From Newsgroup: comp.databases.theory

As a result of the exchange with Dan, I have updated the Transaction Sanity doc, to be generic SQL (eg. readily MS SQL which is the most common), and removed the SG specific notes. Of course, that excludes the freeware and Oracle, which are not SQL-compliant.
To be fair, it must be mentioned that MySQL is the least SQL-non-compliant, and it does not have the vulnerability of the gamut of locking problems (such as being wide open to deadlocks and app code interfering with the internal locks). I have not examined it recently, but it certainly appears that it is possible to maintain ACID compliance in MySQL, whereas PissGres and Oracle do not provide ACID in the herd of programs ("server"), and thus it is not possible to make the app code ACID compliant.
__ https://www.softwaregems.com.au/Documents/Article/Database/Transaction/Transaction%20Sanity.pdf
Nicola
Are there any facts, in my Transaction Sanity doc, p1 and p2, that you dispute ?
Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

On 2021-08-15, Derek Ignatius Asirvadem <derek.asirvadem@gmail.com> wrote:

Are there any facts, in my Transaction Sanity doc, p1 and p2, that you dispute ?

How could I dispute facts?

Thank you for that document, and for the whole discussion around ACID
and transactions. It has made me understand how you use those terms (as
opposed to their "textbook" use), which in turn, has helped me clarify
many aspects of your critique of MVCC.

I'll gladly follow your developments on Dan's data model, if there are
any, and think how they (fail to) apply to MVCC-based systems.

Nicola

--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

Nicola

On Monday, 16 August 2021 at 19:29:04 UTC+10, Nicola wrote:

On 2021-08-15, Derek Ignatius Asirvadem wrote:

Are there any facts, in my Transaction Sanity doc, p1 and p2, that you dispute ?

How could I dispute facts?

Good to hear.

Thank you for that document, and for the whole discussion around ACID
and transactions.

You are welcome.

It has made me understand how you use those terms (as
opposed to their "textbook" use),

As you can see, I do not have private terms or re-definitions of terms, I use the definitions that have been established in the industry, in chronological order (SQL, genuine SQL Platforms of forty years). I trust you appreciate that technical terms are established in order to facilitate correct communication between people. When the academics use terms to mean different things than the established terms:
a. it is an act of dishonesty,
b. it guarantees that the communication will be laboured and confused (which is a substantial component in the volume of our discussion)
c. it is the base they use to promote some primitive and non-compliant feature as the feature (eg. anti-SQL as rCLSQLrCY; anti-ACID as rCLACIDrCY; etc)
d. their students are trained in such falsity, which breaks down when they are employed in industry

which in turn, has helped me clarify
many aspects of your critique of MVCC.

Great.
Perhaps some day, you too, will admit the utter falsity of MV-non-CC, the fantasy, and thus the insanity of perceiving an online shared active database as something, anything, that it is not, let alone a snapshot frozen in time, which is impossible. And the terrible consequences thereof.

I'll gladly follow your developments on Dan's data model, if there are
any,

Assuming you mean the data model relevant to this thread, the goal being Optimistic Locking, and showing the progress of the OLTP Transaction Template code, the GitHub Gist is here. This contains the DDL and obsolete stored proc code:
__ https://gist.github.com/DanielLoth/76d241515655e76cadddef6ed2d373aa
My submission, the latest version of code against that db, is here. It is now generic code for SQL-compliant Platforms (Sybase; DB2; MS; and Informix):
__ https://www.softwaregems.com.au/Documents/Article/Database/Transaction/Visit_Add_tr%20DA.sql
Because it is for DanrCOs db, it does have some nuances: if it is at all confusing, ie. you donrCOt clearly see the difference re Template vs nuances, let me know and I will post code for the db in the Transaction Sanity doc.
If you mean the data modelling exercise for the Shooter db, please post to that effect. I donrCOt know how far Dan is interested in taking it. That is in another GithHub rCLPull RequestrCY.

and think how they (fail to) apply to MVCC-based systems.

After having reached the status that you have, per your post, it would be interesting if you comment on that. The MV-non-CC plus the manual locking (you said rCLI love 2PLrCY) that you have to do vs the Lock Managers in real SQL platforms, that wouldnrCOt dream of allowing the user to interfere with locks.
Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.databases.theory

Nicola
In this thread, you posed questions re "Serialisation" and "Schedules", which I found very odd:
- why on earth should a developer be concerned about such things (internal operation of the server) ?
- a Schedule implies single-threaded operation (we have been fully multi-threaded since 1975, not to mention Sybase is massively so at all levels)
Could you please enlighten me,, in a few words.
I found this, it appears it is being taught at Berkeley, as rCLcomputer sciencerCY about rCLdatabasesrCY. Why ???
https://dsf.berkeley.edu/dbcourse/lecs/22cc.pdf

Cheers
Derek
--- Synchronet 3.21d-Linux NewsLink 1.2