This is an idea I have wanted to implement myself for a long time. I was very sad that Google Bookmarks doesn't provide full page search and still requires me to think up and maintain my own hierarchy of tags/labels/whatever. Nice work, I'll definitely support you.
My suggestion is that you start looking into ways to group and organize people's bookmarks automatically. You could start with some simple things, but I'm very curious to see how more advanced methods would work out.
My biggest feature request: let me import my google bookmarks quickly and easily.
Thank you, we're glad you like our product! We have a few people in our team (including me) who come from a machine learning background, so we will definitely experiment with some clustering algorithms to see if we can organise the bookmarks automatically.
You can import your Google bookmarks easily by uploading your bookmarks.html from Google directly on our "historify site" page. However, this is currently disabled for performance reasons, but we will soon (i.e. in the next few days) re-enable it for the users' convenience. Please follow our Twitter feed to be informed on the status of this, and thank you again for your preference!
This looks like a useful service, but just wanted to point out that having a JS embedded button might not be the best way of going about it, or at least in it's current form with the historious_key as the only means of authentication. If a site is expecting historious users, it can have a simple javascript polling loop that wait until the historious bookmark is activated to jack the key.
We came out with something just like this a little while ago http://wheatt.com/ , but nobody here seemed to care.
Although instead if emphasizing text search, which I didn't find that useful in the beta testing, we focues on combinations of tags, dates, domains, including terms like "today" and "month", but there is still full page text search, which can be used in combination with those other search terms.
Maybe its too complicated for most, but I find it useful and use it many, many times a day.
To say nothing about the implementation or usefulness of either, I'd say the "historious" branding and landing page makes it much clearer what the experience would be.
The different in response could easily be the marketing? Cool idea either way - both of them.
I signed up, but didn't install the Chrome extension. I think you need to be a lot clearer about what exactly is going on.
For example, if I install the Chrome extension, and then visit my banking site, are you indexing my bank statements?
Do you only index pages we click Historify on? The answer to that question ought to be in the privacy policy. You're asking us to install an extension and give you access to EVERYTHING we do in the browser, but not explicitly stating the limits on it.
That said, if you're indexing EVERYTHING I browse to, I think this is actually a really killer app. An app with some insane privacy challenges, but a killer one nonetheless. If you're only indexing the stuff I bookmark, that's pretty cool, and a good step forward, but less exciting to me.
We're sorry to be less exciting, but we do indeed only index the historified sites, otherwise the noise would be too much for historious to be of any usefulness. Luckily the cost of bookmarking a page is really low (just one click), so you can do it as often as you like!
The extension is open source, you can download and run it yourself if you don't trust the packaged one:
I didn't think about the "indexing everything" from what they said. But now that you mention it, it does sound like a great app.
You could have an extension that indexes everything the user goes to. The trick would be making it usable; I almost never use the "history" in my browser, because it tends to never find what I'm looking for. But if you try and solve it in a centralized way, you could do several interesting things:
* Make searches across all the sites you visited work better, maybe by creating better indexes of the content.
* Make the search work according to sites you visited the most (e.g. I'd prefer the search to find comments on HN before other things, then articles linked to from HN, etc.).
Such a service might be pretty useful: browser history that doesn't suck, and is persistent and accessible online. Anyone want to build it?
The drawback: it saves your history to local disk, and the search can get arbitrarily slow over time. I archive my history every fortnight. For older history I use grep.
It isn't something for normals, but I basically have every page I've ever browsed to since Aug '06.
"The trick would be making it usable; I almost never use the "history" in my browser, because it tends to never find what I'm looking for."
FYI: History in Safari rocks. Just open the history view and you have fast searching (including in page text - just checked) that shows all the resulting pages in a coverflow view.
I created a Firefox Addon and website for it last year that did everything you just wrote about, but I've since closed the project down. Let me know if you or anyone is interested in resurrecting it.
It's not open source. The Firefox Addon captured a visual snapshot and the entire page contents of every non-https page you visited, and stored it to an online database. The website allowed you to search your history, and results were based on number of visits to a page, length of time spent on page, number of times that page was selected as a search result, whether the page was actually bookmarked, etc. It was a very useful tool, but unfortunately the database grew large very fast. I tried to implement a subscription feature to offset server costs, but I had zero subscribers.
The code is all C#/MSSQL/JavaScript/XUL. If anyone's still interested gmail me at my username, and sorry for the off-topic posts!
- Import from delicious, to help get people started.
- An option to automatically add the tags from delicious as part of the text that you search, because many websites do not contain sufficient text to be findable without tags (ex: youtube videos with bad titles, direct links to images/other non-html content, etc)
Thank you for your feedback! To address your points:
- There is currently a feature to import from delicious (go to "historify site"), but it has been temporarily disabled while we make our crawler multithreaded, because the unexpected popularity and people trying to import thousands of links at once made indexing very slow for everyone. It should be up again in the next few days.
- That is a very good idea! I'm not sure if the delicious dump file has them, but it should, I'll add it to our bug tracker, thanks!
Where are my bookmarks? I added them, but I can't find an easy way to see them- although I can search through them.
The auto tweet option is cool, but what about a "post to FaceBook" option. Might help you adopt a few more people.
URL shortening service might also be something to look at- historio.us/linkid. You will have to switch some stuff up, like the key- historio.us/user/key_value.
Allow me to also change my email, not just password. Might even think about using OpenID in as well.
Look into a real FireFox addon- Reddit's Socialite would be a good place to start. You can also make a tiny addon to include search in Firefox[2].
Examples in the api documentation would be nice. I keep getting a "3" response when I request a key.
Found the answer to most of these in the FAQ, but I would still like to do it all with the UI[1]- who reads FAQs?
I'm free this summer if you need some help. All I ask for is a recommendation when all is said and done, or not if you are not pleased.
Thank you for your feedback, let me address the points one by one:
1) You can't see your bookmark, that's the point. You don't have to see them unless you're looking for them! However, we realise people might need to see the entire list, so you can search for * or "http" to get a list of your bookmarks. The * search is a bit broken right now, but we are working to fix it.
2) Facebook posting is also planned!
3) Thanks for that idea, I will add it to our tracker!
4) That is also a good idea, we will have to resend you a validation email but that's fine. OpenID would also be great, we want to support it very very much and will do after the high-priority items are sorted.
5) A Firefox extension is planned, but is lower priority right now. Any volunteers who wanted to write an open source extension would be welcome!
6) You are probably getting a 3 because you are forgetting the "agent" parameter, if everything else is correct. We do need example, though, thanks!
7) Who wants to clutter the UI? :P
We'd love some help, and would be glad to recommend you! Please contact me at stavros at historious if you'd like. Also, thank you for the Mozilla link, I wasn't aware of that and it is very useful.
Okay, quick feedback. The first thing I looked for is the answer to "Public or private?" Privacy policy has this:
> We will also never share the content of your historified pages with anyone, although we might use the URLs of pages to compile anonymized lists of popular or new links. We will, however, never share any information about which pages you, personally, have historified.
I think that might be a mistake. You'll lose the ability to get people in for social networking/SEO reasons, and nobody trusts a service like this to stay private any more anyways. At first I thought, "Well, that's more useful because it's private" - and then I thought, "Who am I kidding? They'll just make it public anyways if they get big."
You might consider setting the default to public, maybe with an option for private.
To address your points, there are no plans for your bookmarks to be public, or have an option to show a page with them, either. The nature of the site doesn't lend itself to public bookmarks at all, as there is no list of bookmarks to publicise in the first place. To make it public we would have to add a search box for people to search in your bookmarks, which may or may not be useful...
I see what you mean about it, and I agree, but many people have said "I will never use this if I can't have privacy", so we listened to them. It might, however, be a typical case of the users not knowing what they want, because nobody seems to mind facebook...
> The nature of the site doesn't lend itself to public bookmarks at all, as there is no list of bookmarks to publicise in the first place.
Question: Are you building a fun tool for the good of humanity, or do you plan to make money at some point? From everything that I've seen, making a way to publicise a list of bookmarks could help you if you're going the latter route.
> It might, however, be a typical case of the users not knowing what they want, because nobody seems to mind facebook...
Yes, my first thought was, "I'd only use this if it's private" - my second thought was, "Wait, actually, I'd use it either way, pretty similarly, and there'd be additional value if it's public." I might see what people choose instead of what people say. People talk a lot and then do different things. Just something to think about, and congrats on launching a pretty cool service.
Your points are very valid. It is also my personal opinion that people say one thing and do another, as far as privacy is concerned. I will put the policy change on the table and we will think about ways to make the service amenable to shared bookmarks, as it would indeed be a good feature.
Currently, a way of sharing that I, personally, have found works pretty well is to enable AutoTweet, which tweets your historified links with a tag. Then you can search for #historified and you see links people all over the world have historified. That is just an aside, however, and we will absolutely look into link sharing.
We don't want to use your links for nefarious purposes or to divulge anything embarrassing, but having a more lax policy that enables sharing is something we do want.
Thank you again for your feedback, it is sincerely appreciated.
Hi. First of all well done in getting something up and running. I think I'm missing the point here. Whats the difference wwhen compared to delicious? I'm not being rude I'm really interested
Thanks, the difference is that historious searches inside the pages rather than just the titles. So, if there was a page talking about fish called "Sea creatures", you could search for "fish" and historious would find that bookmark.
Diigo (http://www.diigo.com), among other things supports saving page snaphosts and annotations.
I've been using Diigo for 8 months now and it's just great (I have used some other services like delicious previously, but they really can't compare to Diigo which just offers a lot more functionality).
Since this is being positioned as a bookmark search engine rather than repository, I think it would be awful useful to provide a couple-line excerpt under each search result, just like in a standard search engine. As it is, one needs to guess a little bit why any given result is given for a search.
1) it doesn't seem to be working, or at least your 'indexer' is so far behind that I can't tell if its working. At least change it so that when I search for '' I see the things that I added and are queued to be indexed.
2) That whole search for '' thing is pretty lame.. put a link to see 'all'
3) Good on you to make a chrome plugin, but I find it unintuitive. Clicking on it to bookmark, great, but how do I get to your domain which I always mistype? I think I'd rather have a very simple menu pop up (big icons?) when I click the plugin, one with '+' one with a magnifying glass for search. Something like that. The two most common things I'm going to use this for is adding and searching, you only make one of those easy with the extension.
Hope you can figure out the scaling issues, bummer to have it melt when you get press.
it isn't clear what it is. i click a bookmarklet and you save it and i can search from there? i can do that with my history, though?
i don't bookmark sites because i don't realize i'll find them valuable till after i've left.
I really dislike the top image you have ("in three easy steps") because i sat looking at it wondering if it were an image or a video or a heading, can you make it more obvious it's a slideshow?
after signing up: ew email confirmation. also seems like i accidentally signed up twice.
i installed your extension, but have no idea what it does :) also, what does "historify site" mean?
if i could summize my feedback in a line it'd be: looks neat, but tell me (the user) what the bet is i'm taking in using your service. clearly.
Thank you for your feedback, it is very valuable to us as we're always looking for ways to explain what we do better. The beauty of historious is that you actually find yourself using it a lot, because bookmarking with it is very cheap, just one click with no lists to clean up.
The image we do need to change, as it blends in too much. The email confirmation is just for resetting your password, we don't use it for much else right now... The extension just bookmarks sites when you click it.
Thanks, we will change the front page to reflect all this!
Many reasons, it's too slow especially if you have limited bandwidth, data becomes much more sensitive (e.g. password, banking details, etc etc), but most importantly, it increases the noise so much you can't find anything at all in the end...
i'll give it a try, though let me give you an advice which reflect my problem with delicious.
I've used delicious, ever since it was a perl script running on a single server by Joshua. Days went by, delicious went through it revolution/evolution cycles and stationed where it is right now. However, as a facebook and twitter user, I want to have all my links and re-tweets save automatically in my bookmark service. That is, If I tweet about a fascinating page, I'd like to able to easily retrieve it 6 months later.
that is to say, auto-tweet is nice feature but I am looking for the vice versa (twitter, facebook, buzz, wherever I go and save/mark links).
I "historified" this site - HN.
Then searched for it in terms of "Hacker News" and got
No documents matched your search.
To add more documents to your history, just install the historio.us bookmarklet!
http://historio.us/search/?q=hacker+news&search=Search
I like the idea and I was searching as simple as that since a long time. Here are some suggestions to improve it:
First, it didn't work for me.. I bookmark hacker news with the chrome extension and I still see the wheel turning.. and even if I search for all site I bookmarked, I don't see hacker news.
Second, is it really necessary to create an account and check with email, etc. maybe histio.us/d0m could create automatically an account where I can search and put a password if I want it secure.
Third, it could be useful if we can search with the chrome extension, instead of only being able to add bookmark.
The Chrome extension has some problems, sadly, but we can't reproduce the issue and they are not consistent at all (e.g. some users found that reinstalling the extension makes them go away). We have no idea why it's behaving that way, sadly :(
Email confirmation not really that necessary, and I agree that it's an unnecessary burden. We will look into removing the email verification (but an account is necessary as you need to be assigned an API key)...
We're adding historious as a search provider in the extension now!
We believe that we should make a fun product first and monetize it later (it's the wrong way around, I know, but we like it)!
""Second, is it really necessary to create an account and check with email, etc. maybe histio.us/d0m could create automatically an account where I can search and put a password if I want it secure.""
As others have mentioned, Delicious import would be handy. Even better, however, would be to take it a step further and automatically import on an on-going basis. This would achieve the following workflow:
1. In my Historious settings, add my Delicious feed URL
2. Find a cool link and add it via the Delicious bookmarklet
3. Historious polls the feed, picks up the new link, and grabs the page
Now I have the best of both worlds:
1. If I want to see a list of my URLs, I can browse them via Delicious.
2. If I want to search the saved page data, I can do so via Historious.
As I mentioned below, there is a delicious importing feature but it has been disabled while we test the new, faster backend. The workflow you propose would be very handy, but it would have to support more sites... Luckily, we have an API anyone can use to write their own tools to add sites to historious!
As an alternative you can put the Delicious bookmarklet code and the Historio.us bookmarklet code together in one giant bookmarklet to post to both at once. It's not very elegant, but it works.
Certainly a useful service. I know this because the way I used to accomplish this earlier was, create a Google customized search and manually enter the sites that I bookmarked with delicious and search from there.
So you got that covered, and I will use it.
However, I just signed up and imported all my delicious bookmarks and did a search for a tag and the results were not so impressive. Only 5 with historious and 17 on delicious.
Hopefully this will get better.
You seem have a problem in Safari 4. If I select the bookmarklet text, then drag it to the bookmarks bar, then only the first line was added as a bookmark.
[NOTE: I just noticed there is an actual bookmark as I suggested below. OK.]
I ended up editing the breaks out of the text to get the bookmark to work.
You should probably bury the bookmarklet text in a link, and have the user drag that to the bookmarks bar.
Also, you are not sniffing Safari 4 correctly. I get a message that it is an unrecognized browser!
>Also, you are not sniffing Safari 4 correctly. I get a message that it is an unrecognized browser!
I would be careful about this. Sniffing browsers isn't always a good idea. Instead, let users drag their own bookmarklet- that correspond with their browser type- to the toolbar. Or keep sniffing and if you do receive an "unrecognized browser" response fall back to a page of that type.
The bookmarklet is the same, the sniffing is for the instructions. The problem is that we want the code below to be the same as the one in the bookmarklet, so we use the same file for both, but for some reason newlines are not removed. I will change that now to make it work with Safari...
I'm sorry, do you mean that the actual bookmarklet doesn't work, or the text below? You're only supposed to use the text if the bookmarklet fails, we just wrap it for display...
This sounds really cool, I have just signed up. That said, my NoScript firefox plugin gets in the way of the bookmarklet complaining about cross site scripting. I had a quick go at putting in a exception regex, but no dice. Anyway, I thought you might like to know that noscript could trip your users up.
I don't have time right to work out why it's not working, so hopefully I'll get to try out your service a bit later.
You found a great article yesterday - now you can't find it!
I think the above style of copy works far better for connecting with readers and calling out the problem you solve. It's shorter (always good) and assumes the reader agrees with you (the ones that don't agree aren't your target anyway - so why equivocate? Plus, in fact everyone has had this problem).
1) That is planned and will be added soon, it is a very good idea.
2) I have no idea why we don't do this already, we'll add it straight away!
3) We will, but right now we will be changing backends, so we will document the new one (probably Sphinx with EXTENDED2).
We don't reindex documents, as they are cached and never change (when you bookmark something, the bookmarklet sends the source to the server, which is how we can bypass registrations and paywalls). Plus, you need the document to be as you remember it.
We currently use Whoosh with Python, but it turns out to be excruciatingly slow. We will change to either Sphinx or Solr.
We don't know why the extension does that :( We can't reproduce it on any of our computers, but many users seem to have the problem :/
I've created a very similar social bookmarking webapp. You just enter a URL, and my webapp auto-tags it, indexes it for search, and shares it with all of your followers.
Thank you for your feedback, it is much appreciated!
The point of the site is to not have lists of bookmarks, though, so you can add many many sites and then find them easily. I find that, if I have a list, I tend to either try and keep it tidy or, if I don't, I just get frustrated and never look at it. We wanted historious to never show you the extent of your bookmarks, and to just give you the info you want transparently. That's why the notion of "public bookmarks" is hard to apply here...
Thanks, we're trying not to. We're pretty resistant to adding tags, categories, lists, etc. Even the listing methods implemented now are side-effects rather than actual functionality.
How about an import feature from google web history. Doesn't it also index the pages you visit for searching? It has a bookmarklet and can crawl all of your web history if you tell it to
I think it does, but we're sort of against indexing everything, because then we'd just turn into Google search... It's much easier to locate something when you have only three or four documents of that type rather than a hundred...
Fast site. What environment are you using? Also, on the page with the instructions I think the instructions should be at the top. People are less concerned with the other stuff.
You mean the screen you see the first time you log in? That's just a short preamble to tell you how to use historious (many people aren't very technically minded and don't immediately see how to use the service). Hopefully it's not too much to read quickly, though!
Thanks for your feedback!
EDIT: Sorry, I didn't answer your first question: We use a Linode in Georgia, so if you're near there it should be pretty fast (their servers are quite fast too).
One thing that would be awesome is if historify would spider someone's blog. That way, for instance, I could point it towards paulgraham.com and have all of his essays saved.
Great idea. I assume you currently aren't saving embedded resources like audio files or flash movies? It would be even more valuable to a lot of people if you did.
Nope, we don't... Those aren't very searchable, though, or we would. PDFs are a bigger priority, but they we still need to have a way to extract text from them...
can you add a tiny search bar on the browser? or at least have it be an option? I don't like visiting the site every time i want to search my bookmarks, also, it would also be nice having a way to remove the bookmarks once i've added it. Especially for articles I've read and don't care to read again.
You can right-click in the search bar and create a browser search for it, and then you can use your browser search as normal. To remove the bookmarks, just click the "remove" link next to them.
Yes, you can search for * and that will give you all the links. There's a small problem with that, currently, so you can search for http instead to achieve the same result.
I'm not getting any results back, even if I search for just http. There are definitely urls stored in my account - if I try and export I can see them...
The hints have various useful tips, and the FAQ page too. There is no page with the syntax, because we didn't think that users will need anything special, but we will add one!
Importing from delicious exists (click "historify site" on the top right) and will be re-enabled as soon as we finish testing the new, faster backend. Please stay tuned (and follow us at @historrific), we're sciencing as fast as we can!
Simple and potentially very useful. Will give this a try for a while (previously using pinboard.in).
Suggestions:
1) Chrome extension: keybinding to add current page.
2) Search results: time options.
- filtering, eg last 24h/last month
- sorting, results with time headlines and good spacing for quicker overview (looking for some news article read last week)
My suggestion is that you start looking into ways to group and organize people's bookmarks automatically. You could start with some simple things, but I'm very curious to see how more advanced methods would work out.
My biggest feature request: let me import my google bookmarks quickly and easily.