ChatGPT: The Future of AI is Here!

[Isaac]

@Isaac How do you define "good requirements"?

I'll answer that question if you promise to answer the same question as soon as I do?

 

[JonXL]

I'll answer that question if you promise to answer the same question as soon as I do?

I don't know if I can answer "as soon as" but I'll try to respond in a reasonable time frame.

 

[Uncle Gizmo]

@Jon

Is your Blog written by ChatGTP?

5 Common Microsoft Access Mistakes to Avoid – Access World

www.access-programmers.co.uk

 

[Uncle Gizmo]

Interview questions for an MS Access expert:-

Can you explain the difference between a database and a spreadsheet?

How do you create a new database in Microsoft Access?

Can you describe the process of creating tables and fields in a database?

How do you create relationships between tables in a database?

Can you explain the difference between a query and a report in Microsoft Access?

How do you create a query in Microsoft Access?

Can you demonstrate how to create a basic report in Microsoft Access?

How do you import data from an external source (e.g., Excel spreadsheet) into a Microsoft Access database?

Can you describe how to create and use forms in Microsoft Access?

How do you troubleshoot errors or issues that may arise when working with a Microsoft Access database?

 

[Uncle Gizmo]

It's difficult to phrase a sensible question.

 

[Uncle Gizmo]

What do developers hate about MS Access?

Developers may have several complaints or issues with using Microsoft Access as a database management system. Some common complaints include:

Limited scalability: Access is not designed to handle large amounts of data or a high number of concurrent users, so it can struggle with performance as the database grows.

Limited security: Access does not have advanced security features, so it may not be suitable for databases that need to be highly secure or compliant with certain regulations.

Limited programming options: Access does not offer as many programming options as other database management systems, so developers may find it difficult to implement certain features or functions.

Lack of integration: Access does not integrate well with other systems and technologies, so developers may have to spend extra time and effort to get it to work with other tools and platforms.

Poor database design: Access can be prone to poor database design if the developer is not careful, which can lead to issues with data integrity and performance.

Overall, many developers may find that Access is a limited and inflexible tool that does not meet their needs, especially as the database grows in size and complexity.

 

[Isaac]

A few things come to mind when I think of requirements that make me happy:

1) I can tell the person writing them knows just enough basics about how data works, generally speaking, to phrase things in a way that satisfies both the mind of the general population as well as supplies me with what I know to begin designing - to both our understanding
2) the requirements do not depend overly on using lingo that is not known to the technical side, or could be open to interpretation, but instead references the Business Systems which are involved - again, in a way that anyone can understand who understands the basic Business Systems. The lingo used should be lingo that has a written definition itself
3) the requirements should not be contradictory in nature
4) whatever internal lingo is used, the level that is used should remain consistent throughout multiple documents, projects, and over the course of time for many reasons but among them so as to create documentation that remains synched and accurate
5) the requirements should clearly state if at some point, the writer isn't sure of something that requires discussion - in other words, they should know just enough to have a sense of when their knowledge may be incomplete enough to write it perfectly and stimulate discussion questions
6) a PM might want this, although I don't care as much: the requirements should demonstrate how they are linked to and support the business process goals/rules
7) the requirements should be written in just a sensitive way that uses great exactness with regard to the desired outcome, but stops short of touching how it should be designed, beyond the outcome desired....this one is a bit fluid, most dev jobs have allowed me considerable latitude MOSTLY in the area of FE design, which frankly, I enjoyed.

the requirements leave me with no questions in a perfect world

Edit trying to answer your question humbled me Jon, I realized that frankly I struggled to do so.
Reminds me of something a supreme court justice said (which is roundly mocked, but makes a lot of sense to most people) once many years ago about public obscenity: something like ... I can't define it, but I know it when I see it!
Which used to be true of course, when the majority of the general population was on a similar moral wave length.

In all fairness to me, however, I suspect the problem isn't that "good" requirements cannot be "defined" ... I suspect that
1) It is extremely context-sensitive; i.e., I can only define what Good means in the context of a particular role at a particular company
2) I'm just not smart enough to articulate it, but the definition could exist

 

[AccessBlaster]

The problem is the person being interviewed is likely to know more than the interviewer, especially if the questioning centers around Microsoft Access.

I suspect most businesses don't use it.

 

[Isaac]

One thing is sure. AI can easily write basic ANSI sql, and probably has been able to for quite a while.
And why not ...
Even beyond ANSI, I'm sure it can be easily trained to write it with the commercial database wrappers... T-, PL-, etc.

And once AI is built into SSMS, I imagine it can write good queries, too ... taking advantage of indexes, considering any aspect that the human writer would have been able to discover, well it can discover too.

Wait, now that I waste all this typing, I think I just described what already exists in the form of every damn add-on there is.
Redgate's tools, even SSMS's built-in QBE (yes, it actually has one, but hopefully you don't use it or even hardly know about it LOL)

Oh wait, I just unwittingly described the query execution plan optimizer thing-y. Silly me.

But dammit, someone please STOP progress BEFORE it gets to where a tech-dumb Operations Manager can convince it to write a complex, useful SQL query from nothing but the garbledy-gook that proceeds from his/her/its mouth! Then we'll all be living in cardboard boxes.

Or, we'll just do what a lot of other real programmers have done: Be the guy who CREATES the "no-code" tool, then convinces everyone else that code is obsolete. (Them not realizing that something very complex had to be coded to create their supposed no code tool, of course!)

 

[Isaac]

"Access is not designed to handle ... a high number of concurrent users...." rubbish

"Lack of integration: Access does not integrate well with other systems and technologies, so developers may have to spend extra time and effort to get it to work with other tools and platforms." lacks context

@Uncle Gizmo , was that the AI response or ... ?
Everyone please humor us and tag if it's AI - we're all enjoying this fun new toy but help me out by saying if it's AI or "you" LOL

Out of all the topics mentioned in your reply, Yes, all with some degree of truth/basis in reality, but I would say if I'm going to make a sweeping generalization--as they certainly took the liberty to--the 2 quotes I've included above are probably the two I would view as the least accurate, and mostly myths spread by people who have tried Access about as diligently as I tried my last Gym & Diet.

VBA is quite friendly to COM-level DLLs, any of which can be made in c# or vb .Net and probably other platforms too. It's just a matter of building the "bridge".

Concurrency is another BIG MYTH I hear way too much of. Now before I say this, I'll preface it by saying Sure, maybe you can get to such a high theoretical number that it makes it DIFFICULT (but still not impossible) to design for such considerations as record locking etc. Maybe you can achieve a number that makes it near impossible. Maybe.

But for the most part, all concurrency concerns are myths propagated by:
- People who are not using a Split database, (again, going back to, they haven't used to properly use Access, obviously)
or
- People who are using Access as a back-end, which, yeah, you're deliberately destroying a huge % of Access's usefulness if you're doing that IMHO.

Can't tell you how many ignorant co-workers, managers, or I.T. people I've heard tell me stuff like "I thought you can't have more than 5 people using an Access database at the same time", or rubbish like that.

I tell them: "Oh, you mean the exact same accdb Access file? Well, of course not!! How many people do you think can use the exact same instance of SSMS on the exact same machine at once? I'd say, approximately one, right?" That either gets me kicked out of the conference room or stimulates a discussion about split databases and FE vs. BE

 

[Uncle Gizmo]

help me out by saying if it's AI or "you" LOL

That was the AI...

I'm wondering if it will be useful for creating blog posts to boost your website performance. Obviously it's not getting everything quite right, but it gives you a bloody good start on a blog post....

 

[Isaac]

That was the AI...

I'm wondering if it will be useful for creating blog posts to boost your website performance. Obviously it's not getting everything quite right, but it gives you a bloody good start on a blog post....

Very true. In fact, I'm almost inspired to try it out. Start my first ever Blog. Possibly even get around the dishonesty part by disclosing, in some subtle or humorous way, that the blog is "experimental", or "about AI". Crank out some articles and see what happens.

 

[Uncle Gizmo]

Crank out some articles and see what happens.

I was sat here racking my brains trying to think of something I could blog about, then I remembered, I've got a couple of pages of blog ideas I've collected!

I can't see that plagiarism is a problem with a bot, can you?

 

[Isaac]

I can't see that plagiarism is a problem with a bot, can you?

I certainly don't see how any plagiarism-detection-software, like Turnitin, could detect it - at least not without major advancements and seemingly not without explicit partnership with the AI company tokenizing and storing its output or something.

Hell, 99% of the internet is just someone repeating what someone else said anyway. Frankly, plagiarism by using a bot would probably be an improvement over most content.

 

[JonXL]

A few things come to mind when I think of requirements that make me happy:

1) I can tell the person writing them knows just enough basics about how data works, generally speaking, to phrase things in a way that satisfies both the mind of the general population as well as supplies me with what I know to begin designing - to both our understanding
2) the requirements do not depend overly on using lingo that is not known to the technical side, or could be open to interpretation, but instead references the Business Systems which are involved - again, in a way that anyone can understand who understands the basic Business Systems. The lingo used should be lingo that has a written definition itself
3) the requirements should not be contradictory in nature
4) whatever internal lingo is used, the level that is used should remain consistent throughout multiple documents, projects, and over the course of time for many reasons but among them so as to create documentation that remains synched and accurate
5) the requirements should clearly state if at some point, the writer isn't sure of something that requires discussion - in other words, they should know just enough to have a sense of when their knowledge may be incomplete enough to write it perfectly and stimulate discussion questions
6) a PM might want this, although I don't care as much: the requirements should demonstrate how they are linked to and support the business process goals/rules
7) the requirements should be written in just a sensitive way that uses great exactness with regard to the desired outcome, but stops short of touching how it should be designed, beyond the outcome desired....this one is a bit fluid, most dev jobs have allowed me considerable latitude MOSTLY in the area of FE design, which frankly, I enjoyed.

the requirements leave me with no questions in a perfect world

Edit trying to answer your question humbled me Jon, I realized that frankly I struggled to do so.
Reminds me of something a supreme court justice said (which is roundly mocked, but makes a lot of sense to most people) once many years ago about public obscenity: something like ... I can't define it, but I know it when I see it!
Which used to be true of course, when the majority of the general population was on a similar moral wave length.

In all fairness to me, however, I suspect the problem isn't that "good" requirements cannot be "defined" ... I suspect that
1) It is extremely context-sensitive; i.e., I can only define what Good means in the context of a particular role at a particular company
2) I'm just not smart enough to articulate it, but the definition could exist

Well let's simplify a bit - a good requirement is a requirement that is useful for one thing and one thing only.

Which is why there are so many different types of requirements depending on the intended audience, as you note in your second point #1.

 

[Uncle Gizmo]

Hell, 99% of the internet is just someone repeating what someone else said anyway

If you read a Microsoft Access book published by Microsoft, and then you do a blog about something you learnt in the book, that's only plagiarism if you write word for word what you read in the book. If you put it in your own words it is no longer plagiarism, it is your interpretation of what you read.

The artificial intelligence application is working under your direction. It is like an employee or an agent. There's no plagiarism unless the AI takes the text directly from somewhere else on the internet.

Interesting idea, if the AI gathers some text which is considered to be someone else's work, what would happen?

 

[Jon]

Maybe an AI legal bot could answer the question regarding plagiarism. I believe the CEO of Open AI said something like the value of intelligence will trend to zero, because AI will provide endless amounts of it. Or in other words, cognitive jobs might end up being lower paid than say a hairdresser! Just thought after writing that sentence that no doubt there will be some sort of pudding bowl device that will go over our heads in the future and the bot will snip away frantically, putting hairdressers out of business too. Its a bots life.

 

[Uncle Gizmo]

In this insightful YouTube, Tim Corey explores the benefits and pitfalls for programmers using ChatGTP.

My takeaway:- Google has "Authority"

 

[Jon]

I watched an interview yesterday with the CEO of OpenAI. It was interesting and hugely optimistic (or should I say realistic?) about the future of AI and what is coming. Major disruption is just around the corner and people just don't know it. But he has an inside seat on what is going on.

 

[JonXL]

My brother and I were chatting about Wheel of Fortune. He commented that if you could figure out how to program putting the puzzles together, you could make your own game. I noted that the programming would be relatively easy and that the real work is coming up with the puzzles - not slotting letters into squares.

And that got me thinking about this AI. Can it be asked to generate WoF puzzles? I'm not going to set up an account to find out, but I would be curious to know whether it can generate a sensible puzzle if asked, say, to create a Wheel of Fortune puzzle for a Phrase category...