boundaries, local examples, and clear constraints. In other words, not broad autonomous building, but guided assistance inside a well-defined area.

Since I started using AI with coding I have used it in Clarion on tens of thousands of lines of code. Once I figured out that scope and context was everything (as well as providing an example of what “good” looked like and giving it the ground rules relative to where we were working) it shined.

So yes, I suspect your two-agent method will still be interesting in Clarion, but I would expect it to need tighter supervision and much better grounding than it does in those other languages.

jslarve:

Well, I have seen my share of “I see the problem!” lies

In my case AI often channels Johnny Nash - “I Can See Clearly Now”
(https://www.youtube.com/watch?v=b0cAWgTPiwM)

I’ve written a personal app three times now using AI. The first time I wrote it using Flutter because I wanted to see what a good Flutter app looked like. I used the technique I mentioned in another thread - got one agent to write the original code and then got another agent to write it better. Seems to have worked well.

I used the same technique to write it again using Swift on a Macbook and a third time using Kotlin for an Android phone. I know nothing about any of these languages, but AI agents DO know them so using one agent to monitor the other seems to be working?

I’m going to try the same technique today with Clarion. I suspect it will be a different story?

BS in America is AKA bullshit as well. Maybe it should be NobsQuery. LOL

More than a couple of decades ago I used to provide my power users with a free copy of WinSQL, configured with a read-only login, so they could run some management queries whenever they liked. It was a really simple tool that proved invaluable for them and also helped me learn how to write SQL queries.

I couldn’t remember the name of it today and did a search for it - there are now scores of these tools available. I’ve looked at several of them recently, for other projects and most of them are pretty crap to be honest. So to my way of thinking, WinSQL is your best competitor but you’d also be competing against quite a few others.

PS: I’m not sure how devoted you are to the name? I’m sure bsQuery has some useful meaning, but over here in oz, “bs” is the polite way of saying bullshit! I guess that could make the name memorable, but might not convey the message you intend?

Guide:

https://askgoodquestions.dev/guides/the-hidden-cost-of-magic

Field note:

https://askgoodquestions.dev/field-notes/when-working-code-isnt-enough

The idea is simple, but I think it’s important:

AI makes it easier than ever to get working code. But that doesn’t automatically mean you have something you can safely own.

The moment where things get interesting is not when the code first works. It’s when something needs to change.

That’s where lack of understanding shows up, and where the real cost starts to become visible.

I’d be interested to hear how others are handling this. Are you finding that AI-generated code is holding up well over time, or are you seeing friction when it comes to maintenance and changes?

A tool such as bsQuery can provide that missing layer. With SELECT-only access, power users can build their own ad hoc queries, exports, lightweight reports, and charts while remaining safely locked out of inserts, updates, and deletes. That gives customers more flexibility without exposing the system to the risks that come with full database tools.

For the software vendor, the benefits are meaningful. It can reduce report-request backlog, cut down on support time spent creating minor custom report variations, improve customer satisfaction, and provide a new paid add-on offering. It can also strengthen the vendor’s product positioning by showing that their system supports both everyday users and advanced analytical users.

Just as important, most end users do not have SSMS installed, and many vendors would not want them using it even if they did. A lighter, faster, simpler, branded query tool is far easier to deploy, support, and productize. That makes it a practical companion to the main application rather than a replacement for enterprise developer tools.

Viewed that way, a read-only tool like bsQuery is not just a utility for developers. It can also serve as a revenue-generating extension for software vendors who want to offer customers more reporting freedom in a controlled and supportable way.

Greg
https://datafortress.app/bsquery/bsQuery_User_Guide.html

Thank you for the thoughtful feedback. That is genuinely helpful, and I think your comments are fair.

To clarify, bsQuery is not intended to replace SSMS. SSMS is an excellent full-featured tool, and most developers already have it. My goal with bsQuery is different: a lightweight, fast-loading SQL query tool for situations where a simpler, more focused tool makes more sense alongside SSMS.

The value proposition I am aiming for is:

• fast startup and a simpler interface
• optional lockout mode where only SELECT statements are allowed
• easy query building for users who are not full-time SQL developers
• built-in export, lightweight reporting, and charting
• a practical tool for power end users who would never have SSMS installed on their PCs
• potential OEM/add-on use by vertical market software vendors who want to give customers controlled query/report access

So rather than “Why switch from SSMS?”, the better question may be, “Where does a lightweight, safer, easier-to-deploy SQL tool make sense alongside SSMS?”

Pricing will need to reflect that reality. I agree that developers are not likely to pay a high price for something in a space where SSMS is free. My thinking is to keep bsQuery affordably priced, especially for developers, while also making it attractive for software vendors who may want to deploy it to end users as a branded add-on or bundled companion tool. For software vendors, that can also create an additional revenue opportunity by giving their customers an easy way to build their own queries, exports, reports, and charts without having to install SSMS.

I am also considering a branding option so the tool could be included with, or offered as an optional add-on to, vertical market SQL Server-based products.

In addition, I am considering a source code licensing option for developers or software vendors who want to customize the product further. bsQuery is a Visual Studio C# / WinForms application, so it is built on a mainstream stack that can be extended, rebranded, simplified, or enhanced for tighter integration with a vendor’s own product line. With today’s AI-assisted development tools, that kind of customization is becoming faster and more practical than it used to be.

Also, the current page is intentionally just an early interest / preview page, not the final sales site. It is meant to give people a taste of the direction and help me gauge interest. The fuller documentation is here for anyone who wants more detail:

https://datafortress.app/bsquery/bsQuery_User_Guide.html

I also think your suggestion about a short video is a very good one. A brief demo showing the standout features, intended use cases, and real-world benefits would probably communicate the product much better than a short web page alone, and that is something I am seriously considering.

Thanks again. Feedback like this is exactly what I was hoping to get at this stage.

Greg Berthume

For me, the key questions are: what does this cost, and what additional value does it provide over SSMS? You’d also need to factor in the learning curve of adopting a new tool.

I’ve looked at your page, but I have to admit it doesn’t clearly show the value of switching - although that may be because I don’t fully understand the tool yet. If one of my technical team brought me that page, I wouldn’t feel confident approving the purchase based on it.

A short video demonstrating the standout features and real benefits would likely make it much easier to evaluate.

Thanks

Nardus

The code use was as per Carl Barnes advice:
The procedure was a source code wizard
The Prototype was: (string,string)
The Parameters were: (PROCNAME,MOREMSG)
The code entered to the processed code embed point was:
! Messages
!---------

Message('The '&Clip(ProcName) & ' option is a Work In Progress.' & |
           '||' & MoreMsg , |
           'WORK IN PROGRESS', Icon:EXCLAMATION)
Return

 - change to allow leniency where group name is given as a number.
   Some regex engines allow an alternative syntax ${1} as a numbered
   reference so that has been implemented now in VitRegex in addition to the
   normal $1 form.

 - fix bug where variable lookbehind was not constraining end position.
   The lookbehind sub-match needed a hard end position constraint passed into
   tryMatch. This required a new optional parameter on tryMatch such that the
   candidate start position *and* the required end position are both enforced.

 - lots more tests added

Ankit_Purohit:

07.03:7900 C:\KeyProj\Clarion8\Municipal\run\claelp.dll

Everything else is 8.0.9759
I don’t think 7.3 is compatible with 8.0

Tonymarshall:

create a sort key on age

Just sort on date of birth.

No need to calculate age for this as the lower the DOB the older they are.

If SQL you can get the database to sort it using a VIEW.

If TPS you can create the field and key in the table. If you have existing data and you want to update it and create the key you could use a PROCESS procedure to scan the file and update the values.

• NetTalkDllMode=>0, NetTalkLinkMode=>1
• StringTheoryLinkMode=>1, StringTheoryDllMode=>0
• jFilesLinkMode=>1, jFilesDllMode=>0
• ReflectionLM=>1, ReflectionDM=>0

I’m looking for developer feedback on a new Windows desktop tool I’ve been
building called bsQuery: www.datafortress.app/bsquery

bsQuery is a focused SQL query, data inspection, reporting, and
troubleshooting utility currently aimed at Microsoft SQL Server. It is
designed for developers, support teams, consultants, software vendors, and
trusted power users who want a faster, cleaner workspace for day-to-day
database work without the overhead of launching a full IDE every time.

I personally use bsQuery every day in my own software development projects
and for end-user support, so this is a tool being shaped by real-world daily
use rather than just theory. I simply got tired of waiting for SSMS to load
and wanted something that works the way I want it to work after 26 years of
SQL Server development.

A few of the things it includes:

• multi-tab SQL editor
• schema browser w/fast double-click multi field SELECT build
• results grid
• exports (json, xml, excel, csv, pdf, html w/formatting options)
• report generator (report query results w/chart embedding, lots of
  configuration parameters)
• charting
• Ask A.I. assistance (have it write the script)
• script/snippet support (pinned queries option)
• comparison/helper tools
• safer SELECT-first workflow options
• much more!

I’m also exploring OEM / embedded use for software vendors that want to
include it with their vertical applications.

For more details, screenshots, and a printable one-page overview, please
take a look here:

www.datafortress.app/bsquery

The page is easy to share internally with a manager, product owner, business
owner, or anyone involved in purchase approval. They can also print it or
forward the link to others who need to review the product.

At this stage, I’m mainly trying to gauge interest and hear what other
developers think:

• Would a tool like this be useful in your environment?
• What features would you want added?
• What would make it more valuable than the tools you already use?

I’m also considering two possible editions:

bsQuery Standard - SQL Server only
bsQuery Pro - All engines (PostgreSQL, MySQL, Oracle added)

I’d genuinely appreciate any feedback, criticism, feature requests, or
thoughts on positioning.

Thanks for your interest!
Greg Berthume
www.datafortress.app

Thankfully, it came in source form. I have it all over my applications.