Disclosure: On this site you won’t find specific advice on “how to call function xyz()”. Interpreting C++ ARM and #pragma dwim is also out of scope.

We’re treating our readers as intelligent beings who can use Google and/or StackOverflow, where all such specific questions were answered more than once.

What you will find is opinions, more opinions, and even more opinions on all the aspects of software development - and with a large chunk of them based on real-world experience too.

Your mileage may vary. Batteries not included.

MECHLOVE Blueprint - 2/7. Storage (ZFS-style) and WAL (template-based logical with physical hints)

October 7, 2026 by “No Bugs” Bunny

Quote

One important feature of our WAL is that (because our Data Storage is CoW, and our checkpoint pages are perfectly consistent) it seems that we do NOT need to undo during recovery

Filed under: On.System ArchitectureDesign decisions

Tagged With: DatabaseOLTP
Read more

Radical MVCC and Replay-Based Rebasing OCC (Re2OCC)

October 5, 2026 by “No Bugs” Bunny

Radical MVCC and Replay-Based Rebasing OCC (Re2OCC) - Post preview image

Quote

Write-sets published at the point of obtaining CSNs are immutable and represent THE db history. Everything else is just views over this history.

Filed under: On.System ArchitectureDesign decisions

Tagged With: OLTPDatabase
Read more

MECHLOVE Blueprint - 1/7. General: Pretty Much Classical RDBMS at Heart, with Each and Every Component Rewritten

September 30, 2026 by “No Bugs” Bunny

MECHLOVE Blueprint - 1/7. General: Pretty Much Classical RDBMS at Heart, with Each and Every Component Rewritten - Post preview image

Detour

We interrupt our regular programming to bring you this special report.

Filed under: On.System ArchitectureDistributed systems

Tagged With: OLTPDatabase
Read more

Infographics: Operation Costs in CPU Clock Cycles, Take 2

June 19, 2026 by “No Bugs” Bunny

Infographics: Operation Costs in CPU Clock Cycles, Take 2 - Post preview image

Quote

The cost of thread context switch, including cache invalidation, can reach up to 3 million CPU cycles

Another Quote

RTTI can be quite expensive, with dynamic_cast<> being up to 5x more expensive than a simple virtual function call

Filed under: On.ProgrammingOptimizations

Tagged With: C/C++
Read more