Hacker News Reader

Comment by Tooluka | Hacker News Reader

Sure, tags are clearly better than folder trees. But would you tag correctly 1000 bookmarks? Or 10000?

Also if we imagine some person starting bookmark list from scratch, with tags - he has to A) make complete tag system from the beginning (I've tried, it's not easy) or he'll be forced to re-tag old ones later; B) type in tags (or select) every time he bookmarks site.With folder trees I use one click to drag&drop bookmark to the intended place.

Tags are the future but only if someone would be able to automate them completely.

On sets with 10-100 bookmarks type of index system is completely irrelevant - you can even keep them in a plain list and still don't waste time.

View on HN

"Sure, tags are clearly better than folder trees. But would you tag correctly 1000 bookmarks? Or 10000?"

I would if it were easier to do so.

The extra 2 seconds to do it is the bottleneck, so trimming that is the solution. When I click the star or hit ctrl-d in Chrome (I'm sure it's the same in Firefox), the title of the bookmark is in focus. I almost never change that, so that's useless.

Put the tag field in focus, and make tags autocomplete. That'd put me most of the way there.

Of course you're right -- prepopulating the field with tags other people used would be a much further step.

Or try to auto-guess tags you've used previously?

The tag field is focused if I ctrl+d in Firefox. Which is very handy indeed.

Firefox has tag auto-complete as well.

> Tags are the future but only if someone would be able to automate them completely.

Which is why my dream bookmark manager has a "download and index bookmarked page" checkbox. Tags become optional - but you can still search for your stuff easily. A small-scale local information retrieval system would be way better suited to how we use bookmarks than manual tags, and we have some pretty awesome information retrieval models for closed sets of data that we can't use on the web.

If you have 10000 bookmarks you need a search engine.

That's what "automated tags" are... they don't need inventing.

Essentially you could make hierarchical tags, so your folder could be emulated by a tag.

As for the re-tag, the only way for that to happen is either some advanced AI or simply using crowd-source to get what others tagged that site as.

Reply on news.ycombinator.com