WordPress 7.0 drops PHP 7.2 and 7.3 support
The January 2026 Dev Chat confirmed WordPress 7.0 will drop PHP 7.2 and 7.3. Here is what that frees up in core, where the filter extension argument stands, and how to get on the release squad.
Tag: Backend
Historical data analysis when MySQL runs out of room
Running decades of data through wp_postmeta will take your server down. Here is how I structured a 500 million row temperature dataset as an OLAP schema and queried it with MDX instead of raw SQL.
Transformer high-norm artifacts and the Softmax bottleneck
High-norm spikes in Vision Transformers come from Softmax, not from bad training data. Here is why the sum-to-1 constraint turns background tokens into attention sinks, and what registers, sigmoidal gating and self-distillation actually do about it.
Enterprise financial models in Power BI need a star schema
A flat Excel sheet dropped on the canvas does not survive real financial reporting. Here is what I do in Power Query and the model layer instead: profile the columns, reference rather than duplicate, build a dynamic date table, and keep relationships single-direction.
Semantic Models date filtering that doesn’t leak future data
Budget and prior-year lines running past today make a report look broken. Here is the Date Filter table I build in SQL, plus the cross-filter setting it needs, so a single slicer switches future data on and off.
WooCommerce REST API special characters: what 10.5 fixes
WooCommerce 10.5 fixes two decoding bugs that broke Arabic, Persian and Chinese attribute names and values on variation writes. What changed, how to test it with cURL, and which old workarounds to pull out.
Production ML failures usually start in the data, not the model
Perfect offline metrics mostly prove the model fits your assumptions. Point-in-time joins, zero-filled features the model reads as a signal, and cohorts that change while the histograms hold still.
Federated learning: train the model where the data lives
Instead of hauling data into one bucket, federated learning sends the model to the device and gets weights back. A look at FedAvg in NumPy, horizontal vs vertical setups, and the non-IID problem.
How WooCommerce 10.5 picks a product permalink category
WooCommerce 10.5 changes which category ends up in a product URL: the deepest one in the hierarchy now wins instead of the lowest term ID. Here is what that does to existing URLs and how to put the old sorting back.
Why supply chain data science is worth your time in 2026
Why supply chain data science beats another LLM wrapper in 2026: heatmaps before neural networks, Lean Six Sigma for diagnosing processes, PuLP for network design, and Streamlit so the model actually gets used.