Where do y'all hang out?

#1 · December 22, 2022, 11:27 am

Seeing as this forum is not so very active nowadays, I was wondering where you all go to for stimulating ideas, discussion and debate on relational matters?

Quora, Stack Overflow, LinkedIn, Wikipedia are some of the place I go to nowadays, but I fear I'm missing out elsewhere.

There is a Slack workspace provided by dbt that is great to keep up with the "modern data stack" and observing everyone's ongoing life with SQL, but there is no real discussion of fundamentals there.

Any other suggestions?

#2 · December 23, 2022, 10:04 am

Bit quiet, isn't it? I was hoping you'd get at least someone to answer.

For me, the attraction of this list has been the smart people on it. I'm not a data practitioner and I can do without a dose of matters relational for a while. Each of the places you mention has its role to play I guess, but not for deep discussion. I dislike Slack and have no interest in dbt. I follow a couple of AI lists and some games dev on Discord.

If you find anywhere worth a look, feel free to drop a link.

Andl - A New Database Language - andl.org

#3 · December 24, 2022, 2:55 am

Quote from AntC on December 24, 2022, 2:55 am
Bit quiet, isn't it?

I wouldn't expect a prompt response to an idle question in the few days immediately before Christmas. This is more the sort of topic to muse on in a post-prandial stupor Boxing Day afternoon. (Or a post-Christmas Eve market stupor, in my case.)

I was wondering where you all go to for stimulating ideas, discussion and debate on relational matters?

It seems to me the problem with The Relational Model (by which I don't mean SQL nor NoSQL) is it's too absurdly simple. A few crystal-clear logical concepts (a Relation is the extension of a predicate). Done.

You're not going to support a discussion forum/lecture tour/series of textbooks/etc on that. So Codd went down the rabbit-hole of a plethora of Nulls. McGoveran went down the rabbit-hole of update-through-views (and seems to be still stuck there). Date also is rather far down that same hole. D&D (not sure which) have a rabbit-hole of Type inheritance.

the relational model actually needs no extension, and no subsumption, and above all no perversion, [Date, only he doesn't seem to abide by his own slogan]

Pascal seems to have somewhat drunk the Nulls Kool-aid and somewhat the McGoveran whatever, and now seems to be engaged in dredging through ancient postings in relational theory, re-editing to tell questioners even more bluntly that they're stupid. This is also not a way to engender a discussion forum. (I notice nobody these days actually approaches him with a question -- he steals the q's from elsewhere.)

I do vaguely keep a tab on StackOverflow. But those q's are mostly from people who are unaware SQL is _not_ Relational and/or who seem unable to search for previous answers to pretty much the same question. (I guess it needs a level of understanding to recognise what is 'the same question' in that context.) Also a lot of q's from programmers who've never been trained to be dba's/model designers, and think the data model should just map to the user's screen layout somehow (insert magic here). There's also you-know-who ensuring that SO newbies to RM questions are met with a maximum level of toxicity.

there is no real discussion of fundamentals there.

I did come here a few months ago, trying to get discussion going on a concrete topic. (Concrete as in, I was aiming to build relation-like functionality into a programming language.) I rather thought it touched on fundamentals. But I didn't manage to stimulate any discussion of fundamentals. Rather, people re-iterated entrenched (and not really applicable) ideas/we ended up talking past each other. So I won't be bothering to do that again.

And sorry, I have no other suggestions. The (true) Relational Model is just going to die out/become a historical curiosity, taught briefly in Undergraduate courses by lecturers who never understood it in the first place/think it's just a point of pain on the way to (No)SQL.

Bit quiet, isn't it?

I wouldn't expect a prompt response to an idle question in the few days immediately before Christmas. This is more the sort of topic to muse on in a post-prandial stupor Boxing Day afternoon. (Or a post-Christmas Eve market stupor, in my case.)

I was wondering where you all go to for stimulating ideas, discussion and debate on relational matters?

It seems to me the problem with The Relational Model (by which I don't mean SQL nor NoSQL) is it's too absurdly simple. A few crystal-clear logical concepts (a Relation is the extension of a predicate). Done.

You're not going to support a discussion forum/lecture tour/series of textbooks/etc on that. So Codd went down the rabbit-hole of a plethora of Nulls. McGoveran went down the rabbit-hole of update-through-views (and seems to be still stuck there). Date also is rather far down that same hole. D&D (not sure which) have a rabbit-hole of Type inheritance.

the relational model actually needs no extension, and no subsumption, and above all no perversion, [Date, only he doesn't seem to abide by his own slogan]

Pascal seems to have somewhat drunk the Nulls Kool-aid and somewhat the McGoveran whatever, and now seems to be engaged in dredging through ancient postings in relational theory, re-editing to tell questioners even more bluntly that they're stupid. This is also not a way to engender a discussion forum. (I notice nobody these days actually approaches him with a question -- he steals the q's from elsewhere.)

I do vaguely keep a tab on StackOverflow. But those q's are mostly from people who are unaware SQL is _not_ Relational and/or who seem unable to search for previous answers to pretty much the same question. (I guess it needs a level of understanding to recognise what is 'the same question' in that context.) Also a lot of q's from programmers who've never been trained to be dba's/model designers, and think the data model should just map to the user's screen layout somehow (insert magic here). There's also you-know-who ensuring that SO newbies to RM questions are met with a maximum level of toxicity.

there is no real discussion of fundamentals there.

I did come here a few months ago, trying to get discussion going on a concrete topic. (Concrete as in, I was aiming to build relation-like functionality into a programming language.) I rather thought it touched on fundamentals. But I didn't manage to stimulate any discussion of fundamentals. Rather, people re-iterated entrenched (and not really applicable) ideas/we ended up talking past each other. So I won't be bothering to do that again.

And sorry, I have no other suggestions. The (true) Relational Model is just going to die out/become a historical curiosity, taught briefly in Undergraduate courses by lecturers who never understood it in the first place/think it's just a point of pain on the way to (No)SQL.

#4 · December 26, 2022, 6:27 pm

Quote from Dave Voorhis on December 26, 2022, 6:27 pm

I was wondering where you all go to for stimulating ideas, discussion and debate on relational matters?

It seems to me the problem with The Relational Model (by which I don't mean SQL nor NoSQL) is it's too absurdly simple. A few crystal-clear logical concepts (a Relation is the extension of a predicate). Done.

You're not going to support a discussion forum/lecture tour/series of textbooks/etc on that. So Codd went down the rabbit-hole of a plethora of Nulls. McGoveran went down the rabbit-hole of update-through-views (and seems to be still stuck there). Date also is rather far down that same hole. D&D (not sure which) have a rabbit-hole of Type inheritance.

the relational model actually needs no extension, and no subsumption, and above all no perversion, [Date, only he doesn't seem to abide by his own slogan]

Pascal seems to have somewhat drunk the Nulls Kool-aid and somewhat the McGoveran whatever, and now seems to be engaged in dredging through ancient postings in relational theory, re-editing to tell questioners even more bluntly that they're stupid. This is also not a way to engender a discussion forum. (I notice nobody these days actually approaches him with a question -- he steals the q's from elsewhere.)

I do vaguely keep a tab on StackOverflow. But those q's are mostly from people who are unaware SQL is _not_ Relational and/or who seem unable to search for previous answers to pretty much the same question. (I guess it needs a level of understanding to recognise what is 'the same question' in that context.) Also a lot of q's from programmers who've never been trained to be dba's/model designers, and think the data model should just map to the user's screen layout somehow (insert magic here). There's also you-know-who ensuring that SO newbies to RM questions are met with a maximum level of toxicity.

there is no real discussion of fundamentals there.

I did come here a few months ago, trying to get discussion going on a concrete topic. (Concrete as in, I was aiming to build relation-like functionality into a programming language.) I rather thought it touched on fundamentals. But I didn't manage to stimulate any discussion of fundamentals. Rather, people re-iterated entrenched (and not really applicable) ideas/we ended up talking past each other. So I won't be bothering to do that again.

And sorry, I have no other suggestions. The (true) Relational Model is just going to die out/become a historical curiosity, taught briefly in Undergraduate courses by lecturers who never understood it in the first place/think it's just a point of pain on the way to (No)SQL.

That sums it up nicely.

I've been working on a transpiler to Java which is moving along nicely, but I'm not in any hurry. My goal is to define a useful set of semantics which are sufficient to let the user define relational operators, integrated persistence, etc., within the language. I.e., what's the minimal ergonomic set of functionality a general-purpose programming language needs, not to necessarily be a D, but to write a D without being forced to step outside the language (e.g., create a standalone language processor) to work around the language limitations?

So, yes, I'm creating a standalone language processor to avoid, uh, having to create a standalone language processor.

But other than that...

I've been posting Quora answers, though most have nothing to do with databases or the relational model. I've mainly been avoiding Internet bickering in favour of photography, DIY house upgrades, and restoring a growing collection of vintage Tektronix oscilloscopes.

Here, have a pic I took of a bird.

I was wondering where you all go to for stimulating ideas, discussion and debate on relational matters?

It seems to me the problem with The Relational Model (by which I don't mean SQL nor NoSQL) is it's too absurdly simple. A few crystal-clear logical concepts (a Relation is the extension of a predicate). Done.

You're not going to support a discussion forum/lecture tour/series of textbooks/etc on that. So Codd went down the rabbit-hole of a plethora of Nulls. McGoveran went down the rabbit-hole of update-through-views (and seems to be still stuck there). Date also is rather far down that same hole. D&D (not sure which) have a rabbit-hole of Type inheritance.

the relational model actually needs no extension, and no subsumption, and above all no perversion, [Date, only he doesn't seem to abide by his own slogan]

Pascal seems to have somewhat drunk the Nulls Kool-aid and somewhat the McGoveran whatever, and now seems to be engaged in dredging through ancient postings in relational theory, re-editing to tell questioners even more bluntly that they're stupid. This is also not a way to engender a discussion forum. (I notice nobody these days actually approaches him with a question -- he steals the q's from elsewhere.)

I do vaguely keep a tab on StackOverflow. But those q's are mostly from people who are unaware SQL is _not_ Relational and/or who seem unable to search for previous answers to pretty much the same question. (I guess it needs a level of understanding to recognise what is 'the same question' in that context.) Also a lot of q's from programmers who've never been trained to be dba's/model designers, and think the data model should just map to the user's screen layout somehow (insert magic here). There's also you-know-who ensuring that SO newbies to RM questions are met with a maximum level of toxicity.

there is no real discussion of fundamentals there.

I did come here a few months ago, trying to get discussion going on a concrete topic. (Concrete as in, I was aiming to build relation-like functionality into a programming language.) I rather thought it touched on fundamentals. But I didn't manage to stimulate any discussion of fundamentals. Rather, people re-iterated entrenched (and not really applicable) ideas/we ended up talking past each other. So I won't be bothering to do that again.

And sorry, I have no other suggestions. The (true) Relational Model is just going to die out/become a historical curiosity, taught briefly in Undergraduate courses by lecturers who never understood it in the first place/think it's just a point of pain on the way to (No)SQL.

That sums it up nicely.

I've been working on a transpiler to Java which is moving along nicely, but I'm not in any hurry. My goal is to define a useful set of semantics which are sufficient to let the user define relational operators, integrated persistence, etc., within the language. I.e., what's the minimal ergonomic set of functionality a general-purpose programming language needs, not to necessarily be a D, but to write a D without being forced to step outside the language (e.g., create a standalone language processor) to work around the language limitations?

So, yes, I'm creating a standalone language processor to avoid, uh, having to create a standalone language processor.

But other than that...

I've been posting Quora answers, though most have nothing to do with databases or the relational model. I've mainly been avoiding Internet bickering in favour of photography, DIY house upgrades, and restoring a growing collection of vintage Tektronix oscilloscopes.

Here, have a pic I took of a bird.

Uploaded files:

I'm the forum administrator and lead developer of Rel. Email me at dave@armchair.mb.ca with the Subject 'TTM Forum'. Download Rel from https://reldb.org

#5 · December 27, 2022, 1:04 am

IMO TTM reached a dead-end because of internal failings and external factors.

The internal failings are:

Adoption of a type system that is inherently hostile to all major GP languages (SQL is equally hostile to being hosted)
Hostility to NULL, to the point of making SQL look good (for handling missing data, outer joins, etc)
Hostility to ordered data sets, to the point of making SQL look good (for handling reporting, aggregation, running totals, etc).

A TTM with a host-language-friendly type system and good approaches to NULL-related problems and ordered-dataset-related problems might have stood a chance. I have presented a suitable type system and ways to handle ordered datasets here, and others have presented other aspects, but too little too late I fear.

The external factors of course are obvious:

The strengths of SQL as a comprehensive server-side data-oriented language
The rise of GP-hosted data manipulation techniques such as LINQ and Java Streams.

From my own experience, there is nothing I ever need to do with data where the RM would work a whole lot better than LINQ. But I don't write business apps, so I can stay clear of SQL.

If we don't use TTM-derived products, who will?

And in this happy period between Xmas and the NY, with the temperature hovering well about 35 (C that is, of course), and in like spirit, I offer a picture of a bird taken from where I now sit.

[obviously I have no idea what I'm doing because the upload does not work at all]

Andl - A New Database Language - andl.org

#6 · December 27, 2022, 12:06 pm

Quote from Dave Voorhis on December 27, 2022, 12:06 pm

Quote from dandl on December 27, 2022, 1:04 am

IMO TTM reached a dead-end because of internal failings and external factors.

The internal failings are:

Adoption of a type system that is inherently hostile to all major GP languages (SQL is equally hostile to being hosted)

Hostility to NULL, to the point of making SQL look good (for handling missing data, outer joins, etc)

Hostility to ordered data sets, to the point of making SQL look good (for handling reporting, aggregation, running totals, etc).

A TTM with a host-language-friendly type system and good approaches to NULL-related problems and ordered-dataset-related problems might have stood a chance. I have presented a suitable type system and ways to handle ordered datasets here, and others have presented other aspects, but too little too late I fear.

The external factors of course are obvious:

The strengths of SQL as a comprehensive server-side data-oriented language

The rise of GP-hosted data manipulation techniques such as LINQ and Java Streams.

From my own experience, there is nothing I ever need to do with data where the RM would work a whole lot better than LINQ. But I don't write business apps, so I can stay clear of SQL.

I write business applications and use SQL extensively. From my point of view, the real problem is that SQL is just good enough, to the point that a hypothetical TTM alternative simply doesn't make a big enough difference to warrant the effort to learn something new, let alone adopt it. Indeed, there are innumerable examples of Tool x being notionally superior to Tool y, but this-is-what-I-know-and-it-works is a powerful force. To achieve escape velocity, you need a paradigm shift, not just incremental improvement.

Unfortunately, TTM is incremental improvement over SQL.

As I've mentioned before, I think the opportunities for TTM thinking lie well outside the database arena. SQL won that one, and significant change there probably won't happen in our lifetimes. But for manipulating clusters of n things -- nodes, agents, virtual machines, containers, whatever -- there is still plenty of interesting work to be done. I note that a lot of the heavy lifting in -- for example -- the CI/CD / DevOps world is shell scripts manipulating collections of n things. Opportunity for TTM thinking to create new tools?

Maybe.

If we don't use TTM-derived products, who will?

Well, I do. I've been using Rel for desktop data-crunching for years. I like it.

And in this happy period between Xmas and the NY, with the temperature hovering well about 35 (C that is, of course), and in like spirit, I offer a picture of a bird taken from where I now sit.

[obviously I have no idea what I'm doing because the upload does not work at all]

Huh. Now there's a real problem. You did 'Choose File' and picked a jpg, jpeg, gif, png, bmp (or pdf, or zip), yes?

Here's another of mine, from last week in Edinburgh...

Quote from dandl on December 27, 2022, 1:04 am

IMO TTM reached a dead-end because of internal failings and external factors.

The internal failings are:

Adoption of a type system that is inherently hostile to all major GP languages (SQL is equally hostile to being hosted)

Hostility to NULL, to the point of making SQL look good (for handling missing data, outer joins, etc)

Hostility to ordered data sets, to the point of making SQL look good (for handling reporting, aggregation, running totals, etc).

A TTM with a host-language-friendly type system and good approaches to NULL-related problems and ordered-dataset-related problems might have stood a chance. I have presented a suitable type system and ways to handle ordered datasets here, and others have presented other aspects, but too little too late I fear.

The external factors of course are obvious:

The strengths of SQL as a comprehensive server-side data-oriented language

The rise of GP-hosted data manipulation techniques such as LINQ and Java Streams.

From my own experience, there is nothing I ever need to do with data where the RM would work a whole lot better than LINQ. But I don't write business apps, so I can stay clear of SQL.

I write business applications and use SQL extensively. From my point of view, the real problem is that SQL is just good enough, to the point that a hypothetical TTM alternative simply doesn't make a big enough difference to warrant the effort to learn something new, let alone adopt it. Indeed, there are innumerable examples of Tool x being notionally superior to Tool y, but this-is-what-I-know-and-it-works is a powerful force. To achieve escape velocity, you need a paradigm shift, not just incremental improvement.

Unfortunately, TTM is incremental improvement over SQL.

As I've mentioned before, I think the opportunities for TTM thinking lie well outside the database arena. SQL won that one, and significant change there probably won't happen in our lifetimes. But for manipulating clusters of n things -- nodes, agents, virtual machines, containers, whatever -- there is still plenty of interesting work to be done. I note that a lot of the heavy lifting in -- for example -- the CI/CD / DevOps world is shell scripts manipulating collections of n things. Opportunity for TTM thinking to create new tools?

Maybe.

If we don't use TTM-derived products, who will?

Well, I do. I've been using Rel for desktop data-crunching for years. I like it.

And in this happy period between Xmas and the NY, with the temperature hovering well about 35 (C that is, of course), and in like spirit, I offer a picture of a bird taken from where I now sit.

[obviously I have no idea what I'm doing because the upload does not work at all]

Huh. Now there's a real problem. You did 'Choose File' and picked a jpg, jpeg, gif, png, bmp (or pdf, or zip), yes?

Here's another of mine, from last week in Edinburgh...

Uploaded files:

I'm the forum administrator and lead developer of Rel. Email me at dave@armchair.mb.ca with the Subject 'TTM Forum'. Download Rel from https://reldb.org

#7 · December 27, 2022, 12:56 pm

I don't write shell scripts either, but I do use the odd drop of PowerShell and that has much the same power as C#.

The strength of a TTM-inspired SQL substitute lies where the data is not and cannot be server-based, and clusters of things may well lie in that domain. But as I said the lack of a hostable type system and features akin to NULL and ORDER BY are a huge barrier for TTM as it stands.

I'm using Edge this time, so let's see if it works any better.

Nope. I do choose file, pick a JPG, then Submit. Is there a step I'm missing?

Andl - A New Database Language - andl.org

#8 · December 27, 2022, 2:44 pm

If Dave's trying to tempt me with his nice bird pics (robin, eagle owl), then he's succeeded. This one's a nice video taken by my trailcam, fortuitously set to take 30 seconds per shot.

Hugh

Coauthor of The Third Manifesto and related books.

#9 · December 27, 2022, 4:15 pm

Quote from dandl on December 27, 2022, 12:56 pm

Nope. I do choose file, pick a JPG, then Submit. Is there a step I'm missing?

Hmmm... Perhaps a bug or permissions issue. I'm using a user-level, non-admin test account and have installed some pending upgrades so everything should be up-to-date. I shall try to upload a picture of a duck, from yesterday.

Uploaded files:

#10 · December 27, 2022, 4:34 pm

Quote from Hugh on December 27, 2022, 2:44 pm

If Dave's trying to tempt me with his nice bird pics (robin, eagle owl), then he's succeeded. This one's a nice video taken by my trailcam, fortuitously set to take 30 seconds per shot.

Hugh

Two users with the same issue, looks like. My previous post-upgrade attempt appeared to work, but that was using Safari. This attempt is using the same standard user account and Chrome, which is also the same core engine used in Edge. Another duck from yesterday...

Uploaded files:

The Forum for Discussion about The Third Manifesto and Related Matters

Where do y'all hang out?