> But still this seems far less flexible for me then what is currently set up
It totally is. It's also far less costly in time.
> i nearly always have time to highlight something..
I don't. I guess we have different usecase
> My memory is not perfect
Exactly! And it's a feature!! See that as a mechanism to purge an entry from the cache if it hasn't been refreshed for a while.
The problem with bookmarks (and TODO, and everything) is that they tend to accumulate, since there's no natural pruning mechanism. So you need to only use them as a data source, or augment them with some algorithm to extract what remains relevant.
> you still have a mental model for your system, same as i do
My model runs in the background, and doesn't depend on any action.
If at any point in time I need something, I'll fire up a query that will hit all datasources to give me a list of possibilities sorted by this rough score, so something in edge collections (level 4) will be above something in level 3 etc.
Someone here gave stats: in 4 years, for 7000 entries, they used their system 20 times. That's like a 0.28% usefulness. It's not worth spending much time on.
> I wouldnt switch to manually writing SQL queries of bookmarks and relying on my memory if you paid me.
Well, if your system makes you happy, stick to it.
Personally, I prefer to optimize my workflow for 99.72% of the usecases than the 0.28%
> this stuff ultimately is a preference
Totally
> your tone almost seems to suggest everyone should only set bookmarks and search their contents using SQL queries
If you got that impression, sorry. I think your system would work for people who are say researchers and need to carefully read and categorize everything.
My personal belief is that isn't the case for the modal HN reader, and that they'll be best served by being lazy and avoiding premature optimization - and if they really want to optimize, maybe they can start tapping into the data that's already collected (bookmarks, timestamped visit) some of it automatically instead of creating a parallel workflow that's highly dependent on their ability to do some manual tasks (highlighting, reading, annotating) that will require a lot of effort considering it will be useless about 99.72% of the time
> My personal belief is that isn't the case for the modal HN reader, and that they'll be best served by being lazy and avoiding premature optimization - and if they really want to optimize, maybe they can start tapping into the data that's already collected (bookmarks, timestamped visit) . . .
I also want to do as little work up front as possible, but I have spent so much time trying to find something I remember that I read that I try to optimise the process.
I do that (on the Mac) by automatically indexing all the websites that I visit via the app HistoryHound. That way I am pretty sure that I can find a website again by searching by words (that I think) I remember from the content.
I have also tried a lot of external bookmark managers and have now settled for an app called BookMacster. Reading a webpage I then just press a key combination and get up a small window where I type in tags (with autocompletions from earlier tags) and if I want, copy and paste an interesting part into the note section of the window. This way I get tagging that is non existent in Safari and can automatically push bookmarks back to all the browser that I am using. (Not affiliated in any way with these apps by the way).
Finally, sometimes I wonder when I read a page for the first time. I can't see that directly in Safari as the browser only shows the last visit in its history function. So then I access the Safari History via sqlite using something like this:
sqlite> select datetime(history_visits.visit_time + 978307200, 'unixepoch', 'localtime') as dantime, history_visits.title, history_items.url, history_items.domain_expansion, history_items.visit_count FROM history_visits INNER JOIN history_items ON history_visits.history_item=history_items.id WHERE history_visits.title LIKE "%My Hacker News knowledge assimilation stack%" ORDER by dantime;
It totally is. It's also far less costly in time.
> i nearly always have time to highlight something..
I don't. I guess we have different usecase
> My memory is not perfect
Exactly! And it's a feature!! See that as a mechanism to purge an entry from the cache if it hasn't been refreshed for a while.
The problem with bookmarks (and TODO, and everything) is that they tend to accumulate, since there's no natural pruning mechanism. So you need to only use them as a data source, or augment them with some algorithm to extract what remains relevant.
> you still have a mental model for your system, same as i do
My model runs in the background, and doesn't depend on any action.
If at any point in time I need something, I'll fire up a query that will hit all datasources to give me a list of possibilities sorted by this rough score, so something in edge collections (level 4) will be above something in level 3 etc.
Someone here gave stats: in 4 years, for 7000 entries, they used their system 20 times. That's like a 0.28% usefulness. It's not worth spending much time on.
> I wouldnt switch to manually writing SQL queries of bookmarks and relying on my memory if you paid me.
Well, if your system makes you happy, stick to it.
Personally, I prefer to optimize my workflow for 99.72% of the usecases than the 0.28%
> this stuff ultimately is a preference
Totally
> your tone almost seems to suggest everyone should only set bookmarks and search their contents using SQL queries
If you got that impression, sorry. I think your system would work for people who are say researchers and need to carefully read and categorize everything.
My personal belief is that isn't the case for the modal HN reader, and that they'll be best served by being lazy and avoiding premature optimization - and if they really want to optimize, maybe they can start tapping into the data that's already collected (bookmarks, timestamped visit) some of it automatically instead of creating a parallel workflow that's highly dependent on their ability to do some manual tasks (highlighting, reading, annotating) that will require a lot of effort considering it will be useless about 99.72% of the time