Flow of words

This was my experience getting php and apache to run on termux apt update && apt upgrade -y apt install php-apache I was following an old blog post that used php 7 and got these errors Can't locate API module structure `php7_module' in file /data/data/com.termux/files/usr/libexec/apache2/libphp.so: undefined symbol: php7_module Cannot load /data/data/com.termux/files/usr/libexec/apache2/libphp7.so Corrected in apache config file vim $PREFIX/etc/apache2/httpd.conf LoadModule php_module /data/data/com.termux/files/usr/libexec/apache2/libphp.so Apache is running a threaded MPM, but your PHP Module is not compiled to be threadsafe.  You need to recompile PHP. Comment out mpm_worker and use mpm_prefork #LoadModule mpm_worker_module libexec/apache2/mod_mpm_worker.so LoadModule mpm_prefork_module libexec/apache2/mod_mpm_prefork.so This still does not work as apache (httpd) was failing silently. According to Chatty LoadModule php_module ...libphp.so often fails in Termux becau...

I had this idea to create my own custom youtube search page. On youtube you cannot block a channel or have it removed from future search results. I find this useful for example when I search "Tobago vlog" I get the politics spam videos also showing up from one channel in particular. Sometimes others. I want to filter those out. ChatGPT prompt I want to build a youtube search page that shows each result with a "add to watch later" button and a "remove channel from search results" button that removes that channel from current results and future results. Removed channels are added to filter.dat. Search results shows videos from the last week. Folder structure youtube-filter-app/ │ ├── public/ │   ├── index.html │   ├── app.js │   └── styles.css │ ├── data/ │   ├── filter.dat │   └── watchlater.dat │ ├── server.js ├── package.json └── .env I created an MVP and it works beautifully. I had to log into Google Developer Console and create an API...

Why is the Caribbean not a tech hub? I saw this question online and it got me thinking. Who should be asking this question? Who should be answering this question? Do we have the data to answer this question? Besides data, what else do we need to answer this question? And when we have answers, who is going to do the work to change things? This is what my friend Chatty had to say: The question "Why is the Caribbean not a tech hub?" is really a question about history, capital, education, governance, culture, and coordination, not just technology. The Caribbean is often unfairly compared to places like Silicon Valley or Shenzhen without acknowledging that those regions benefited from massive investment, large markets, and long-term industrial planning, while Caribbean nations were shaped by colonial extraction, small fragmented markets, debt, and brain drain. We have some useful data, but not enough to fully understand the region’s digital potential, and data alone cannot answer ...

I had this idea to create a big digital clock display using shell script in termux just for the fun of it. This is what I came up with with the help of Chatty.

I love static site generators. Ease of use and simplicity and speed. I remember trying 11ty and Middleman before. Today I am trying nanoc. Nanoc is written in Ruby. https://nanoc.app/doc/installation/ I am also using the siimple css framework Where the documentation could not help me, I turned to ChatGPT, for example: I am trying out nanoc static site generator. Can I have a list of images and captions and generate a page for each photo with caption and have a list of these pages on the homepage? I asked my friend Chatty about some of the considerations when choosing static site generators. When choosing a static website generator, consider ease of use, performance, customization options, and compatibility with your workflow. Evaluate factors such as the programming language and templating system it uses, support for plugins and themes, build speed, SEO features, documentation quality, community support, and how well it integrates with hosting platforms and version control systems. The...

These were the ChatGPT prompts I used. It is amazing how quickly simple apps can be coded with the right prompting and proper testing and sometimes troubleshooting and debugging. I want to create a play whe payout calculator html/javascript page User enters under their wager Base1 Wager Mega Wager Base2 Wager MU Wager Base3 Wager MX Wager Then multi selects what balls have called Mega, MU, MX, MBonus, MUBonus, MXBonus Bonus pays 36 to 1 Base pays 26 to 1 Mega pays 80 to 1 MU pays 100 to 1 MX pays 125 to 1 Show calculations Base always pays out and user can only select mega or mbonus, mu or mubonus, mx or mxbonus Tax is 10% on winnings 1000 and over I tested it myself and this was the results

This is ChatGPTs suggestions on what could be some next experiments. 1. Add a queue system for async job processing and retries 2. Automate scheduled scraping with cron-based GitHub Actions 3. Create notification rules/alerts for exchange rate thresholds 4. Add authentication with Supabase Auth or OAuth 5. Build an admin dashboard with logs and controls 6. Version your API endpoints (/v1, /v2) 7. Implement webhooks for outbound event notifications 8. Add rate limiting and API security protections 9. Experiment with serverless/edge functions 10. Add real-time updates using websockets or Supabase realtime 11. Build a small ETL pipeline for transforming scraped data 12. Add monitoring, logging, and failure alerts 13. Intentionally test failures and recovery handling 14. Compare JSON storage vs databases like SQLite/Postgres 15. Add analytics and metrics collection 16. Store historical exchange rate snapshots and graph trends 17. Add caching layers and cache invalidation strategies 18. Cre...