Dailog #1
https://hudsonreview.com/2025/10/the-gypsy-life-of-robert-louis-stevenson/
Found the above link in the hacker news. Seems like a book.
I also found out about a tool called - disco.cloud from hacker news. I might be experimenting on that later.
I made a choice to never open HN in mobile. I loose my intentional mindful living. So I will open it in my laptop only.
I will be continue learning go lang and experimenting with it.
https://www.nan.fyi/database
I love this blog. Explained how LSM trees work by building a KV store. We write the data in disk files. That acts as our DB. Updates, search and deletes are harder. So we go to the approach of append only log. Now we face the size issue and search will take longer issue. Now we go to segmentation and compaction process.
To make search easier we go with indices. but for range queries indices won't work. Because the records can be anywhere. So we maintain sorted list of records in memory. When it gets too large we flush to disk. To make sure we are not affected by crashes we write to log file so that we can reconstruct whatever we had in memory.
For search, first check in memory then using index search in the db. Given value we can search in memory if yes return, else look in index and find the value. For ranges, take the range value indices and look around them. How this will work in multiple files and all I am not sure and unclear but let's see.
This is enough for now. I will explore of LSM later. I will probably go through their other blog posts as well.
https://llm-brain-rot.github.io/
exploring this, might continue later...