> Now, assume I have a generic tag programming, and a couple of more specific tags: python and c. I definitely want my bookmarking service to be smart enough to figure that if I tag some article with either python or c, it means programming as well; I don't want to add the tag programming manually every single time. So, what we need is a hierarchy of tags. Surprisingly enough, I failed to find a service which would support that.
I created tagly ( still in development though), which has tag inheritance, an api and can be used for bookmarking / courses / dms / paywalled articles / .... Support for RSS is included ( everything / per tag / per user (depending on their tags, ...))
When adding a url, if the title hasn't been filled in, it also fetches the <title> tag of the html page
PS. Most of my items contain machine-learning, datasets, AI, ... It's also not released yet, but i'll do it here if there's enough intrest https://github.com/NicoJuicy/Tagly/issues ( currently only used as public issue tracker)
I appreciate the motive but this is just (currently) too difficult for saving things from mobile.
I have a CRUD app that I run on Heroku that paste links into. The app fetches the page, populates the title and some tags. Saved links are organised in set of hierarchical collections - works well for me.
Thanks for the comment. Yeah, as I indicated in the article, I actually plan to implement a native client for Android (and probably iOS), for a better user experience. Not that I'll be able to do that soon though...
Interesting. I currently have a nasty little Android app which hooks into the sharing menus. It allows me with one click to append a link into a file managed through dropbox. That file is in my vimwiki.
However, I do want to have tags, and I like the idea of automating it. Scraping the title seems easy, but tagging automatically? It works well?
I wrote it to host my own bookmarks, with tagging, and searching. As it is just a simple HTML file, and some support media, it's trivial to handle with git and allows me to keep my bookmarks up to data across browsers and hosts.
Yeah it's good. I basically paste a URL into a section and press enter. The app then uses https://github.com/jaimeiniesta/metainspector to get the title and guess at some tags based on the meta tags and title.
I have the same concerns as you. I currently use Diigo (after failed attempt on number of others, Google Bookmarks to be the worst in my experience) but it suffers from not having hierarchical tags which I find to be major problem. It does however let your recent tags be reapplied in single click so it covers a lot of scenarios for me - instead hierarchy I prefix my tags with 'tag category'. Lets say I research if PostgreSql is good enough so I have number of URLs all start with `database postgresql more other tags`. First 2 I will have on all URLs and will change 'more other tags' part. Diigo lets me quickly click latest tags and I manually add others. Later I can quickly drill down on tags I want and limit on hierarchy. Diigo also recommend tags to make this even easier.
One other thing I find indispensable is Diigos annotations, here is an example of your article with 2 tags and 2 annotations:
Since it allows for sharing, I can quickly share important bits from documentation that are relevant for colleges or lets me personally remember important aspects of entire page. This is major feature for me and if geekmarks eventually supports it I would switch to it without thinking (especially as Diigo is not free, although I use free variant that doesn't limit me in most important aspects).
Other things that Diigo offers such as caching, sharing, outliners etc. I can either live without or use some scripting to allow with geekmarks. Recent update of Android client finally made it usable and it works on phone too, something that is nice but it could be easily reimplemented using your API.
Actually, all memory-aiding tools should have full-text search capability. In this case, a bookmarking tool should fetch and store text-only version of URL content to search it later. And optionally, scan it for tags.
Are you doing selective import from the browser? I'd like to split out and upload some of my bookmarks, to see if I can stick around. I generally use browser's native bookmarks manager (which does have this problem of folder/tag hierarchy you talk about) and I sometime use Slack to bookmark non-serious stuff (the stuff I don't much care about, but still would like to bookmark).
I didn't exactly understand - what was the benefit of using programming/python as opposed to programming, python? Was it that when the tag is registered then doing python from then on puts it automatically in programming?
It basically creates a hierarchy. Similar to folder structure. So you could have a general Programming tag, that would show you anything tagged in Programming/, but you could look at just the Python tags inside Programming by going to Programming/Python.
There is none really if you can use AND condition when searching tags.
It may be more practical tho. Currently as I see it, if you type 'python' it will offer you both tags and more i.e. 'programming/python' and 'animals/python' so you can choose quicker.
In many services you can actually use '/' in tags or something else such as '-'.
Interesting… (from comments on the article): "Thanks for the comment. Honestly, I've never heard of Pinboard before, so I can't say anything about it at the moment, I'll look into this."
I've been recently using shaarli, and so far so good. But as a fan of open source, it never hurts to have other options. I look forward to any self-hosting options, and wish the project author good luck!
What about Pinboard's tag bundles?
https://blog.pinboard.in/2012/08/every_day_i_m_bundlin/