MozillaZine

Proposals for Incorporating Machine Learning in Mozilla Firefox

Friday June 18th, 2004

Blake Ross writes: "I will be doing research this summer at Stanford with Professor Andrew Ng about how we can incorporate machine learning into Firefox. We're looking for ideas that will make Firefox 2.0 blow every other browser out of the water. People who come up with the best 3-5 ideas win Gmail accounts, and if we implement your idea you'll be acknowledged in both our paper and in Firefox credits. Your idea will also be appreciated by the millions of people who use Firefox :-). We'll also entertain Thunderbird proposals."


#1 Please respond on his blog

by Waldo_2

Friday June 18th, 2004 6:46 PM

Reply to this message

He wants the responses as comments to the actual blog post, I'm sure.

#2 Re: Please respond on his blog

by AlexBishop <alex@mozillazine.org>

Friday June 18th, 2004 7:04 PM

Reply to this message

Quotage: "I'll read all ideas posted to the forthcoming MozillaZine story"

Alex

#4 Re: Re: Please respond on his blog

by Waldo_2

Friday June 18th, 2004 11:03 PM

Reply to this message

Oops, dummy. And I thought I'd read the article well when I responded. Someone hit me. Hard.

#3 Intelligent Prefetch

by Catfish_Man

Friday June 18th, 2004 8:44 PM

Reply to this message

This isn't really a UI feature, but here goes: Extend the cache so that while the browser is idling it prefetches the pages you go to the most. The idea being that if (like me) you surf maybe 10-30 pages really regularly (say, forums and news sites), then those few sites will be almost instantaneous. The tricky part of this, of course, is doing it in such a way that it never interferes with whatever the user is doing.

/me goes off to think up more ideas...

#5 Related Sites

by Catfish_Man

Friday June 18th, 2004 11:08 PM

Reply to this message

Have something that finds sites related to the current one and displays them in a useful (possibly adaptive, since we're thinking about machine learning) way. This could possibly be done by piggybacking off of google's related sites feature.

#6 URL guessing

by Catfish_Man

Friday June 18th, 2004 11:16 PM

Reply to this message

Is there any way of editing these posts? I feel like I'm spamming.

Anyway, feature idea: a URL bar that corrects mistyped URLs. As an example, mozillazine.ogr is definitely wrong, and in a fairly easy to figure out way. mozillazine.com is harder, but still might be possible using some combination of checking if it exists, and user's past browsing habits (i.e. if they've hit mozillazine.org 50,000 times and mozillazine.com 0 times, they probably aren't going to mozillazine.com). Similarly, perhaps if the server returned a 404 page it could trigger the browser to try some other variants. Unfortunately I can see how parts of this would get amazingly annoying if implemented improperly (and there may be no good way of doing it).

#95 Re: URL guessing

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:00 AM

Reply to this message

I think that'd be a good improvement for Firefox's error pages. A bit like the corrections Google suggests:

"The website <http://www.mozillazine.ogr> could not be found, did you mean <http://www.mozillazine.org>?"

An improvement would be if it would automatically check if the alternatives exist and then display only those that are actually existant (if you enter xyz.ogr and xyz.org doesn't exist, it shouldn't be suggested -- maybe offer xyz.com or xyz.net as alternatives?).

Based on the user's location and language preferences a lookup for relevant alpha-2 country code TLDs could be made too: en-US would trigger a .us search, en-CA would trigger a .ca search, DE would trigger a .de (as would de-DE), .at (as would de-AT) and maybe .si (or whatever switzerland is) (as would de-SI or whatever) search, EN would trigger a search for all english speaking countries, and so on. So, if the user specifies en-CA as well as EN as language preference, only english country codes would be searched, if she only specified en-CA as preference, only the Canadian alpha-2 code would be searched. I have no idea how to cover multinational country codes tho -- although .eu seems to be the only extension which is multinational right now. If no language preference is set, alpha-2 country code TLDs would not be searched at all (except for the next situation -- read on).

A levenshtein check could work wonders as well. Like, if the specified TLD extension is non-existant (ogr for example), the most similar (i.e. those with the lowest levenshtein difference) existing extensions could be suggested ("xyz.nx was not found. Did you mean xyz.nu, xyz.nz, [..] ?").

All that would be needed would be a list of existing TLD extensions along with a language->countries lookup functionality. If a language, country or non-country TLD extension is not recognised, it should be ignored, but not trigger an error, as new valid extensions might be defined by IANA in the future.

A lookup using something like WHOIS to find out whether the site exists at all would probably work better than trying to connect to a domain. The error page would otherwise take too long to load, I guess.

The user should be able to chose whether he wants to use this feature at all and whether he wants it to check country TLDs as well, only do levenshtein checks, etc. Power users might want to disable it to speed up their browsing.

#7 Machine Learning

by Neutrox

Friday June 18th, 2004 11:26 PM

Reply to this message

Would it also be possible to make it so that sometimes we can erase stuff in the inline auto complete, like I might type something by mistake but the inline autocomplete will always bring that up and there is no way to erase one address, we have to erase all addresses. But this machine learning is a great idea.

Also I was wondering if it would be able to upgrade the pop-up, lately sites have been putting pop-ups that come on the website instead of a new window and are really annoying, is there any chance to make it so that those pop-ups don't show up at all. Microsoft is going to put pop-up blocker in the next IE version and in the SP2 so getting one step ahead might help.

#15 Re: Machine Learning

by Grauw

Saturday June 19th, 2004 9:14 AM

Reply to this message

1. You can already remove entries in the list showing as of Firefox 0.9. Try selecting it with the cursor keys, then pressing SHIFT-DEL.

2. Those on-site popups really are pretty much impossible to avoid without also seriously limiting JavaScript or Flash functionality.

#39 Re: Re: Machine Learning

by arielb

Saturday June 19th, 2004 10:05 PM

Reply to this message

i would like to limit javascript features (or dump the js completely) and flash and other plugins on a per site basis just like images and cookies

#97 Re: Re: Re: Machine Learning

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:07 AM

Reply to this message

I think there's already an extension which allows disabling cetain Javascript functionality or Javascript in general per site.

#96 Re: Machine Learning

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:05 AM

Reply to this message

What's more annoying than mistyped auto-fill-ins is actually that you have to select a login if you have used more than one login on the site. What would be better would be a check which login is used the most (on the site, subdomain, directory or in general) and then automatically fill that in. Maybe even giving inputs select functionality if there's more than one login? Like an input with a small select-ish arrow next to it which pops up the login list?

#8 Column View

by pixelcort <pixelcort@pixelcort.com>

Saturday June 19th, 2004 2:34 AM

Reply to this message

In the Mac OS X Finder, you have a view called Column View. This can apply to FireFox, also:

When clicking on a link, it opens in a new column to the right. Clicking the back button moves the focus to the previous column. Forward button moves to the next column. Going back and clicking on a new link clears the columns to the right and puts the selected link into the next column to the page with the link as normal.

Rationale: Having a visual linear view of your trail of link-clicking is great, and it removes much time in using the back button.

#9 Multiple-Tabs

by pixelcort <pixelcort@pixelcort.com>

Saturday June 19th, 2004 2:37 AM

Reply to this message

If the user Shift clicks on another tab, the content will split vertically or horizontally to display both tabs. Clicking on another tab returns the view to normal.

Rationale: Many times a user will want to be able to have visual access to multiple tabs, especially when referring to documentation in one for the work in another. Being able to shift-click on multiple tabs and have them all visible in the same view would alleviate this naturally and functionally.

#11 Re: Multiple-Tabs

by atopal

Saturday June 19th, 2004 2:43 AM

Reply to this message

That's a great idea. I thought, we have this with loading in sidebar, but it's not the same.

#13 Re: Multiple-Tabs

by wiredearp

Saturday June 19th, 2004 3:52 AM

Reply to this message

That's a *fantastic* idea - hardly on the subject of machine learning, but an obvious next step in the evolution of tabbed browsing. Hope it's gets picked up by developers!

#23 Re: Multiple-Tabs

by elzahir

Saturday June 19th, 2004 12:43 PM

Reply to this message

I agree, that is a really good idea.

#47 Re: Multiple-Tabs

by travise

Sunday June 20th, 2004 3:33 PM

Reply to this message

The funny thing is, you can do this with the old MDI UI that was once commonly used in Windows applications. In a way, Mozilla's tabbed browsing is just a more limited form of MDI.

#10 Full content History search

by pixelcort <pixelcort@pixelcort.com>

Saturday June 19th, 2004 2:41 AM

Reply to this message

This one is a little more difficult to implement. Basically, by using a combination of the cache and history, one can search the content of pages available in the history.

Rationale: Say you want to return to a page but can't remember the Title nor URL. You remember a few words from the page, though. By searching the contents of the history by means of using the cache, you can find the pages that have said text in their contents.

#16 Re: Full content History search

by Grauw

Saturday June 19th, 2004 9:16 AM

Reply to this message

This sounds *very* useful.

#29 Re: Full content History search

by richman555 <richman@ptdprolog.net>

Saturday June 19th, 2004 2:41 PM

Reply to this message

Thats a great idea. A new way to search your previously viewed content. Brilliant!

#30 Re: Re: Full content History search

by richman555 <richman@ptdprolog.net>

Saturday June 19th, 2004 2:49 PM

Reply to this message

Actually, after some thought, anything should be searchable in your cache, including images and cached video clips. I know the about:cache function works pretty good, but making it searchable would rock!

#44 Re: Re: Re: Full content History search

by pcabellor

Sunday June 20th, 2004 10:26 AM

Reply to this message

It could be implemented as a "simple" checkbox next to the search button in the History sidepanel. Very handy indeed!

#81 Re: Full content History search

by clip

Wednesday June 23rd, 2004 12:03 PM

Reply to this message

I would find this very useful.

#12 Session Persistance

by pixelcort <pixelcort@pixelcort.com>

Saturday June 19th, 2004 2:45 AM

Reply to this message

When closing the app, save the state of every window and all it's tabs. When opening the app, restore to this environment. See Gnome's session management and OmniWeb 5's implementation for ideas.

Rationale: Sometimes people really do need to close their environment. Being able to return to exactly where you left off is very user friendly and reduces the stress put on users when they need to log out for whatever reason.

#43 Re: Session Persistance

by Melv

Sunday June 20th, 2004 8:42 AM

Reply to this message

There's already an extension that does that - Session Saver (<http://texturizer.net/fir…/extensions/#sessionsaver>).

#58 Re: Re: Session Persistance

by aztkgeek

Monday June 21st, 2004 9:47 AM

Reply to this message

But why not package the Firefox with this extensions, because no all the people know that extensions or why not made a "Today Suggest" with an abstract extension with a random order, with that the people can know the extesions, using the RSS.

#14 Reply

by napolj2

Saturday June 19th, 2004 5:02 AM

Reply to this message

Hmm... there's been this debate about how many preferences to keep in the UI. IMHO, the ideal UI would start out with a relatively small number of options, and as the user got more comfortable and familiar with the software, more options/preferences would gradually be revealed.

And maybe we could have an intelligent user agent to guide the user through common tasks... for some reason a talking paperclip comes to mind. yes, everyone would love that... ;)

#27 "We will always leak"

by gwalla <gwalla@despammed.com>

Saturday June 19th, 2004 2:07 PM

Reply to this message

"the ideal UI would start out with a relatively small number of options, and as the user got more comfortable and familiar with the software, more options/preferences would gradually be revealed."

That doesn't sound like a good idea to me. A constantly-shifting options dialog sounds about as bad as menus with a mind of their own.

#38 Re: Reply

by arielb

Saturday June 19th, 2004 9:59 PM

Reply to this message

easier just to have a basic and advanced mode

#98 Re: Re: Reply

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:12 AM

Reply to this message

Some kind of extended mode would be useful indeed. Maybe even a developer mode for debugging. A power user mode allowing to configure everything that can be configured would be darn useful tho.

#17 Suggestions...

by WillyWonka

Saturday June 19th, 2004 9:42 AM

Reply to this message

I'm thinking this could best be utilized in the bookmarks. Create a list of the most visited/least visited sites in your bookmarks.

Do the same thing based on your history. Sometimes you don't bookmark sites that you visit regularly.

Have it automatically sort your bookmarks into groups. These X pages are similar in topic to one another - group them.

Have the browser learn which topics you're interested in and then find similar pages on google by having the computer try to piece together search queries which it thinks would be good. I'm thinking you might have to train the browser for this one though.

Have it learn which features you haven't been using and then display a list with information on what they do and how to access them. Sort of like a tip of the day. This could be useful for showing people who are afraid of using the computer what useful features they aren't using.

Look at the extensions they have installed and then suggest similar ones that they might be interested in. If the person has a lot of web developer related extensions and a new one comes out, tell them about it. The same could possibly be done for themes and plugins. This should probably be run when the user clicks the check for updates button.

I'm done for now :)

#99 Re: Suggestions...

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:14 AM

Reply to this message

I don't like the idea of the browser messing with my bookmarks on its own. Grouping of the history by topic (there's a search engine with a similar functionality already) might be handy for some people tho.

#18 Remembering paths for file types

by levik

Saturday June 19th, 2004 10:33 AM

Reply to this message

It would be great if Mozilla/Firefox could learn to remember what kinds of files you generally save to/upload from where on your drive.

For example, if you always save image files to My Images, and audio files to My Music, if would be great if the file save dialog would default to the appropriate location based on the mime type of the file being saved. Even cooler would be a feature finding "greatest common" path. For example, if images are saved to My Pictures/photos and My Pictures/paintings, the dialog should open to My Pictures.

For uploading, when you click the "Browse" button, the browser should look at the name and the accepted types of the file input field to make the determination where to open the file selection dialog (also based on previous usage history, and independently of the file save dialog)

#73 Re: Remembering paths for file types

by ivanii

Tuesday June 22nd, 2004 12:57 PM

Reply to this message

>>It would be great if Mozilla/Firefox could learn to remember what kinds of files you generally save to/upload from where on your drive.

There is extension for customizing download folders for file that matches some criteria (DownloadSorter is the name as far as I remember), and also many download managers are able of doing this.

#84 Re: Re: Remembering paths for file types

by fuqnbastard

Wednesday June 23rd, 2004 11:16 PM

Reply to this message

Nevertheless file open/save dialogs are an area which might quite usefully be tackled with machine learning.

#100 Re: Re: Re: Remembering paths for file types

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:20 AM

Reply to this message

Indeed. Something like the select folder dialog when adding a bookmark would be great, especially if the listed directories were the most used ones, not just the ones you recently used. What might become a problem if sorting it by filetype would be that most sites still deliver videos as text/plain (i.e. no special MIME type defined), therefore I'd suggest not sorting it by filetype or IF SO only to sort it by extension.

On a related note ever since I installed Firefox 0.9 MPEGs automatically receive the additional extension MP2 (I blame WinAmp actually), which might be intended to be useful but becomes annoying fast. If Firefox wants to add MIME type based extensions to a file it should provide an option to disable that "feature". Useful for images served via PHP or other dynamic scripts, but annoying for everything else.

#19 Smart prefetching

by levik

Saturday June 19th, 2004 10:39 AM

Reply to this message

I was sure this was already mentioned somwhere, but couldn't find it so here goes:

How about "smart" prefetching. For example, if the browser sees that you were just lokoing at page1.html and page2.html, and are currently on page3.html, and the page also contains a link to page4.html, I think it's pretty safe for the browser to begin fetching this in the background (it could be only the page's HTML if you are on a dialup, images and flash if you are on a broadband connection)

IFrames and remote javascript can be left up-prefetched to avoid making extra ad impressions in case the user doesn't continue to the pre-fetched page. (If they were loaded, such deceptive practice would end up hurting the web publishers in the long run)

#85 Re: Smart prefetching

by fuqnbastard

Wednesday June 23rd, 2004 11:17 PM

Reply to this message

Yeah, I agree. Put that together with the "Fast Forward" Feature like in Opera (just better ;o), in which the logically next page can be accessed by the usual forward-shortcut.

#101 Re: Re: Smart prefetching

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:26 AM

Reply to this message

I think that should only be done if the URL contains a number (that has increased by 1 or something from the last page to the current one) or if a "Next" <link/> is provided.

Something like the "Next/Previous" bookmarklet which increases the number in the URL by 1 would be useful. Just a little smarter. EX: You have been viewing the page <http://www.example.com/bl…s/2004/06/001/index5.html>, then <http://www.example.com/bl…s/2004/06/002/index5.html>, and are now viewing <http://www.example.com/bl…s/2004/06/003/index5.html>. Firefox could recognize the pattern and suggest <http://www.example.com/bl…s/2004/06/004/index5.html> as the next page "in line" (rather than <http://www.example.com/bl…s/2004/06/003/index6.html>, which the bookmarklet would open). This should also work when entering the URLs directly. Just call it "increment / decrement" instead of "forward / back" to avoid confusion.

Whether it should be able to recognize timestamps in URIs is a different question -- I think that'd be going to far for now.

#20 Grouping peoples interests

by solhell

Saturday June 19th, 2004 10:50 AM

Reply to this message

(This post is not about suggesting new web sites based on user's browsing behaviour) If you think of it, a persons topics that s/he browses on internet is not diverse. Bulk of it is maybe 3 or 4 categories and 5-10 websites in each of these categories. If you are interested in technology news, you might be browsing slashdot, mozillazine, etc. If you are interested in nba, you go to nba.com, sports.yahoo.com/nba, and some discussion forums on your favorite team. Having established that, from learning perspective, it would be interesting for firefox to gather the 100 (for example) main websites I visit regularly and divide them into several categories automatically. Let me give more details on what I think: - Everytime I visit sports.yahoo.com/nba/ , I go there to see if there is new commentary. When there is one, I click on it and read it. I want to know if there is a new commentary on that web page (it might be couple days till they put a new item there). But I don't want to know if the sports.yahoo.com/nba/ web site changed or not. It is updated hourly with new news items, but I don't care about the news items. - Another example: I like <http://www.kingsfans.com/forums/> web site. I have it bookmarked. But my interest is not this web site, which lists a bunch of discussion forums. Besides it is updated every minute as new posts are coming. I am interested in "NBA fans forum" on that website. So I directly click on this link when I go there. Now "NBA fans forum" lists a bunch of most recent forum posts. I am not interested in all posts. I am interested if it is related to the Grizzlies, or if it is related to Tim Duncan.

- Going back to categorization, I have several categories of interest. I have several main entery points for each of these categories. I have a main web page in each of these web sites (might or might not correspond to the main entery point). And I have certain behaviour on which kind of links/news/posts/items I am interested on this main web page.

Bottom line is, I want firefox learn this behaviour and then see if there is something interesting for me there...

#102 Re: Grouping peoples interests

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:29 AM

Reply to this message

Reminds me on Google's Alert feature. Interesting idea, although the customising ("only alert me if the change includes the words 'foo', 'bar' or 'qux'" etc) would be a hell to implement.

#21 Preload websites

by Smigit

Saturday June 19th, 2004 10:50 AM

Reply to this message

this is a bit like Catfish_Man inteligent prefetch. However the follow will concentrate not on your favourites so much but on your usage of one site.

anyway what I reckon would be good if whenever a user visits a site, their route through that site will be saved into a local file (encrypted for security). After a few visits firefox will gradually develop a sense for what is the most common route through the site so when i visit the front page, once it has dloaded the content for the page i am currently on it will load up the 1, 2 or 3 most likely pages i will next visit, and perhaps the next 3 for each of those based on the route firefox predicts that I will take based on my normal route through the site. This would require good bandwith management but could make loading pages faster, especally if the page one wishes to visit is quite large and they stop on the before hand page long ennough, say reading an article. THen in the background itll be loading the next page in.

SO just to give an example. Say that 90% of the time i visit mozillazine.org i click the Fx forums link that takes me to the forums. However i dont read the build forums so i next navigate to the main forum page and then finally into the firefox features topic. Assuming i do this regularily enough, as soon as i arrive at mozillazine.org it will load the news page as normal, but once its done that and assuming i havent requested anything else to be loaded, instead of sitting idle it will precache the FX Builds forum (not out of cache but by dloading the page) and once thats dont it may then cache the forum index. As soon as I request a page to load it stops preloading (that data will now be in cache anyway). Basically firefox would dramaticlly learn by browsing habits and would cater.

theres two issue here. people not on broadband who hapen to be dloading other files and dont have the bandwith....but you could say only precache if the user has "so and so" bandwith available for use. secondly on dynamic sites theres no point precaching a forum if i dont then navigate it for 5 mins, The browser would have to make inteligent decisions such as what to cache when (as a forum ou want it to be as up to date as it can be). That said, you could very well load the images and the like so that they are cached.

#22 Untrusted websites

by Smigit

Saturday June 19th, 2004 11:09 AM

Reply to this message

sorry for the double post.

I was thinking about how Junkmail is a great feature of thunderbird and that you could implement a similar feature into firefox. Basically have the ability to somehow flag sites (or have them automatically flagged maybe) that you dont trust or dont want to be viewable for whatever reason. Like the junkmail all the non trusted sites will be analysed against certain criteria and then when you next go to a site that looks like one you wouldnt normally trust the browser will give you a prompt warning you before you view it that it looks like something the user wouldnt want to visit.

Simple case...you flag 30 porn sites, the next time a site that has the word porn in the title is loaded firefox will warn you that the site may be unsuitable or unsafe before you view it. I think this could be a rather useful parental control feature, say you could have it so that once a prompt is up, to actually continue to the site you need to enter in some sort of master password, allowing parents to block sites off to users or children. Some parents may wish to block their kids from any suites that supply file sharing clients for instance or illegal mps's or warez sites. could be a replacement for aps such as net nanny, or at the very least a way to stop ppl having to view trashy pages they may otherwise not want to see.

issue is youd have to visit the site 1st, and once wouldnt be enough and by the time your permissions are set up you would have been better off just avoiding those typses of sites anyway. lol could work but, would need to be refined.

#24 Download similar FlashGet!.

by cuastecomate

Saturday June 19th, 2004 12:58 PM

Reply to this message

Hi! mi english is poor, sorry! I live Mexico. =)

The system of download similar at FlashGet, with mirrors, resume, control bandwith, etc....

#25 Bookmark organization

by Draklyne

Saturday June 19th, 2004 1:12 PM

Reply to this message

-By watching how you move your bookmarks, Firefox could imitate you and learn to place bookmarks where you want whenever you hit "Add Bookmark". It would probably be best if you could toggle its meddling on or off while it's in learning mode, or turn meddling and learning modes on at the same time. You could attach a visual aid in the form of a leash to the cursor, and give it a Leash of Learning and teach it to eat townsfolk...

-Don't know how on topic this is, but you should be able to make Bookmark Manager organize your bookmarks. For instance, Firefox could organize bookmarks by category (tech, health, cooking, etc. determined intelligently), by time accessed, or just alphabetically.

-Firefox could have a resident program that learns when you access the internet and what sites you go to on a daily basis and what time - directly after booting up the computer, at around 4PM, right before shutting down the computer (well...that last one would require an advanced feature, namely seeing into the future). Then it could eventually open a browser window by itself tabbed to the websites it has learned at the time you normally open them. Maybe it will even prefetch the websites. All the settings would, of course, be editable. This would open up another avenue in pranks.

-Intelligent bandwidth detection would be a plus for laptops. You could set low bandwidth to not load images or flash animations, or any object over x kilobytes. Firefox could try to render every page under a certain amount of seconds, removing excess until you're left with text. I'm assuming it would learn your habits. Actually, I'm assuming Firefox would learn your habits in each of the above.

#26 Firefox could let you choose good or evil...

by Draklyne

Saturday June 19th, 2004 1:14 PM

Reply to this message

-The Firefox we have today is sorely lacking something...a digital pet! Firefox 2 should have a built in digital pet for the sole purpose of having a use for machine learning. At least one of the creature choices should be a fox (duh). It should go from an infant to a towering titan. Except don't make the stupid thing crap all over the screen. Oh, excuse my language I meant to say, "intelligent thing". Firefox 2 should have upgraded AI and allow you to teach your Creature to lead armies in massive crusades. And a better graphics engine.

Go here for inspiration: <http://www2.bwgame.com/>

#37 Re: Firefox could let you choose good or evil...

by arielb

Saturday June 19th, 2004 9:47 PM

Reply to this message

yes perhaps firefox could have a little fox that pops up and says ""It looks like you are browsing a webpage. Would you like help?" Maybe it could even do a trick

#61 Re: Re: Firefox could let you choose good or evil.

by brobinson

Monday June 21st, 2004 2:47 PM

Reply to this message

And it could be named Flippy....

#71 Re: Re: Re: Firefox could let you choose good or e

by zookqvalem

Tuesday June 22nd, 2004 10:54 AM

Reply to this message

Ha ha. Just like the paper clip in M$-Office suite... Far out....

#103 Re: Re: Re: Re: Firefox could let you choose good

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:36 AM

Reply to this message

A paperclip fox that could eat townspeople and crap all over your browser -- sounds like a great toy.

It's all but useful tho. I don't think this was an honest suggestion.

#28 Smart mail filtering

by gwalla <gwalla@despammed.com>

Saturday June 19th, 2004 2:25 PM

Reply to this message

This is really just an extention of existing Thunderbird/Mailnews behavior: instead of having a single Bayesian filter just for junk mail, have unlimited Bayesian filters for your inbox subfolders. The system would learn what sort of mail goes in which folder based on mail you move there manually--if a message is moved to a folder, count that as a positive; if a message is moved out, count it as a negative. Eventually you should only rarely have to move messages on your own, without ever having to create a mail filter. Obviously, there should be some way of disabling it one a folder-by-folder basis, for example if you have a "saved" message for specific messages you especially want to hold on to, you might not want the system trying to decide for you.

Another mail & news possibility: an implementation of the Insidious Big Brother Database <http://www.jwz.org/bbdb/> or something like it, integrated into the Address Book. You could see stats about when you last got a message from a certain person, what you were talking about, etc. Extentions that enable other means of communication could also hook in to it, e.g. if you installed ForumZilla you could track discussion with an individual ine email and blogs. In the Suite, there would even be potential for integration with Chatzilla (you could track a thread of discussion with somebody across email *and* IRC...) or the browser (...*and* web forums *and* blogs *and* so on).

#31 Three suggestions

by RazorX

Saturday June 19th, 2004 7:09 PM

Reply to this message

Here are a couple suggestions to incorporate in Mozilla Firefox future version, which IMHO will be helpful. A) Add a total privacy mode. Acceptno cookies, keep no websites in the history, no cache, etc. This is perfect for paranoid people or public internet browsing. B) Add a kiosk viewer thing. What I mean is that Firefox has the option of serving the curent webpage that its viewing on a certain port, another computer can view the Firefox session as it goes a long. Perfect for web presentations, making friends watch you play an internet game, or the like. C) Certain popups or just new windows which you request make a window appear with no address bar, nav bar, etc. Add an option to display that in the standard Firefox right-click menu. Well..... those were my two bits.

#104 Re: Three suggestions

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:40 AM

Reply to this message

Your "kiosk" feature seems like a critical safety risk to me. It might be useful, but it should not be part of the default install. Make it an extension, if anything, but for the sake of truck-sized security holes please don't make it part of the default install.

The last thing you want is remote viewing of your browsing habbits -- although this might be great as a tool for parental control (yes, I know, about three million teenagers will hate me now).

#32 Learned Screen Scraping ?

by alan8373 <alan8373@deronyan.com>

Saturday June 19th, 2004 8:00 PM

Reply to this message

I have an idea that's best described as Learned Screen Scraping. Here's the gist of it... go to any news site, and you'll invariably find about a dozen or more of what are basically useless portions of the page -- banner ads, links, login info, etc, and other junk that surrounds the actual contents of the story you want to read. What if it were possible to teach firefox to scrape screens by you highlighting the actual body of the news story on a web page and clicking on a 'scrape' or 'learn to scrape' button on the toolbar. The next time a news story is read from that site, firefox could, based on what you taught it, either choose not to display what was not highlighted maybe through some CSS magic, or by doing an in-memory screen scrape and then applying some kind of other CSS stylesheet to just display the body of the story itself. This would work for most sites because of content management systems that typically generate the same html structure for every page on the site with the content only being what changed. So, when you highlight a story with the mouse, and tell firefox to 'learn', it can maybe learn that everthing highlighted, would actually mean somethine like ... go to the html ... body ... table ... tr ... td element, and scrape everything in there. for advanced users, maybe we could even do it with the actual page's html source. For me, I would LOVE a feature like this if I were able to go to sites like eweek.com, and linuxtoday.com without the ads and the other extraneous and useless junk on the news pages. I'm sure the advertisers would not be too happy about it, but I'm also sure other people would like this feature. I know I would!

#33 Re: Learned Screen Scraping ?

by alan8373 <alan8373@deronyan.com>

Saturday June 19th, 2004 8:13 PM

Reply to this message

It's me again -- I had some more thoughts on my idea above - this could really be applied to almost ANY site - not just news sites. We could teach firefox to 'scrape' any site and just get the guts of what we actually want. Also, maybe we could setup a central repository to store other people's settings for scraped sites so that firefox could refer to this site whenever it goes to any page. Imagine going to a site you've never visited before, and only seeing GUTS - no banners no junk. All of this could be collectively worked on by everyone in the community for every site on the Net. Way cool!! Also, what if we started to incorporate these into an RSS type engine. Imagine getting RSS feeds with the actual body of a news story - and nothing else. You'd never have to visit the actual site again to get the whole story after getting only an RSS teaser or headline. Imagine how the advertisers would be pooping in their pants if Firefox and some sister web site / service provided the Net's only completely ad-free web browsing experience. Sorry for rambling, but I'm thinking more and more that maybe this is good stuff here ... What if this scraping ability was extended into a completely automated feature such that the browser would use bayesian type filtering at the html element level to determine which portions of a web page were 'spam' - ads and such, and what was actually the real guts of the page. Think of SpamAssasin - but run once for each html block on a page. I apologize for the long-winded-ness of this, but I just had to get these ideas out.

#34 Re: Learned Screen Scraping ?

by alan8373 <alan8373@deronyan.com>

Saturday June 19th, 2004 8:20 PM

Reply to this message

It's me again -- I had some more thoughts on my idea above - this could really be applied to almost ANY site - not just news sites. We could teach firefox to 'scrape' any site and just get the guts of what we actually want. Also, maybe we could setup a central repository to store other people's settings for scraped sites so that firefox could refer to this site whenever it goes to any page. Imagine going to a site you've never visited before, and only seeing GUTS - no banners no junk. All of this could be collectively worked on by everyone in the community for every site on the Net. Way cool!! Also, what if we started to incorporate these into an RSS type engine. Imagine getting RSS feeds with the actual body of a news story - and nothing else. You'd never have to visit the actual site again to get the whole story after getting only an RSS teaser or headline. Imagine how the advertisers would be pooping in their pants if Firefox and some sister web site / service provided the Net's only completely ad-free web browsing experience. Sorry for rambling, but I'm thinking more and more that maybe this is good stuff here ... What if this scraping ability was extended into a completely automated feature such that the browser would use bayesian type filtering at the html element level to determine which portions of a web page were 'spam' - ads and such, and what was actually the real guts of the page. Think of SpamAssasin - but run once for each html block on a page. I apologize for the long-winded-ness of this, but I just had to get these ideas out.

#46 Re: Re: Learned Screen Scraping ?

by slxception

Sunday June 20th, 2004 2:31 PM

Reply to this message

Interesting thought. I thought of something similar involving the user to basically custom-choose how pages should show up, but similar to the Windows advanced display appearance control panel. (I just left my comment in the blog instead of here to keep it simple.)

#66 Re: Re: Learned Screen Scraping ?

by bmacfarland

Monday June 21st, 2004 3:41 PM

Reply to this message

I like this idea a ton, except for the part where it kills the free web, but we can gloss over that for now. I'd like to see it extended to the point where you databased all the text. The text could then be searched (kind of like the history function that people suggested) or used as input on what topics the user likes via a Tivo "thumbs up" or "thumbs down" (or maybe a Hot or Not ten point numeric scale). Let's see what else you could do with this text. You could categorize it and present it to the user as a customized dmoz.org type hierarchy of directories. Then people could edit their hierarchy (maybe delete their categorization of looking at porn sites or whatever) and opt-in to share their hierarchy with a centralized website. From that site they could meet new people, make new connections with people they have stuff in common, discuss their interests, whatever...

I realize this is a combination of ideas. It would cross a few different areas of machine learning. The only other thing that I can think of in machine learning would be something like a talkback feature, but instead of report bugs, it would report what people are doing (mouse trails and the like). Gathering this information and using it effectively, you could reduce the tasks that appear to be taking the longest to complete and/or those done most often. I'm not sure how this could work, but I'm putting it out there for you to run with.

#50 Nuke Anything + Bayesian Filter

by superyooser

Sunday June 20th, 2004 11:15 PM

Reply to this message

I think the solution is to make something like Nuke Anything <http://update.mozilla.org…mp;category=Miscellaneous> that you could train with a Bayesian filter to pre-emptively nuke what you don't want to see. This would be good not just for clearing out the clutter, but for blocking offensive content like pornography, hate speech, etc.

#70 Re: Nuke Anything + Bayesian Filter

by phaasz <phaasz@hotmail.com>

Tuesday June 22nd, 2004 7:51 AM

Reply to this message

Hmm sounds cool. I think some thought would need to go into identifying the "unit" for the engine to use though. For email, it's easy; the unit is a single message. It seems that su is suggesting a scheme where you could select arbtrary text (eg "hate speech"), which points to a switch to tokenising arbitrary numbers of consecutive words (which would be resource intensive)... Or are you suggesting some other tokenising scheme?

How about something which blocks entire urls based on bayesian filtering (whether the resource be an image, web page, etc )? But I would not just look at the url string itself (as most adblocking software currently does), as Bayesian filtering is most effective when more information is available - eg for spam, taking into account entire email content including headers, etc. Which means we need more than just a url in order to be smarter...

We could also parse parts of (or all of) the linking page: - the html element in which the url appears (a, img, etc) - the parameters of that element (eg height, width, alt text ) - the style of that element - other contextual aspects?

And for the resource itself: - http header fields (after which we could possibly drop the connection?) - the entire page (for web pages)

This scheme could then be trained for: - blocking offensive content - blocking ads - blocking certain content types (eg swf) - blocking large content (by reading http header)

To do all of these things, of course, the engine would need to be semantically aware (eg treat a number in the content-length http field differently from that same number in a url), but this shouldn't be too hard using a scheme similar to that suggested in Paul Graham's second attempt: <http://www.paulgraham.com/better.html>.

...enough rambling!

#78 Re: Nuke Anything + Bayesian Filter

by lump1

Wednesday June 23rd, 2004 9:09 AM

Reply to this message

Yes, this is by far the best idea here. I'm thinking more of AdBlock with a Nuke Anything interface. You'd start with a very conservative set of "priors" for the Bayesian filter which would block the most obvious ads that nobody wants to see, and have no false positives. Then, any image that would display would have a right-click option like "this is an ad". Clicking that would not only nuke the image/flash/iframe but all the data about that nuking would be stored and adjust the priors for the Bayesian filter based on as much data as possible, including (obviously) URL, disconnect between the URL of the image and of the page, dimensions of the image, the link target (if the image is a link) and position in the page. A bonus would be the ability to easily import/export and share training data.

An important part of the interface would be an icon that appears in the status bar which indicates that images are being blocked. Clicking it would override the blocking and show all the content of the page, with a thin red border around images that the filter thought should be blocked. If any of them are false positives, you could choose "this is not an ad" and train the filter to allow them. What would be an especially nice touch is to have that status bar icon appear in different colors depending on its perception of the "safety" of the blocking decisions that it made. So if the blocking icon shows up and it's green, it means the filter blocked something, but it's absolutely certain that only worthless stuff was blocked. If the same icon is yellow or orange, it means that at least one of the blocked objects is only about 90% or 85% likely to be worthless. That would encourage the user to click the icon and double-check the decisions of the filter. If the user decides the decisions were correct, that would be further data for the filter, and it would make similar decisions in the future with more confidence. This is a perfect problem for machine learning, and it would give people a sense of ownership over Firefox, since it would be personalized for them. It would also make them want to evangelize Firefox to friends, and to share with them their training data.

One more thing: this, or any result of this project, should be made an *extension* and no attempt should be made to cram it into the mainline code. That way it can be tested by users long before 2.0 and development won't slow down the main browser development effort. The extensions framework is beautiful, and whatever additional feature anyone can think of should be made into an extension, unless there is a very powerful reason why making it an extension would compromise its functionality. For example, bayesian filtering for Thunderbird should be an extension, so that several competitive bayesian classification schemes could be used. On the other hand, I think that a post-1.0 Firefox and Thunderbird installer wizard should include a screen with checkboxes for auto-installing the most important extensions, like this one, mouse gestures, advanced preferences, and maybe two more. It's really the extensions that make Firefox such a superior browser, and this should be showcased. So don't worry that I'm trying to diminish the prominence of this important work by suggesting it should be an extension. I'm trying to save the sleekness of Firefox and to increase the prominence of the greatest extensions.

#83 Re: Re: Nuke Anything + Bayesian Filter

by phaasz <phaasz@hotmail.com>

Wednesday June 23rd, 2004 8:42 PM

Reply to this message

I *really* like the implementation suggestions. I would argue that there should be no set of priors, because false positives mean different things to different people. Your default set of priors might be annoying to an avid porn hunter, for example :) Import/Export is essential. I don't think we should just limit ourselves to images, however, as per my previous post. You could achieve all suggested in your comment with a more generic scheme that allows blocking of any resource (eg swf, html, etc) whether it be embedded in a page or not.

Slightly off topic but I concur that regardless of which idea(s) are chosen, they should be implemented as extensions because this effects a "survival of the fittest" environment in which the best extensions will reign supreme, and because users are allowed to construct the browser of their dreams. An argument against this, of course, is that take-up by the non-technically savvy could suffer if one needs to manually install most of the things that make Firefox the best browser (current handling of Java and Flash in particular are inexcusable - at a minimum, a linked xpi should do all of the work). Having a "starter kit" download bundled with a set of the most popular extensions according to update.mozilla.org (which can then be chosen between at installation time) seems like a good compromise here.

#86 Re: Re: Re: Nuke Anything + Bayesian Filter

by superyooser

Thursday June 24th, 2004 4:01 AM

Reply to this message

You could have predefined sets of priors profiled for different kinds of users: "ad hater," "protective parent," "no flashing stuff," etc.

As for implementation, I was thinking that you could click into "Train Nuke Mode" and draw shapes around what you like or don't like. You know how a lot of news pages have the article in the center column and cruft everywhere else? Draw the rectangle around the article and "crop" the rest out. All the page elements you excluded get fed to the learning filter as *bad*.

Better yet, look at the Web Developer extension for Firefox and see the Outline menu. It draws colored boxes around block level elements (all), frames, and table elements. In Train Nuke Mode, you could select individual elements easily this way. Perhaps a gesture or hotkey could enable and disable outlining elements. There should also be DOM tooltips in case there are lines all over the place and you can't tell what's what.

This is something that the regular Nuke Anything should have. When it says "Remove this object" it's often not clear what *this* object is. Am I nuking a whole table? or just a tr? or td? A div? Or the span, a (href), or font within the div? Sometimes I accidently nuke the main content of the page.

#89 Re: Re: Re: Re: Nuke Anything + Bayesian Filter

by lump1

Thursday June 24th, 2004 10:19 AM

Reply to this message

Your idea about specific-use priors is very good, but the one about the Train Nuke Mode implementation is absolutely great. I want to see this at work... and soon! The extension should be called "Learned Intolerance" (just kidding). It looks like web pages are getting more annoying, useless crap all the time. With some effort, it's not hard to defeat AdBlock, currently the best ad filter on the internet, IMHO. Few have started trying to do so, but they will... unless we can pre-emptively develop such a powerful garbage removal system that they stop trying. The spammer vs. filter arms race is already quite far. We can still prevent a new arms race of hiding vs. forcing unwanted web content.

#105 Re: Re: Re: Re: Re: Nuke Anything + Bayesian Filte

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 6:52 AM

Reply to this message

As I'm against internet censorship (*cough* <http://eff.org> *cough*), I'm strictly against the idea of presets. I also think that you would be, technically, violating the Terms of Service of some sites by filtering out all advertising. Especially if they provide memberships which disable the advertising. Self-promotion is indeed a common problem (imagine every site you went to had dozens of pop-ups and banners about how great it is and why you should buy a ten-year membership right now), but advertising, how unpleasant it may seem is an essential part of surviving as a free service provider. Don't even think about claiming donations would make up for it -- the reality is that only a fraction of the users are willing to donate money, and even they will only donate small sums. Few sites can survive by donations.

I do not think Firefox should promote the blocking of legitimate advertisings (excluding malicious code and techniques, eg. pop-ups, pop-unders and illegal content).

#110 Re: Re: Re: Re: Re: Re: Nuke Anything + Bayesian F

by phaasz <phaasz@hotmail.com>

Tuesday June 29th, 2004 1:02 AM

Reply to this message

Arguing that a generic learning content filter will spell the end of the free web is a little disingenuous, me thinks.

I tend to agree that presets should not be provided (at least not official ones), but only because the whole point of a learning filter is that it should learn from an individual in order to reflect their choices. Remember that this self-same facility would be useful for objectionable/illegal content, popups, etc. I, for example, work at a company with very stringent web quotas, and I would attempt to teach it to block images above a certain size, flash animations, etc. In summary, blocking legitimate advertising is one of many uses of such a filter, just as downloading kiddie porn is one of the many uses of firefox. The moral dimension belongs not to the tool, but to how it is used.

To look at it another way, what we are talking about are individual resources that are accessed individually via HTTP by our client tools (eg firefox) and rendered accordingly. I could be using a phone or a text-only browser (eg lynx) to browse a site, and if that involves removing all images, then that's not a moral question AFAIK. Don't underestimate the resilience of commerce - perhaps (puts on optimist hat at this point) stronger web shit-filters would encourage advertisers to package their messages into a format that their target audience don't want to block (think virals, etc).

#118 try an ad filtering program

by GlenCoakley

Thursday November 4th, 2004 9:30 AM

Reply to this message

You could try a flexible ad filtering program like Proxomitron: <http://www.proxomitron.info/> It is no longer developed but, I'm not sure what more you would add.

#35 Font size in new tabs

by sjmuniz

Saturday June 19th, 2004 8:44 PM

Reply to this message

Well... i use glasses... big ones. Sometimes i need to use tree times ctrl + to be able to read. And i _hate_ when i go to the next link within the site (the same site where i pressed ctrl + 3 times ) and the new tab has lots of ants that seems like text, because Firefox opened the new tab with the default font size. I'd like to open the new middle clicks with the same font size as it's parents. Thanks for this great browser :)

#45 Re: Font size in new tabs

by pcabellor

Sunday June 20th, 2004 11:27 AM

Reply to this message

or, since we're on the learning topic, Firefox could learn how big you prefer your fonts and autotune them.

#79 Re: Re: Font size in new tabs

by jgraham

Wednesday June 23rd, 2004 11:09 AM

Reply to this message

Well I have the beginnings of some code that works along those lines (the idea is to adjust the zoom level so that the most common font size for paragraphs on a page is the preferred size that you specify in prefs). I'll post it as an extension after I have time to finish writing it (annoying distractions like a graduation cermony and making sure I have somewhere to live and money to pay the rent are getting in the way at the moment ;) )

#36 Font size in new tabs

by sjmuniz

Saturday June 19th, 2004 8:55 PM

Reply to this message

Well... i use glasses... big ones. Sometimes i need to use tree times ctrl + to be able to read. And i _hate_ when i go to the next link within the site (the same site where i pressed ctrl + 3 times ) and the new tab has lots of ants that seems like text, because Firefox opened the new tab with the default font size. I'd like to open the new middle clicks with the same font size as it's parents. Thanks for this great browser :)

#40 Regular sites

by RazorX

Saturday June 19th, 2004 10:27 PM

Reply to this message

I'm sorry, but my last post missed the scope of the suggestions. What Firefox can do is observe regular sites which cintai news articles, forums, webcomics, basically a similar webpage whose contents changes over time. Then, on startup, you can choose for a tab to come up whose screesn is split up into all your forum sites into one tab (expanding on an earlier post) or the choice to have them all startup in different tabs.

#41 Regular sites

by RazorX

Sunday June 20th, 2004 12:02 AM

Reply to this message

I'm sorry, but my last post missed the scope of the suggestions. What Firefox can do is observe regular sites which cintai news articles, forums, webcomics, basically a similar webpage whose contents changes over time. Then, on startup, you can choose for a tab to come up whose screesn is split up into all your forum sites into one tab (expanding on an earlier post) or the choice to have them all startup in different tabs.

#42 Page Rendering Settings

by shodan

Sunday June 20th, 2004 12:35 AM

Reply to this message

Draklyne touched on it but I understand that there are a couple of settings for optimising the way that Mozilla renders new pages as they are downloading. Right now they are hidden prefs that are set for a balance of bandwidth, lower page draw & time to useful page on a hypothetical "most common" system but there might be some gain in making the setting more dynamic, based on the bandwidth that is experienced either for the browser in general or possibly for the site. This isn't going to save a lot but I think that this sort of behind-the-scenes optimisation is going to have a more positive, if less visible, impact than messing with my bookmarks... although I do like some of the ideas for making history more useful.

#48 Possibly impossible?

by johnlar <johnlar@tfn.net>

Sunday June 20th, 2004 7:49 PM

Reply to this message

How about using something like "whats related" to generate suggested sites for when a person is bored and wants to explore new stuff. Basically the browser gets an idea of the type of sites the person visits and can suggest other sites that the person may be interested in based on their history.

#49 Highlighting

by nogwater

Sunday June 20th, 2004 10:16 PM

Reply to this message

It would be great if Mozilla/FireFox could automatically highlight text and or paragraphs that it thinks you'll be interested in. It could either be based on words/phrases that in the past you've shown interest, or based on recent search queries (or any text submitted in a form?)

This would be especially useful with links. If there was a way for FireFox to know what links you're most likely going to want to click (again based on past history, or by pre-crawling) and then highlight those links somehow, it would make surfing that much more fun. Of course, it could get you in a bit of a rut. Imagine doing a Google search and having the 4th result down light up because it's actually the one you want.

Maybe there could even be a way to mix this with something like stumbleupon.com.

#51 Autocomplete

by theseer <theseer@gmail.com>

Monday June 21st, 2004 2:06 AM

Reply to this message

okay some people that type fast often make mistakes when typing in the url. me i always end com with an n or ,. *.con, *.co, if mozilla firefox detected this was happening quite often for the certain extension it could be automatically fixed?

#106 Autofix

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 7:02 AM

Reply to this message

I think that's pretty much covered in the auto-fix (or rather "suggest corrections") thread a bit higher. A protocol correction would be handy as well. Oftenly, when dealing with "sponsored" links, you'll get URLs like <http://www.example.com/fo…p;http://www.example.net/> When copying the URL and removing the first part (which simply redirects you or opens a different URL on a random basis) you might accidently leave the & in place and get something like this: &<http://www.example.net/> which sometimes only triggers an alert() in Firefox. If you have switched to a different tab in the meantime, it might happen that the URL is gone and the tab only shows a blank page. A good solution would be an error page like "Error: &http is not a registed protocol. Did you mean <http://www.example.net/> ?" or so. This would fit nicely with the domain extension sniffing/fixing suggested earlier.

#52 Chat Bot

by theseer <theseer@gmail.com>

Monday June 21st, 2004 2:08 AM

Reply to this message

a chat bot that learns from what it analysis from certain websites. maybe a button that tells it to read. so u can just ask it questions that it has analysed in a website and it will recall it straight away.

<http://www.somesitewithgmailinformation.com> *clicks analyse* *opens bot* *says when is gmail becoming public* *replied: it has not been announced yet*

#53 crawler

by theseer <theseer@gmail.com>

Monday June 21st, 2004 2:15 AM

Reply to this message

something that crawls the internet for things related to the current website u are browsing. u acn then view the info showing u website urls and a snippet of text from them to see if u want it. over time if u visit a certain site a lot of times it will crawl that when u r idle isntead of when u are visiting a certain site.

#54 crawler

by theseer <theseer@gmail.com>

Monday June 21st, 2004 2:27 AM

Reply to this message

something that crawls the internet for things related to the current website u are browsing. u acn then view the info showing u website urls and a snippet of text from them to see if u want it. over time if u visit a certain site a lot of times it will crawl that when u r idle isntead of when u are visiting a certain site.

#55 Thunderbird automatic filtering

by leafdigital

Monday June 21st, 2004 5:24 AM

Reply to this message

My mail would benefit from being categorised, but I'm too lazy. The software should do it for me. (This is not quite the same as somebody else's suggestion for using manually configured categorisation that can take advantage of the Bayes engine.)

There are a number of ways you could create an automated filtering system. One very simple way would be to use the 'to' address; if you receive a number of messages to a specific 'to' address (e.g. <webmaster@example.com>) that isn't your main account address, it could automatically create a new folder and filter those messages for you. You'd want this to work in a 'sensible' way so that it didn't bother creating a new folder if you only get 1 message per month to that address, or if somebody sends a bunch of spams there, but only if you actually would be likely to want it.

(For the UI I would implement this as 'categorisation' - the AI system 'categorises' messages according to these rules, makes the categories itself, and may create automatic filters, but you can change the filters by assigning messages from a category to go into a different folder etc.)

Just using 'to' addresses would actually be a really good start for some users; then extending it to 'from' addresses might also help (e.g. you get a lot of mail from your boyfriend, why not filter that elsewhere). You might then start looking at other factors such as text classification (the existing Bayes system), but I'm uncertain as to how best to implement that. Vector quantisation techniques might help. Perhaps this could only work in conjunction with user feedback.

This kind of system does rather reveal the flaws in the 'folder' system of mail storage in the first place; clearly a better kind of mail database ought to be used which *doesn't* categorise mail insofar as its physical storage location (i.e. just store by date received in some easily backupable manner) but applies one or more categories to each piece of mail, as set by the user or by AI systems. I can see that being a UI research topic, though it's not really AI. (Basically, you can still see mail as folders if you like, but it could be in more than one folder.) I'm sure it has been implemented before.

Incidentally, I know this isn't the place for it but bloody hell Thunderbird really needs to store the 'filter' setting per-folder... in my inbox I only view unread, everywhere else I want view all... it's cool that you can create your own filters and stuff, great, very nice, but sort of missing on that basic functionality.

Another obvious flaw in the current program is the spam detector (which is good, but could be improved); the solid Bayesian text-only-ignore-the-headers approach that appears to be used means it misses out obvious opportunities for learning like 'mail sent to <real.address@example.com> or <another.real.address@example.com> is more likely to be nonspam', after user marks some mails to those addresses [and as a converse, mail sent to <chat@example.com> or <sdgsgsdgsdg@example.com> is more likely to be spam]. Isn't there some way to take account of at least these basic headers? It doesn't appear to do so currently.

I don't think Fire* really needs AI to any great extent except for one issue which somebody else mentioned - automatic bookmark categorising. Yes please. Could be tricky, though... The history search is a great idea too.

--sam

#68 Re: Thunderbird automatic filtering

by phaasz <phaasz@hotmail.com>

Tuesday June 22nd, 2004 7:25 AM

Reply to this message

Creating categories automatically would probably be quite unintuitive because they are so subjective. I certainly concur that categorisation is much more useful than foldering. Perhaps a learning Bayesian filter that learns categories instead of spamminess (as currently implemented) or folder (as suggested by gwalla above)? The user sets up whatever categories make sense (with some default categories, one of which would be "Spam").

The engine would automatically assign categories to each piece of incoming email. Marking a message as spam would function as it does now. Similarly marking a message as "work related", or "funny", or "legal action" or "job applications" (or some combination) would work. For every category, every message either belongs or does not belong. By having the user set all applicable categories with incoming email, the engine would be trained.

I would envisage that as it goes, the engine would guess all of the categories for each incoming mail article, and that this categorisation could be corrected as needed by the user to further teach the filter.

The notion of folders could then disappear. Instead, one would simply configure arbtrary "views" which are effectively stored searches (with category, date, sender, etc available as filters), groupings/sorts (conversation, sender, date, size, etc).

The junk mail folder is simply a view that shows all messages with the junk mail category set...

You'd probably want to attach a pretty mean indexing engine so that it could be searched effectively.

I hope all of this makes sense!

#56 Sidebar Viewer

by Packetrat

Monday June 21st, 2004 6:30 AM

Reply to this message

Most of the obvious and best suggestions have been made: 1) Learned prefetching, either based on relative/absolute linking or learned through link histories and page location. 2) Smart bookmarking, including a cycling top 10/25/50/n most visited links. 3) 'Related To' features that find pages similar to current page. 4) Content filtering based on mouse-selection profiles, either whitelisting or blacklisting. I just wanted to point these ideas out as confirmation that they are relevant.

Unrelated to Machine Learning itself, whether it be prefetching or content filtering, Firefox would do well to have a sidebar that would display the page title, url, and percentage of prefetch or modified construction complete. Additionally, and more importantly, it would have a toggled parameter that would allow/disallow the display of a thumbnail of the page as it would look displayed in the main window. When the mouse cursor moved over these thumbnails, either an enlarged view would raise up or the page would temporarily be displayed in the main window - that is until the cursor was removed.

In the case of prefetching, the user could tell if the prefetch was useful before s/he visited the link or based on a confirming click-over to the main window. With content filtering, the sidebar could display multiple guesses at what the user would like to see or not see. These thumbnails could then be whitelisted, blacklisted smartly to determine future prefetches or content-filtering for individual domains. There should also be a UI, whether a sidebar or dialog, that allowed the user to delete site profiles so that excess learned data can easily be discarded.

#90 Re: Sidebar Viewer

by Packetrat

Thursday June 24th, 2004 11:05 AM

Reply to this message

Afterthought: when reviewing links with the cursor in the actual page, the corresponding prefetch should be highlighted in the sidebar.

#57 FileManager

by aztkgeek

Monday June 21st, 2004 9:42 AM

Reply to this message

I think if firefox is small, quick and is only a navigator and made it like a FileManager like Nautilus or Windows Explorer with a FTP client, FTP-SSH Client, CSV Client, etc, Firefox can be the best navigator, and why not, transparent window style and resize window style.

#107 Re: FileManager

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 7:10 AM

Reply to this message

I agree. Firefox urgently needs a proper FTP browser (with a drag and drop interface). Either make it part of Firefox or produce a new product (which could be linked with Firefox -- I think Thunderbird is going to be linked with Firefox, at least that is what the "Mail" icon in Firefox 0.9 suggests). It bugs the hell out of me that MSIE is the only way to go if you want an in-browser FTP interface.

#59 Thumbnail tabs

by aztkgeek

Monday June 21st, 2004 9:51 AM

Reply to this message

Made a Thumbnail sidebar with the previews of the tabs, like MacOS Dock, but in a sidebar and in a tooltip over the title tab.

#60 Remembering DOM/format on page reload

by Racer

Monday June 21st, 2004 12:35 PM

Reply to this message

Often times, when you go back to an existing site, even if its text content has changed, the format of the page stays the same. Why not have Mozilla assume the page still has the old format until something changes it. You could even pre-render using that format and any old cached images (and cached css or js files if that is possible). Of course, if the page's format has changed, Mozilla would reflow to the new format; but that would be the exception, rather than the rule.

#62 Bookmark Suggestions

by Shinglor <shinglor@optusnet.com.au>

Monday June 21st, 2004 2:59 PM

Reply to this message

I like the prefetching idea, what I would really like to see is a way to submit your bookmarks and have suggestions returned chosen from other people with similar bookmarks ala iRate. Also star ratings for bookmarks like iTunes. Oh and bookmark sorting BUILT IN! Seriously, it's crazy that I have to download an extension for that.

#63 Bookmark Suggestions

by Shinglor <shinglor@optusnet.com.au>

Monday June 21st, 2004 3:00 PM

Reply to this message

I like the prefetching idea, what I would really like to see is a way to submit your bookmarks and have suggestions returned chosen from other people with similar bookmarks ala iRate. Also star ratings for bookmarks like iTunes. Oh and bookmark sorting BUILT IN! Seriously, it's crazy that I have to download an extension for that.

#64 Bookmark Suggestions

by Shinglor <shinglor@optusnet.com.au>

Monday June 21st, 2004 3:00 PM

Reply to this message

I like the prefetching idea, what I would really like to see is a way to submit your bookmarks and have suggestions returned chosen from other people with similar bookmarks ala iRate. Also star ratings for bookmarks like iTunes. Oh and bookmark sorting BUILT IN! Seriously, it's crazy that I have to download an extension for that.

#65 Whoops

by Shinglor <shinglor@optusnet.com.au>

Monday June 21st, 2004 3:03 PM

Reply to this message

Sorry for the triple post.

#67 More GMail

by Kommet

Monday June 21st, 2004 5:24 PM

Reply to this message

Blake, if you need some additional GMail invites to use as prizes, I'll be happy to contribute some. I have a few left (I think 2) and you can have them all.

I might be able to get you more than my meager supply, though. In fact, I might be able to get access to a whole boatload. My roommate (who works for Google) has 50 or so left, so I can probably snatch a few.

Just leave a PM for user Kommet in the MozillaZine forums or reply to this post. I'll check back in a while...

#69 project managed bookmarks

by smkatz

Tuesday June 22nd, 2004 7:33 AM

Reply to this message

If I go back and forth between 3 bookmarks for a consideradable time period, or open up the same tabs. perhaps an "ad-hoc bookmarks bar" could appear or could be created in the bookmarks menu (with an icon in the status bar). Keep in mind that this machine-learning could be extensions. Firefox is perfect for that.

My reasoning: Piles of paper stay on your desk. They are eventually filed into drawers. Bookmarks are really counter-intuitive as a metaphor. With tabs, I like "piles" better.

Netscape has the concept (or had the concept) of a default folder. I'd like to expand on that idea.

Dare I say, "It looks like you are working on a project. We've auto-created the following temporary bookmarks.."

Don't show this dialog again--its a feature, not a bug.

--Sam

#72 Thunderbird multiple identity management

by smithjt

Tuesday June 22nd, 2004 11:32 AM

Reply to this message

One place where AI would help me greatly on a daily basis is via the handling of email addresses in Thunderbird. I own and entire domain and use unique aliases for anyone not white-listed (newsgroups, web forums, internet shopping, etc.) to sort, filter, and track address dissemination. I want to be able to not only reply automatically with the approriate alias but to have the proper alias guessed from the recipient whenever I compose an entirely new message. I also want TB to be able to strongly incorporate sender address to recipient alias in the spam filtering, as a leaked address would be perhaps the strongest correlation to being spam. Being able to automatically determine whether a recipient address represents a mailing list to which I am subscribed would also be great. Finally, it would be great if the 'To:' column always showed my alias or mail-list address first in a list of addresses. --Jeff

#92 questions

by smkatz

Thursday June 24th, 2004 3:51 PM

Reply to this message

I use this exact alias and domain name system. It currently defaults to the last used identity--which is a silly practice. It should default the To: line to match the reply-to address.

Your idea that it would know if an address was leaked and correlate that to higher spam is intriguing, but faulty. It is possible someone on Usenet would want to reach you, and if I use amazoncs@doman for Amazon Customer Service communications.. they need to get past the filter too.

Besides, you would need a way of telling the client, either if !my default address { to==alias; }

or you would need a statically user-updated list of addresses, which is impractical.

#108 Re: Thunderbird multiple identity management

by Ashmodai <ashmodai@mushroom-cloud.com>

Friday June 25th, 2004 7:16 AM

Reply to this message

Allowing the user to fill in the From: line in new mails and replies would be useful to. I own an entire domain and have a lot of e-mail addresses which all use the same POP3 account and therefore I don't need more than one account in TB, but it won't let me use anything but my default address as From: line. Quite annoying when you can't chose between private and official addresses.

#74 Link suggest based on user behaviour

by ivanii

Tuesday June 22nd, 2004 2:09 PM

Reply to this message

It would be very interesting to mark somehow links on the page (and other tabbed pages) that are interesting for user (or to put them into sidebar).

As many users have already insisted on extended prefetch function, Firefox 2 could also do some analysis in the background and mark which links are interesting for user.

I guess this should be condidered for prefetched pages:

1. Bayesian based guessing what is interesting to user (user should previously mark interesting and uninteresting pages). 2. Page where the visitor already was, where he spent some time and which is updated 3. Page where user showed some activity (filling forms, clicking links, etc). 4. Pages that have link from pages that user marked as good or where he showed some activity. Yes, that's right, Firefox could compute something like personal version of Google's PageRank. Finally it may cooperate with Google to use this personal rankas a feature to compute personal search results!

Firefox should also consider probably to present separatly pages that user usually visits after the page he now visits (this requires tracking). As this links cannot be marked on the page, they should be presented in sidebar.

#75 Link suggest based on user behaviour

by ivanii

Tuesday June 22nd, 2004 2:10 PM

Reply to this message

It would be very interesting to mark somehow links on the page (and other tabbed pages) that are interesting for user (or to put them into sidebar).

As many users have already insisted on extended prefetch function, Firefox 2 could also do some analysis in the background and mark which links are interesting for user.

I guess this should be condidered for prefetched pages:

1. Bayesian based guessing what is interesting to user (user should previously mark interesting and uninteresting pages). 2. Page where the visitor already was, where he spent some time and which is updated 3. Page where user showed some activity (filling forms, clicking links, etc). 4. Pages that have link from pages that user marked as good or where he showed some activity. Yes, that's right, Firefox could compute something like personal version of Google's PageRank. Finally it may cooperate with Google to use this personal rankas a feature to compute personal search results!

Firefox should also consider probably to present separatly pages that user usually visits after the page he now visits (this requires tracking). As this links cannot be marked on the page, they should be presented in sidebar.

#76 Intelligent Forum Browsing

by Pied

Tuesday June 22nd, 2004 3:54 PM

Reply to this message

Hi! I think that this idea might be some great time saving for those who are used to reading forums. Often, good (or often used) forums show when there is a new message or a new thread. Therefore, i d like mozilla to be able to tab-open all new messages in the whole forum, or say, just threads from the page I am browsing. It would either use cache for forums who only give the date of the last message, either parse the code to see if the thread has a "new message" icon or a different color link. Then, it would open the last page of this thread.

Having the possibility to see _all_ new messages can be nice for little forums, and "just new messages from this page" nice for bigger ones...

#77 Record/playback actions

by lacostej <coffeebreaks@hotmail.com>

Wednesday June 23rd, 2004 6:49 AM

Reply to this message

I proposed a way to record/playback actions here: <http://bugzilla.mozilla.org/show_bug.cgi?id=242114>

It was marked as WONTFIX, but I still think it will be a good idea. Automating things like entering a web site, or making some more complex actions might have some interesting capabilities that could only be revealed once the functionality is in there.

That combined with XUL could make an interesting functionality available to many applications.

#80 Smart Cookie Manager

by clip

Wednesday June 23rd, 2004 11:49 AM

Reply to this message

I prefer to maintain control over cookies consequently I am frequently having to respond to cookie prompts, clean up stored cookies and clean up my cookie exceptions. Perhaps I suffer from cookie paranoia but I would like to see smart cookie management that did some of this for me. There are at least two aspects to this.

First learn which types of cookies I typically allow and which types I typically block. For me this is not always 'site' based. When the allow cookie prompt is displayed set the default based on learned preferences. For example a cookie with the word "ad" or the word "count" might default to block. This allows me to maintain control via the cookie prompt but a quick hit of the Enter key and I'm moving on. I might even be more inclined to disable the cookie prompt message once Firefox smart cookies caught on to my likes and dislikes. That implies when the prompt message is disabled, smart cookies would decide for me if the cookie is allowed or blocked.

Second learn how I manage stored cookies and exceptions. There are various categories of cookies that I will allow but I do not want to store indefinitely. Learn what cookies I repeatedly delete from my stored cookies. Learn what cookies I repeatedly delete from my exception list. For example while hunting down the best price on an item I might allow some sites to save cookies. But when done I will clear them all out. These are redundant tasks that could be automated. This could be done each time Firefox loads or better yet an option to schedule it; on startup, on shutdown, every 10 days, or every 30 days.

I suppose the above idea could be extended to the various pieces that make up Firefox security including cache, passwords, etc. For example Smart Cache Clear, Smart Cache Update, Smart History Clear etc.

#117 Re: Smart Cookie Manager

by mjgallow

Saturday September 11th, 2004 11:38 AM

Reply to this message

I think a Smart Cookie Manager is a great idea, but I would also like it to be easier to change the status of cookie exceptions in Mozilla Firefox. I shouldn't have to retype the name of a cookie in my exception list in order to allow it to be blocked or allowed. I should be able to click the cookie name, then click block or allow. It also should be clear that removing a cookie and blocking a cookie are not the same.

It would be nice if there were shortcut keys for the buttons in the cookies exceptions dialog window as well.

Also, sometimes it is a little difficult to know what cookie you want to allow so as to enter a site or utilize a site's functions, especially if you have already decided to block all cookies from that site. If there was a way for your browser to list the cookies of the current page being viewed, that would be really helpful, especially if it could tell which cookies originate from the web site and which do not.

#119 Re: Smart Cookie Manager

by pannabecker

Tuesday February 15th, 2005 10:33 AM

Reply to this message

There are a few really annoying web sites that drop cookies that upon a second visit to that web site prompt you to fill out user info or join their membership role. it would be very haandy to have a script based manager, or a rule based manager that would with the exception list. to do some thing like 'for wishingtonpost.com delete all (or COUNT...) cookies when exiting site...'

#82 what technis do they use?

by lima1

Wednesday June 23rd, 2004 2:38 PM

Reply to this message

I think, to find some possibilities, it is necessary to know, which technics they want to use. Support vector machines? Boring bayes networks ;)? ... please, enlighten us...

#87 Daily page visit

by andybauer <andy@bauer-oe.de>

Thursday June 24th, 2004 8:22 AM

Reply to this message

I usually visit some news pages once or twice a day after each other. It would be cool if mozilla could detect such pages, that are visited after each other regularly, and provide a "go to the next page" button. When a certain page is visited often after the tour through these pages is complete, it should be attatched to the tour and if the user doesn't stay long on a page, it should be removed. Maybe multiple tours wold be possible (e.g. job, computer news, private forums,....) The pages on the tour could be sorted, by the time spent on it.

#88 bad search engine results

by lima1

Thursday June 24th, 2004 8:32 AM

Reply to this message

if I search in google, I try the first most promising hits, open it, and close it very fast, if I can't find the informations I search. when I search for mathematical formulas for example, I get often very popular sites, that are not very detailed. the best results are the one with many formulas, often produced by latex2html, hosted on university servers. in the trainingsphase, the browser could learn, what pages I find very interesting (long opened, but active (scrolling, return to this tab, ...). After that, browser could prefetch links (maybe this could be machine learned, too) and examine them in the background, while I read the first hit, which other links are maybe more interesting and give me a suggestion.

#91 Broken bookmarks

by lijil <mikecaines@speedymail.org>

Thursday June 24th, 2004 3:38 PM

Reply to this message

When I click a bookmark only to find out the URL is no longer valid the first thing I do is track down the new location. Mozilla could detect similar URL's that are visited a short time after, or get info from a redirection page if it was provided, and then prompt to update your bookmark.

#93 Community learning

by pa_bryan

Thursday June 24th, 2004 5:21 PM

Reply to this message

My idea is for a community based learning system. It's more of a framework than anything else though. Servers are set up to store browsing habits. You configure your browser (firefox of course!) to use a particular server (e.g. community.mozilla.org). Then, as you browse, it updates the server with your browsing habits. For example, while looking at pages about apples, I also looked at pages about oranges.

If a lot of people use this system, the system will learn from a whole variety of different users with different tastes. Thats all good and well, but your then at the mercy of everyone else's browsing habits? Not so! The system could also store you habits locally, and only refer to the community site when needed. For example, one day I browse around looking for sites about bananas. I've never done this before. It just so happens that other people have already done this, and the community site has browsing habits based on bananas. Hey wow, look at that, it tells me that in Australia, a nasty fungus that effects frogs is being spread all around Australia, because infected frogs are hopping into the banana boxes that are subsequently shipped all around Australia.

Now, imagine another day, I'm browsing pages about apples. Remember, I also like to look at pages about oranges at the same time. The community site however, relates apples to a computer company instead (based on other people's browsing habits). That's okay, my local system knows to show me information about blood oranges rather than the new G5's. Maybe sites about using Macs to build computer models on the properties of citric acid would be ranked higher ;-)

Like I said, my idea is more about the framework. The actual mechanics of how the system learns could incoporate many of the suggestions already posted (e.g. Bayesian based learning). Perhaps, a side bar could be used, that lists sites relevant to what you've been looking at recently. It could update as you browse, and tune itself as you follow (or don't follow) its suggestions. Maybe it does Google searches on your behalf, and show the results in the side bar.

Privacy is a concern with my idea. I certainly wouldn't be using a server that's run by certain large companies. All habits stored would have to be anonymous. Pehaps "open" sites would be trust worthy. By open, I mean ones that let anyone see the information stored at any time, has a clear privacy policy etc. All data should be sent in clear text so you can use monitor the traffic your browser is sending. Perhaps, XML would be a good choice for the transfer of data between clients and the server.

There's also bandwidth concerns, and possibly storage. Bandwidth has been discussed already (see the pre-fetching posts for example), and storeage costs are low enough that it's probably not a problem to store the habits of millions of people - after all, it's only meta-data... how much could there be?

#94 Community learning

by pa_bryan

Thursday June 24th, 2004 6:02 PM

Reply to this message

My idea is for a community based learning system. It's more of a framework than anything else though. Servers are set up to store browsing habits. You configure your browser (firefox of course!) to use a particular server (e.g. community.mozilla.org). Then, as you browse, it updates the server with your browsing habits. For example, while looking at pages about apples, I also looked at pages about oranges.

If a lot of people use this system, the system will learn from a whole variety of different users with different tastes. Thats all good and well, but your then at the mercy of everyone else's browsing habits? Not so! The system could also store you habits locally, and only refer to the community site when needed. For example, one day I browse around looking for sites about bananas. I've never done this before. It just so happens that other people have already done this, and the community site has browsing habits based on bananas. Hey wow, look at that, it tells me that in Australia, a nasty fungus that effects frogs is being spread all around Australia, because infected frogs are hopping into the banana boxes that are subsequently shipped all around Australia.

Now, imagine another day, I'm browsing pages about apples. Remember, I also like to look at pages about oranges at the same time. The community site however, relates apples to a computer company instead (based on other people's browsing habits). That's okay, my local system knows to show me information about blood oranges rather than the new G5's. Maybe sites about using Macs to build computer models on the properties of citric acid would be ranked higher ;-)

Like I said, my idea is more about the framework. The actual mechanics of how the system learns could incoporate many of the suggestions already posted (e.g. Bayesian based learning). Perhaps, a side bar could be used, that lists sites relevant to what you've been looking at recently. It could update as you browse, and tune itself as you follow (or don't follow) its suggestions. Maybe it does Google searches on your behalf, and show the results in the side bar.

Privacy is a concern with my idea. I certainly wouldn't be using a server that's run by certain large companies. All habits stored would have to be anonymous. Pehaps "open" sites would be trust worthy. By open, I mean ones that let anyone see the information stored at any time, has a clear privacy policy etc. All data should be sent in clear text so you can use monitor the traffic your browser is sending. Perhaps, XML would be a good choice for the transfer of data between clients and the server.

There's also bandwidth concerns, and possibly storage. Bandwidth has been discussed already (see the pre-fetching posts for example), and storeage costs are low enough that it's probably not a problem to store the habits of millions of people - after all, it's only meta-data... how much could there be?

#109 Noticing repeated sequences of user actions

by martinholmes

Friday June 25th, 2004 1:17 PM

Reply to this message

I have a proposal for something that would offer to create a kind of "macro" when it notices that the user has repeated a sequence of actions several times. For instance, if (to enter a particular location on a particular site) I need to hit a URL, click on a link, click on another link, then scroll down to find something, and I do that every day (perhaps unable to bookmark the end-location because it's a dynamic URL), the program would notice after three or four times. It would show me a list of the actions I've just taken, and offer to remember them in the form of a macro, perhaps linked from a bookmark or a toolbar button.

In another example: perhaps, when developing pages, I have a habit of popping up the JavaScript console, and hitting two validation links in the Developer toolbar to get HTML validated through W3C and Bobby, I would be offered the option to save this sequence of actions and run them automatically.

Obviously, this is something similar to an MS Office macro, but instead of my recording it myself (which I may well screw up, even if I know how to do it), the program would notice these repeated sequences and offer to do the job for me.

Cheers, Maritn Holmes (University of Victoria)

#111 Re: Need help? Do it yourself

by lacostej <coffeebreaks@hotmail.com>

Tuesday June 29th, 2004 6:05 AM

Reply to this message

martin, this is like my idea of recording actions. See one post earlier in that page. There's an issue opened in bugzilla with that aim.

Jerome

#112 dashboard

by hno

Tuesday July 6th, 2004 11:57 PM

Reply to this message

Have a look at this: <http://www.nat.org/dashboard/>

Quoting: "Why can't my computer automatically show me things that will help me with what I'm doing, instead of making me search around for them?

The goal of the dashboard is to automatically show a user useful files and other objects as he goes about his day. While you read email, browse the web, write a document, or talk to your friends on IM, the dashboard does its best to proactively find objects that are relevant to your current activity, and to display them in a friendly way, saving you from digging around through your stuff like a disorganized filing clerk.

For example, if a friend IMs you and says "I can't wait for our camping trip this weekend!" the dashboard will show things like your recent emails about the camping trip, your camping bookmarks, and any files or notes you've got on your hard drive about camping."

I don't know how possible this is to do in Firefox, it might even be more appropriate if the OS had this feature. But nevertheless, I think it could be a really useful tool. Cheers.

#113 Machine Learning suggestions...

by faraz

Wednesday July 7th, 2004 8:35 PM

Reply to this message

The first obvious thing, probably mentioned earlier, is pre-fetching of additional pages in an article or any site that requires one to go through a few pages. This can be expanded such that perhaps the remaining pages can open in background tabs; this can be configured. Also, it would be nice if my browser can automatically bookmark some pages for me...if it finds that I visit a certain page often, I can have a "frequently visited" bookmark folder show up. Also, there's one thing that has peeved me in almost every browser though it's not really related to machine-learning... bookmarks in webpages with frames. Often, you would bookmark a page after you've navigated a bit through a frames-based site, and you realize later that you've actually just bookmarked the front page of the frames site, and not the actual page you want, since the URL in the address bar doesn't change. It would nice to see this fixed. However, as a general suggestion, I believe one of the best things about Firefox right now is its simplicity. If implementing these suggestions will bloat my browser, I'd prefer to not have them. In the end, the browser should be configurable as much as possible...and there are some things best left to humans.

#114 Multiple link selection

by amish

Thursday July 8th, 2004 12:32 PM

Reply to this message

Multiple link selection: This would be the ability to hold down the CTRL or SHIFT key to select multiple links in a page with the left mouse. A user can then use the right mouse button menu to select the usual options, e.g. "save link targets", "open links in new tabs", etc. Selecting "open links in new tabs" will open all the selected links/documents in adjacent tabs. This would be great and a time saver is you want to open open multiple links in a page in their own tab or window (e.g. from google's search results). This behaviour is analogous to Konqueror's file manager functionality, which let's you select multiple files in a directory with the CTRL and SHIFT keys. So rather than selecting links in Mozilla or Firefox one at a time and opening them up in a new tab I can SHIFT-leftclick on the first link and SHIFT-leftclick on the 5th consequtive link to select all the links from 1 to 5 and then open them in their own tabs, or I could CTRL-leftclick on the first link and CTRL-leftclick on the 5th consequtive link to and then with a single right link save both targets in one save step. If you wanted to open all the sites shown in a google response page some machine learning :-) may be necessary to avoid selecting the "Cached" and "Similar pages" links as well.

#115 Search database of bookmarked content

by amish

Thursday July 8th, 2004 12:34 PM

Reply to this message

This is an extension of the idea proposed by someone for searching the local browser cache. For each page bookmarked the browser should save keywords and content from bookmarked pages. Common words (e.g. is, be, can, etc.) and duplicate/repetitive words can be dropped to reduce the size of the search database. If you have a large number of bookmarks like me, you can then search the content of the bookmarked pages locally. Its possible that the site content may have changed, and then possibly not or the content could be archived on the site. Saves having to go to Google if a bookmarked site already has what you are looking for.

#116 stop from wndow change itself! (not pop-up!)

by shmulikdge

Wednesday July 14th, 2004 6:16 AM

Reply to this message

i not mean pop-up, i mean a wahole new browser page. in sme site, i open 3 windows for ex, and i wathing one, and other page come to my face with no warning!! i want to chooze waht i see!

and 2 thing abot the favorite. when i delete something, firefox need to ask me IF I AM SURE I WANT DELETE! fter i press 'yes'... and i want to drag a folder to orginze me facorite. and drag from window to the favorite in the broswer like in IE..