I've identified several things a bookmarking service should be able to do.
1. In the old days bookmarks were files stored on your local computer that let you find useful sites quickly. that is still the fundamental function of a bookmarking tool. pretty much all the tools out there do this limited thing.
2. local archive. you should be able to backup your bookmarks to an html file. this allows you to use those bookmarks without having them installed in your current browser. say, if your at a friends house, all you'd need to bring with you is that file and you can access all your links, and search through them via description or tag. it also provides the vital "i-need-a-bombproof-way-of-storing-my-bookmarks-'cause-i'd-kill-myself-if-they-got-trashed" archive function. this archive must fully support any taging methods used. the end file should always be html, so it can be opened anywhere. (firefox screwed this one up latley using a .js file to hold the backups by default. although .html is an option, it doesn't support tags.)
3. online backup. it can be a pain to drag that html file with you everywhere, and like anything else, you will forget it. so now, with web 2.0, we'd like to store our links on an easily accessible webpage. that way the only thing we need to remember is that url and we can access our bookmarks on the go. ala, "johnnyboy.tool.com" .
4. sharing. since the web deals in news, videos, pictures, etc... we often find things that we think a friend would like. so a good bookmarking tool should allow for this. i think the best way would be to designate a folder or tag named "share with john" or "share with marcy".
5. the sharing feature should allow john and marcy to subscribe to that folder/tag via rss. that way you don't need to expressly "send" each page to them.
6. some folks aren't going to figure out the rss thing, so it needs to support sharing via email. the interface should not require anything from the user (not even an email account, god knows we all have too many already). you should just click "share via email", a window pops up, you select you friends email address from the list (or type it, if this is the first time you are sending to that address), type a quick message (ala, "hi john, this is the coolest plane i've ever seen. let me know what you think".), and hit send. voila!
it should be that simple, but so far no service i've seen supports that. stumble upon does an adequate job, but the email it sends is cluttered and annoying.
(also, the email that is recived should come from "tool.com" period, no identifing link ti john's account.)
7. bookmarks are fundamentally private. users don't want to broadcast their interesting in light bondage to the whole world, especialy not when it's attached to a username that their family may know. still, they want to be able to share the cute kittem pics with their moms. the trick here is to use user ids instead of names.
- so john doe would sign up as "johnnyboy",
- his online bookmark page would be "johnnyboy.tool.com" (easy for him to remember),
- he'd share folders or tags with his mom as "rss.13729839.tool.com" (or just "13729839.tool.com") (notice that it's not "13729839.tool.com/kittens", we don't want mom to navigate upwards to 13729839.tool.com and see a list of all folders, get it?)
8. discovery. people like to find cool things. a great way to do this is to index what john likes, find others that like those things, and then alert him to new things found by those like minded folks (or rather, let him flip through the newest pages when he has the time, just like stumbleupon). stumbleupon is great at this, but they lack privacy features. meaning anything, anything you "like" is shown under your username. see point 7. (a sub point here is that people like to make friends online, so i would be a nice feature to allow john to "message" "andy" and talk about new sites or other opinions. it is important that these "friends" not know who john is except as "John", his alias. (notice it is different than "johnnyboy"...because any fool could figure that user "johnnyboy"'s private url is "johnnyboy.tool.com". thi wouldn't be huge since you met via "like" matching... but there is no reason to allow that contamination...a mirror of "johnnyboy.tool.com" could be created if john wanted at "1280237027.tool.com" so he could show "andy" his true interests, but this whole process should empower the user to set privacy as he wishes. the default settings should be strictly private.)
- the discovery feature should not show any user name. thus john has the internal profile (the one the matching engine sees) of "johnnyboy.tool.com". this shows ALL "liked" bookmarks, but it is in no way tracible (to the public anyway) to john's account. this way john is sent pages on both kittens and light bondage. but since no one knows the full access "johnnyboy.tool.com" url, they can't see what he doesn't want them to see. he only gives out the spesific shared folder/tags.
9. all of this must be fast. in browser tools are generally pretty good on speed. web tools generally take about 2 beats longer to sych up than you want. or they skip that and just sych on browser shutdown. that's ok...until you have a browser crash in the middle of a marathon bookmarking session. a better solution is needed.
10. nested tags. us old schoolers worked up elaborate folder hierarchies that put a taco website in food/mexican/spicy. (and some far longer). we don't want to trash the old system just to start using (poorly supported) tags. you need to be able to import your old bookmarks and copy that order with tags. so the old taco site can be found by looking for any of the foldername/tags AND you can still get to it via browsing your bookmarks. if you can't nest tags you can't navigate through the bookmarks to see what you've got, it's all just one huge mess in the "root" of the bookmarks folder. so if you didn't put good tags on a page you cant search for it (at least not now, see point 11).
11. searchable. not just by url, title, description,keywords, and tags or folders...but full text on the page. (you can probably get away with just pulling the text off of the page, you don't need to SAVE it, next point, so the storage needs wouldn't really be that huge).
12. store local copies of the pages. maybe not all of them, but make taking a webpage offline easy. just hit the "readable offline" button. the famous "readitlater" extention for firefox can do this, but it should be hardwired into any decent bookmarking tool.
that's all i can think of for the moment. if you have any suggestions, let me know.
if you know a good bookmarking tool that does all of this, i REALLY REALLY want to know!!!!!
ok?
thanks.
:D
No comments:
Post a Comment
Leave your (sensible, useful, nonspam) comments here. If you want me to reply, you need to leave an email address or something. Obnoxious comments will be removed the next time I log in. Thanks for looking :)