It has good points, but has 2 major problems, IMHO, that we observed when I was part of a team building a RDBMS:
* It again ignore the best solution for nulls: Algebraic types. Once you have this much get solved for free
* The argument against bags:
The best idea of the paper is that a RDBMS in fact internally do different data structure and temporal representations that are not of concern to the user. Correct.
THEN it jumps to the conclusion that bags should not be presented to the user, despite the fact acknowledge it need to display them.
This is wrong, and the major reason that is always overlooked is that assume that 2 identical values are wrong to exist.
I can have "Jhon, Jhon", be 2 separate persons, not know at this time more to disambiguating, adding a Id do nothing to help me, and yet, this data is correct as is.
The language MUST allow me to deal with this. It will absurd to claim an array language, procedural, functional, imperative, declarative, etc can and a relational can't.
It say the lang is intentionally limited for purity, that true, is ideal in some cases, but that purity is what the engine/compiler should track, but not distort MY DATA for this.
And the part where the perf of having bags dismissed is easy to debunk: implement a dbms and profile.
P.D: Is true that set unlocks lots of benefits, and in some cases are ideal. But very fast once you hit the real world you see that need both, similar how btrees alone is not enough, and then there is hash-based indexes and many others.
it sucks, but reality is unfriendly to pure pure languages
P.D.2: And to clear misunderstanding, I agree with the points in the same way Rust say the borrow checker is ON by default, but I argue we CANT make a strict rule, in special because data is not ideological pure
this is an unfortunate take. btrees do not imply bags. in fact extreme bagginess can be a problem when all the keys in your page are the same. lets take this from the point of view of the user, we can certainly understand that some domain models really want to be set like, and other ones baggy.
if the base domain is sets we do get a lot of nice constraints, particularly with semantics around projections and if we want to support bags, its trivial to add a counter for the number of identical records as an additional column.
if we want to support sets in bag-land, we could imagine an extension which support 'row unique' on insertion. or we can use a 'unique results' post-filter. so there are paths, but its a lot less clear in a SQL context.
you insist that your data demands bags. I really prefer to work in sets and bagginess is always just a source of crappy errors for me. shouldn't we really support usable models for both regardless of the underlying storage?
It's unfair, but my first thought after reading the abstract was "oh great, another one".
There have been several languages claiming to fix SQL and none of them (to my knowledge, I'd love counter examples) have achieved widespread enough adoption to be named a proper successor.
Reasons I can guess:
- SQL's 50 years of entrenchment- all RDBMSs speak it! Most monitoring systems speak it! Any successor language needs a good interop story so people can use it with their existing systems.
- SQL is good enough for run of the mill day to day tasks- and these days by the time I'm lost in recursive queries or window functions, I can ask an LLM for help. Maybe a successor language can win on IDE support or other parts of the dev/agent experience
Successor languages also tend to only replace parts of SQL (usually the queries, not the insert/update stuff). I think PRQL does this (once again I'd love to be wrong). Now the dev has to learn two languages?
I guess my point is that a successor language, can't just fix the semantic issues with SQL, to be successful it also has to provide large ecosystem (and maybe political) steps up. I didn't see any of that in this abstract, which kills my excitement.
P.D: I implement parts of the SQL support for a RDBMS. I think the teams expect >5 years to fully do it.
I bet is possible to implement Rust way faster than that (ie: I mean here for both: Parsing and correctly executing, not all optimization and other concerns)
I think one part that is overlooked by people who want to replace sql, is just how popular sql is with programmers. Maybe its just me, but sql is one of my favourite languages that i deal with on a regular basis.
I think its similar to regex that way. There are deficiencies that are hard to objectively defend, but at the end of the day its something i enjoy using and gets the job done.
Can we also move on from speaking about rows as if they're vertices in the conceptual model graph, and start treating them as n-ary associations / facts?
It has good points, but has 2 major problems, IMHO, that we observed when I was part of a team building a RDBMS:
* It again ignore the best solution for nulls: Algebraic types. Once you have this much get solved for free
* The argument against bags:
The best idea of the paper is that a RDBMS in fact internally do different data structure and temporal representations that are not of concern to the user. Correct.
THEN it jumps to the conclusion that bags should not be presented to the user, despite the fact acknowledge it need to display them.
This is wrong, and the major reason that is always overlooked is that assume that 2 identical values are wrong to exist.
I can have "Jhon, Jhon", be 2 separate persons, not know at this time more to disambiguating, adding a Id do nothing to help me, and yet, this data is correct as is.
The language MUST allow me to deal with this. It will absurd to claim an array language, procedural, functional, imperative, declarative, etc can and a relational can't.
It say the lang is intentionally limited for purity, that true, is ideal in some cases, but that purity is what the engine/compiler should track, but not distort MY DATA for this.
And the part where the perf of having bags dismissed is easy to debunk: implement a dbms and profile.
P.D: Is true that set unlocks lots of benefits, and in some cases are ideal. But very fast once you hit the real world you see that need both, similar how btrees alone is not enough, and then there is hash-based indexes and many others.
it sucks, but reality is unfriendly to pure pure languages
P.D.2: And to clear misunderstanding, I agree with the points in the same way Rust say the borrow checker is ON by default, but I argue we CANT make a strict rule, in special because data is not ideological pure
this is an unfortunate take. btrees do not imply bags. in fact extreme bagginess can be a problem when all the keys in your page are the same. lets take this from the point of view of the user, we can certainly understand that some domain models really want to be set like, and other ones baggy.
if the base domain is sets we do get a lot of nice constraints, particularly with semantics around projections and if we want to support bags, its trivial to add a counter for the number of identical records as an additional column.
if we want to support sets in bag-land, we could imagine an extension which support 'row unique' on insertion. or we can use a 'unique results' post-filter. so there are paths, but its a lot less clear in a SQL context.
you insist that your data demands bags. I really prefer to work in sets and bagginess is always just a source of crappy errors for me. shouldn't we really support usable models for both regardless of the underlying storage?
It's unfair, but my first thought after reading the abstract was "oh great, another one".
There have been several languages claiming to fix SQL and none of them (to my knowledge, I'd love counter examples) have achieved widespread enough adoption to be named a proper successor.
Reasons I can guess:
- SQL's 50 years of entrenchment- all RDBMSs speak it! Most monitoring systems speak it! Any successor language needs a good interop story so people can use it with their existing systems.
- SQL is good enough for run of the mill day to day tasks- and these days by the time I'm lost in recursive queries or window functions, I can ask an LLM for help. Maybe a successor language can win on IDE support or other parts of the dev/agent experience
Successor languages also tend to only replace parts of SQL (usually the queries, not the insert/update stuff). I think PRQL does this (once again I'd love to be wrong). Now the dev has to learn two languages?
I guess my point is that a successor language, can't just fix the semantic issues with SQL, to be successful it also has to provide large ecosystem (and maybe political) steps up. I didn't see any of that in this abstract, which kills my excitement.
Here is a whole seminar series on SQL alternatives https://m.youtube.com/playlist?list=PLSE8ODhjZXjbpOIrZheFWxk...
It’s pretty interesting!
The only good reason is that SQL is the only main interface for most of the main DBs.
There is NOT other *good reason*.
P.D: But, why? Well, at first SQL was truly small. Then the cost of "just adding" something look easier than redo another lang.
TODAY, SQL is insanely bad and complex.
To the point that you can, literally, learn Rust faster than SQL.
Seriously! look for example how many reserved words you need to be aware of:
https://en.wikipedia.org/wiki/List_of_SQL_reserved_words
P.D: I implement parts of the SQL support for a RDBMS. I think the teams expect >5 years to fully do it.
I bet is possible to implement Rust way faster than that (ie: I mean here for both: Parsing and correctly executing, not all optimization and other concerns)
I think one part that is overlooked by people who want to replace sql, is just how popular sql is with programmers. Maybe its just me, but sql is one of my favourite languages that i deal with on a regular basis.
I think its similar to regex that way. There are deficiencies that are hard to objectively defend, but at the end of the day its something i enjoy using and gets the job done.
Their proposal: Rel: A Programming Language for Relational Data https://dl.acm.org/doi/10.1145/3722212.3724450
https://docs.relational.ai/build/tutorials/meet-pyrel/
Can we also move on from speaking about rows as if they're vertices in the conceptual model graph, and start treating them as n-ary associations / facts?
I’m not opposed to this but, curious, what do we gain from such reconceptualization?