For example, as a software developer and as I find Python modules useful - is it worth copying the relevant parts of their documentation into PKM?
I have a fear, that if I copy it, the documentation can change after sometime, but I’m not sure of what i’m missing out on if only reference the documentation and not actually include it directly in the PKM.
For me, I find it much more valuable to update my vault with what I’ve learned rather than the documentation I used to learned it.
For example, recently I was studying how to use generator functions and the
yield statement in Python. After I played around with it a bit and created a working example, I put the working example in my reference vault, along with any links to articles and documentation that helped me on my way.
The idea here is that when I want to recall how to use generators and
yield in Python, my vault will show me the most useful thing – where I left off with the results of my previous research. From there I can dig further into documentation or articles if needed.
Hope this helps!
I contemplated this question for a while. My idea is to use the PKM as an entry point to documentation rather than serve as documentation itself. My concern was partly updating it, but also I didn’t find copying content to be very helpful when the docs are readily accessible. So, instead of copying and pasting everything, you might include…
- knowledge you generated (such as Craig mentioned);
- keywords, tags, and internal links to help locate this information in the future;
- a summary of the concept in the documentation (which you may copy, but it might be better if you explain it yourself).
Isn’t documentation easily google-able? What advantages would having it in your vault serve that a bookmark in Firefox would not?
Semantic search tools, “à la” Smart connection, seem to make use of full articles.
Over time, link rot can make great swaths of sources disappear. I’m undecided that full copies need to be kept in one’s vault, versus elsewhere, but keeping full copies someplace can be prudent — at least for important items.
I am mostly like @Craig - save the trigger stimulant and my associated thoughts.
And, as @EvolvingRichie mentions, with CoPilot plugin now, I can stay inside of Obsidian and just converse with ChatGPT; never have to leave the farm. Claude AI is my current favorite LLM.
PDF’s are now the only full docs I stash in Attachments.
There is the argument for stashing whole documents locally if there is concern or necessity about being connected to the internet. I do this when I’m traveling and parsing large documents.