Firefox/Feature Brainstorming:Other or Uncategorized: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
Line 238: Line 238:
=== Editable dropdowns ===
=== Editable dropdowns ===
* Enhanced SELECT tag to allow the user to type new values into the dropdown or accept one of the provided options.  Examples in abound in real applications but in a form, the only solution is to provide an input box and a comment like "select a value or enter a new one".
* Enhanced SELECT tag to allow the user to type new values into the dropdown or accept one of the provided options.  Examples in abound in real applications but in a form, the only solution is to provide an input box and a comment like "select a value or enter a new one".
* Enhance the SELECT more by allowing HTML in an OPTION. Will allow for images, bold/italic text and much more in an option.
* Enhance the SELECT more by allowing HTML in an OPTION. Will allow for images, bold/italic text and much more in an option. (Restriction necessary as e.g. a SELECT in an OPTION doesn't make sense).


== Linux compatibility ==
== Linux compatibility ==

Revision as of 12:09, 24 October 2006

« Firefox/Feature Brainstorming

External editing of text boxes

  • Allow an external editor (emacs, textpad, etc) to edit text in form text boxes
  • Currently supported by plug-ins, but only on some platforms; should be available everywhere
  • As more and more apps move towards web services, the ability to effectively edit large blocks of text (wikis, blogs, forums, webmail) becomes more and more critical.
  • Include ability to save/restore the text box contents.
References

Search and replace for editable areas

  • Search and replace for textareas and editable iframes

"Open page in another browser" option

  • Simple, useful, and easy to implement.

Conditional Comments

Allow users to add CSS filters for version compatibility that doesn't rely on a CSS rule being implemented. Arguably, this is the unofficial standard for dealing with Internet Explorer bugs and it would be nice to have a simple interface for doing this with any browser (just in case things change later on down the line). The syntax should be about the same, so something like <!--[if lte FF 6]> <![endif]--> would be appropriate.

Glow Text

Allows text to Glow when the glow HTML command is used, such as in IE7.

Page autorefresh

Create a new "Automatically page refresh" submenu in page's content menu (and inside "In this frame" submenu). It might contain items like "Stop autorefresh", "5 seconds", "10 seconds", "1 minute", "5 minutes", "Custom...". The page should be automatically refreshed untill "Stop autorefresh" selected or user navigated away from the page. In case of page was loaded by 'POST' command the new item should be "disabled" or user might be asked about "repost data every time".

Synesthetic Surfing

Synesthesia, or the cross-linking of senses, has been long known to be a powerful nmonic for anyone with the ability to tap it. Richard Feynman often thought of complex physics equations in colors, and these powerful associations can be used to assist in surfing the web.

I am a synesthetic of a few different flavors, one of which is letter/number to color associations. Whole words have color averages usually dominated by the first letter. The word "Web" for example, is orange with some blue hues at the end. I found a plug-in for Firefox that colored your tabs, but it was random. I would like to see user-picked colors for specific domains. For example, www.google.com would be a yellowish-orange color on the tab. I would select the specific hue I wanted and the tab would be colored so. That way, if I have a many tabs open at the same time, I will know which one is google at a glance. The ability to associate tonal sounds with the tabs as a mouse-over function or click function would also be very helpful. Those who are more inclined to audio cues would benifit from this immensely! I love designing GUIs, so don't worry, if you don't code this, I will ;) -NeuralZen

References
Commentary
It shouldn't be user picked, but extracted in some ways from the URL domain and optionally changed by a meta tag in the XHTML. This is important to make it a useful feature for both sites and communications (on other computers, the same site will have the same colour). Also it doesn't overload unneeded customization.

Propietary Anchor mode

Many pages don't have anchors so linking to Content that is somewhere on the page can be a pain. Make possible custom anchors like this would make firefox scroll 0pixel to left and 555pixel down: 

url#custom(0,555)

This make firefox scroll to the 5th H1 element on the page and highlight it decently: 

url#custom(h1/5)
References
Commentary
  • Pixels are unsafe (depending on fonts, zooming etc.)
  • This looks like reinventing the wheel. use xpath instead?
  • This implies support for XPointer

Smooth vs Raw Page Transitions

There isn't any *good* point to have the previous page "visible but unusable" until the new page starts rendering, a block a time. This is aesthetically awful and without any useful gain at the UI level.

Details & Commentary

Smooth Page Transition: when someone clicks on a link, the page will slightly fade out, indicating that it is refreshing until the new page is fully ready, when a smooth (really fast crossfade <0.5sec, maybe). On the page there should be some form of interaction that could allow to see - and browse, and use - the previous page until the new is fully loaded AND the possibility to switch to the new one, displaying it a block a time.

  • If implementing this, Firefox could adhere to some of IE page/site transition definitions from HTTP response or meta http-equiv headers.
    • Please no: this is a step back in the past using BAD transitions that were just annoying. This feature, as stated, is conceptually DIFFERENT and must NOT be assimilated with the IE page transitions.
  • The point here is to make the navigation softer and nicer for everyone (cool-factor), not only occasional user, and behaving more like an offline app. The "freeze page, load the other one block a time" is obsolete and should be improved into a new form of page transition: unobtrusive, aesthetically pleasant and without drawbacks (on slow connection, seeing the page loading a block a time is useful, don't avoid it if needed).
  • To be clear: this is COMPLETELY different from the "page transitions" concept seen before on some browsers as site behaviour. This should be an integrated feature, well balanced and without annoying drawbacks. Also, this should be default but disabled.
  • The interaction could be:
    • On a full-page location refresh (i.e. non javascript) the current page should be lightly dimmed until the new page is ready.
    • When the new page is ready, a fast crossfade should display it.
    • If the page loads slowly, clicking on Forward (a button not really used since now) will show the page loading a block a time, like now.
    • If the user clicks on the dimmed "old" page, it could interact with it until the new page is ready (and faded in). Any click event of course blocks the page load (and the Forward button will turn disabled).
References

Mouse gestures

Allow the user to use mouse gestures (like right click followed by left click for 'back' etc)...

References

Link navigation using arrow keys

  • Using shift+arrowKey, the selection "hovers" to the nearest link in that direction
  • With this feature remote controls can be configured to navigate pages
  • Same as Opera feature

Remote Control

Support for Red/Green/Yellow/Blue remote control buttons

  • The ability to configure keyboard shortcuts to select HTML elements that have the accesskey attribute in the format: accesskey="{red}" | accesskey="{green}" | accesskey="{yellow}" | accesskey="{blue}"
  • e.g. configure Alt-F1 to select the element with accesskey="{red}", Alt-F2 selects element with accesskey="{green}" etc
  • With this feature the RGYB buttons of remote controls can be configured to select links
  • This would make Firefox compatible with the Netgem iPlayer facility

Scripting

  • Implement a possibility to record short macros/scripts in order to automate task done on homepages, e.g. login and print a specific table, etc.
    • Support for regular expressions so that pages with broken URLs can be fixed, HTML can be converted to another format, etc.
References

Patching display

Let the user configure the display of their page and 'patch' them.

For exemple, change some image, text disposition or even the whole design of the page using regular expression. (Maybe witch a WYGIWYS editor-like to be more user friendly ?)

(I used a while ago an adblock proxy who do that, but it's really slow, whith something directly coded in firefox, this will be great ! No more forum witch crap animated smilets if we can change them on client side)

I think about a way that users can themselfs modify a whole page, and submit their 'patchs' to a centralized server.

When an other user navigate on the page, he can use a button on the navigation page to switch between the different patches submitted by other people and vote for them.

So for one site, we will probably had a lot of version : Ad-free, Alternative design (Sorry for my bad english, if you can reformulate my proposition more clearly, don't hesitate)

JavaScript Performance

JavaScript engine is too slow. When a script on page is become unresponsive - Firefox hangs, until window, proposing to stop that script appears. Enclosing enhancements are good, but functionality and quality are the main thing, we must to follow.

PDF Reader

Make a quick-opening pdf reader, just for text. In other case, run adobe or similar

Renomear abas / Rename borders

Possibility of rename the borders (abas). When a site in a new border confides, logicaly the border (aba) will be with the heading of the site, to put would be interesting to be able to temporarily change the heading that appears. This would be extremely useful case was with two or more open borders.

Note: (Sorry for my bad english, if you can reformulate my proposition more clearly, don't hesitate)

Incorporate Wizz RSS

To incorporate the Wizz RSS. It becomes it native of the program.

ActiveX

  • Implement full ActiveX support
    • including optional support via WINE under Linux

Web services integration

Automatically create a web services test form by inspecting a URL to a WSDL file. For comparison, ASP.NET can do this by browsing to a web services URL like foo.asmx?wsdl

Web mail integration

Following Firefox 2 support for online web feed readers, it could be extended to web mail services like Yahoo! Mail, Hotmail and Gmail at least. Users should be able to select one or more web mail providers and set a default or just keep the default stand alone email client. The idea is not to develop a complete email client but to provide minimum integration:

  • New mail notification through an alert a la "Downloads complete"
  • mailto: handling
  • File/Send link... overlay
  • More providers can be added with preferences for compose URL and new mail XPath

Email pseudo-url

Presently the mailto: pseudo url simply opens up a mail client. In the old Netscape/Mozilla days, the mail client was part of the package, so it wasn't necessarily such an issue. Perhaps Firefox could incorporate a lite smpt component which could actually send off data, perhaps as part of a form's action.

Explorer files integrated

Firefox 3.0 should integrate a file explorer of your system, while having the same characteristics as a web page: Tabs, mouse gestures, bookmarks. etc.

Explorer integration (Windows)

In Windows allow file system browsing (ie C:\windows), perhaps by calling on MSHTML APIs.

  • Preferably using default file manager -open Koqueror in the current tab when a local directory path is entered

In Windows, allow Firefox to change global settings for preferred browser; not only changing preferred browser for files-types, but also to change what browser applications start when they want to access a webpage.

Link Activation

  • Allow popup windows (ones that don't currently work in tabs) to appear in tabs.
  • Add option to pre-process Javascript (and Flash?) links - enable MMB/CTRL-click to open these in new tab, by tracking the target before obeying the command
  • Improved functionality when opening embedded Windows Media Player video files.
  • When hovering over a link, show its target in status bar (e.g. "Open http://example.com/ in a new tab") as in Safari.
  • Allow the user to draw a box on the webpage that launches all links within the box bounds, as new tabs
  • User key for this could be MMB, Drag or a modifier key with LMB, Drag.
  • if a large number of links are selected, a confirmation box could ask weather this was the users intention.
  • A simple example of where this would be useful is opening a number of items in a list of links, including the results of a search engine.
  • Text highlighted that starts with http:// can be opened in a new tab by dragging the selected text upward.
  • Any link will open in a new tab if dragged in any direction rather than being clicked.
  • Boss key, hot key combo that minimizes and removes from task bar, same combination brings it back.
  • if a link links to an image an preview of that image after 1 second of hovering over that link would be nice. The picture should have some kind of [X] to close it (or clicking somewhere else) and a resizer of that preview image would be nice too.
  • if I mark a text which contains a URL that does is not clickable it would be nice to have a right click option to open that link (marked text need to be parsed. if a collection of links is in that marked text a chooser would be nice (seen in opera) + "open all links in tabs" would be nice too
  • an option to parse all links in a website for dead ends. example: software collection of a 5 month old website. with an option to check (behind the scenes) all links for working results would be nice. If a link returns a "not found" or so, FF could darken that link (or alter an tiny image in front of the link - sad skull - or so.
  • When downloading a pdf (non HTML file) with target="othertarget" in link check mime type first before opening another (empty) tab and download / process the file directly.
  • An integrated prefetching mechanism (use free bandwith to load in background all the static pages linked from the page) - similar to Fasterfox plugin.
  • After i click Shift + click of link i open a link in other window. Now Gecko focuses a loaded window after loading but it's very bad! I suggest to do not focus loaded windows after a loading page through "Shit + click of link". It puts off from browsing :(
  • Same as the one above when reading email with several links: Every click on a link brings FF to focus. Better would be keeping FF in background so that one can read the whole mail undisturbed.
  • Make an option that Ctrl+Alt+Click, or some other combination, that will open a "preview" of the like page in a "bubble" or some other form of preview.
References

Link Context Menu

  • Add a command to open selected text in a new tab/window.
  • Allow to customise the links context menu. I for example have never clicked on the "Send Link", "Bookmark This Link" on purpose and i hate it when it starts my outlook or opens up the bookmark dialog.
References

Additional Mouse Interactivity

  • Whitespace grabs within the document act as grabbing the icon in the address bar.
  • Dragging links to whitespace opens the link in new tab or window
  • Highlighting text causes it to be treated as a link in all of it's functionality, including the I'm feeling lucky Google search. If the highlighted text is a valid URL but not formated link (clickable) then open in new tab.
  • Shift-click, alt-click or whatever on any graphical object (animated gif, flash etc.) hides the object
  • Opera has a nifty feature with it's Mouse Gestures when you visit an Apache Directory Index or a Gallery site, and a few other sites. When you are on the index, you can just "go forward" in an imaginary history and it will start with the first link. Continuing going forward, will result in the second link, third link, etc... And when you've seen them all - you're back at the Index page.

Javascript Links

  • When ctrl+click or middle clicking a void() + onclick link browser opens current page in a new tab or window then performs a click on the requested link.
References

Tab Refresh

  • Allow to refresh Tabs automatically by a counter

Independent tabs

  • The main browser interface should be on a thread that MUST never stop responding (and have a bigger priority). Each tab, should have it's own thread used to fill it's content. If a web page from inside one tab stops responding (or takes too much to load - such as big pdf files or sites that use java - but not resticted to those) the tab should be coloured in a different way (and maybe some text: the web page is not responding). The user MUST be able to either stop , go back, or close the tab, even if the content is not responding. It's VERY anoying now if I open lots of tabs, and the last stops responding, I must close the browser. (true the session restoration is a good step, but it's only a workaround, not the real solution to the problem).

Quick Top of page/Bottom of page

  • Perhaps a Goto Top and Goto Bottom button in the Find bar at the bottom of the browser. This would be faster than using the scrollwheel, which many mice don't have.
  • Duplicates Ctrl-Home / Ctrl-End functionality. Would an extra icon to do this be useful?

Uploading Files

Auto-zip multiple files in upload dialog

  • Implement multiple file selections in the upload dialog that automatically generates a archive (.zip) for easy transmissions of many files.
  • After selecting multiple files a dialog shall appear that asks for the file name of the archive (to be transmitted, auto-naming already included) and confirms the archive type (.zip/.tar.gz/whatsoever)

Support drag and drop files from OS into a text box / text area

  • Dragging a file icon from an OS file browser window should drop the name of the file in the target text box. Main use would be for uploading files, but it can be convenient for any other case when a file name and path must be used. It would be a must-have for web-based Document Management Systems, like eGroupWare's DMS.
References

Dynamic Font Support

Support dynamic font such as prf, eot or svg. Considering the huge advancements made in with SVG in the Firefox browser, I really can't see any excuse for NOT including this technology.

From a designer's point of view it makes just about the most amount of sense - with IE4 onwards supporting font embedding, gifting this technology to FF would just about cover 99% of browsers - the possibilities from a typography point of view are fantastic.

References

Syntax Highlightning

  • Devloper aware syntax highlightning
    • No language includes these words &nbsp;, <th> and such. The spellchecker should know about symbols used by developers, bloggers and wikiwriters and not to highlight them as a syntax error at least.

Enhanced Form Controls

WYSIWYG Text Area

Enhanced the TEXTAREA tag to provide a WYSIWYG editor that generates HTML code. How many different Javascript, ActiveX, Applet and AJAX solutions exist for this problem? How many of them actually work? None. Let's fix this problem once and do it right. Imagine: 

<TEXTAREA NAME="foo" TYPE="wysiwyg">

The user would see a text box with WYSIWYG editing controls. When the form submits, the field would contain HTML code for the user's content. Other browsers would ignore the new flag and display a standard TEXTAREA.

EVERYONE needs this, including this Wiki.

Better: Fix Bug 13474! 7 years for a Bug are quite long!

Editable dropdowns

  • Enhanced SELECT tag to allow the user to type new values into the dropdown or accept one of the provided options. Examples in abound in real applications but in a form, the only solution is to provide an input box and a comment like "select a value or enter a new one".
  • Enhance the SELECT more by allowing HTML in an OPTION. Will allow for images, bold/italic text and much more in an option. (Restriction necessary as e.g. a SELECT in an OPTION doesn't make sense).

Linux compatibility

  • Render embedded buttons with native GTK2 widgets
  • Make keyboard shortcuts work with non-Latin keyboard layouts in GTK2 (currently, GTK1 builds work as they are supposed to, as well as GTK2 under GNOME. But wth, not all people use GNOME!)
  • Make sure GTK2 builds do not screw up the keyboard focus: focus is periodically placed to the improper widgets which effectively disables keyboard shortcuts until mouse click inside browser.
References

Mac OS X compatibility

  • Support Command-Option-F to select search toolbar (standard for Mac applications).
  • Render embedded buttons with native Aqua widget
  • Dragging favicon to bookmark bar: Expand bookmark folders for bookmark placement when hovering -- FF 1.X+ under other OS's have been able to do this.
References

Long term target

On a long term basis Mozilla has come to a cross-road. The current XUL technology can't be used on a PAD/Smart phone nor on an OLPC. Yet there's a need for such a browser/mailer and it's just a matter of time somebody starts such a project. But I'm sure if such a browser/mailer gets built, it will be ported to the ordinary desktops slowly starting to replace Mozilla. And since switching technologies isn't a task done in a few weeks this subject has to be taken care of ASAP.

IMO it's important for Mozilla to start a task force now to think about the long term future. The PDA/Smart phones are already on the market almost exclusively using IE/Outlook and the OLPC and other embedded devices is just on the brink. So the pressure of this task, to prevent Mozilla from lagging behind again, is imminent.

Evaluating the future of XUL

  • It's essential to first evaluate XUL in the light of new requirements, especially about new devices ranging from the OLPC to embedded devices.

Evaluating evaluating possible alternatives

  • Alternatives should be based on the cornerstone "runs anywhere". Possible is IMO only wxWidgets yet others like Java or .Net/Mono should be rated against XUL as well.

More vs. Less

Well it seems like we could devide the users wishes into two overall categories: While the ones can't get enough features, want more of them and through that keep pushing new technologies and conveniences for the user, the other fraction just want a quick and stable browser, supporting all the W3 standards.

I belong the other category, I want a quick and stable browser with a lot of Extensions and at least 5 distinct good looking skins. Still, I've got to admit, that without the more-and-more fraction, I still would not use the Mouse Gestures. So on the one hand, I am by all means interessted in the ideas of the other users, on the other side, I got to admit that I am _saturated_. The new features in version 2.0, I do not want them. Honestly. Yeah. That's it. So, unlike most of the users who posted here, I simply do not have any _new_ ideas thus I would prefer to vote uppon the most wanted features.

Instead, you could focus on Sunbird. And of course some of the more interessting extension could some help to improve faster. I do not want to tell anybody how to spent his or her time, but if I am asked, this is my answer.

"Core Firefox" & "More Firefox" (Packaging options)

  • Distribute a single, lightweight feature-wise version of Firefox WITH a set of Mozilla Firefox developed and branded extensions in a specific extension pack (se the matching suggestion above). With this solution you will have:
    • A Firefox "Core" with just essential features (based upon XULRunner...)
    • A Firefox "More" with many features, released WITH Firefox as Extensions, by the Mozilla Firefox development team, and updated in sync with Firefox "Core".
  • Marketing wise, it could be like now a SINGLE installer, with an optional flag to be turned off during install. Still, after the setup, you could remove the extensions and you'll be "light" again.
References
Commentary
I fully support the idea of a minimal firefox with official extensions from the devs and unofficial extensions from the communitiy. Firefox is getting too big. The first time I used Firefox (around 0.5 ?), I used it, because it was small and fast! Now it just takes too long to load. This also helps the core team to do it's core stuff, and the extensions people to work on thier extensions separatly. If they want to update a feature or remove some bug, only the new extension has to be upgraded and not the full Firefox package. Official extensions would also allow the devs to adopt community extensions, if they are deemed usefull for (nearly) everyone, and add them to the offical Firefox extensions.

Extension packs

How about a extension 'mode' pack or something? This could be a bunch of plugins grouped under a certain name like 'Standard','Ultra Fast','WebDeveloper','Kiosk' etc... and then have a simple drop down where one could select the 'mode' and only plugins for that mode would be loaded. This way we get the fast version, and the features packed versions for those who want it (and we can easily switch). Ofcourse let the user modify the mode pack if they wish, but this can be hidden somewhere so beginners don't stumble upon it by mistake and mess things up.

It really is not a 'fast mode' vs 'feature packed mode' only, as it depends very much on what features a user wants and what they are doing at the time. By creating a couple of common usage mode packs it will make it easy to customize without having to individually download and install each plugin (though one could still do this). Maybe this is what a 'profile' is all about, but then again I still like to share bookmarks no matter how I'm browsing.

I for one sometimes like the browser to be very fast, at other times I prefer all the web developer options, and at other times when browsing for personal use I'd like other plugins loaded.

This way the core can stay small but first time users still get the cool behaviours they like without having to understand this whole plugin install business.(so a standard download and a core download without the mode packs for example)

The use case I see is that first time users will want all the whiz bang toys to atleast match other browsers, and don't want to go through and figure out what extensions are and how to install, they just want to click go and off they go. Experienced users on the other hand may want other features or a slimmed down feature set. I think this approach solves both issues.

I love the idea of official extensions! And I'd definately like the ability to strip down firefox to the core. This method of development and packageing gets my vote!

Perhaps the installer could be a bit more than just Core & More. You could for example have a check box for extra features, and another for web development tools. Each check box could be expanded to get individual extensions like no-script.

And of course the defaults will be well chosen.

The more functionality you add, the more words and submenus have to be created. And both is confusing for beginners and older folks. What is the difference between words, that seem to mean the same thing at first glance ? Do I have to understand "all these new words" like "javascript", "PlugIn" ? Keep in mind, not all of the users are like the Techies who develop !

Firefox already has all the necessary stuff. There must always be an easy way for "lightweight-users" to use firefox without getting confused about words, large submenus or confusing questions.

The more functions, the more words, the more confusion. Please, always offer an easy way for the non-techie user ... !

I agree that a much smaller (memory) footprint is needed (even on desktop systems), and love the extension pack idea. I'd like to also strongly suggest looking at the reason for the memory size with the current feature set - 225 MB with just 3 tabs open just seems like too much.

The subject of More vs. Less is quite interesting. A good starting point might be to ask "what recent features are being used?". In that vein, maybe one of the "pre-release" versions could be "instrumented" to count the features used and (with the permission of the user, of course) report back those. Another way might be to do a web crawl, and see what features are used by various "popular" web sites. As with ANY feature set, one should realize that "growing featureism" is a very slippery slope. I suspect that many web users would be content with the features of an older browser (say back to 1999) to do their work. Remember that each added "feature" involves code, and with any new code comes both bugs and vulnerabilities. The user community wants neither.

Please focus more on stability, robustness, and security than on adding fun features. New features are nice, and fixing bugs isn't so much fun, but I find Firefox to be more of a memory hog and less stable than IE, and that's pretty sad. Stop competing against other browsers because you're afraid of losing market-share. The only way Firefox will ever be truly great if it's designed to be great from the ground up. Focus on changes that are less glitzy and more on changes that are fundamental to making a good browser platform. When something needs to be re-architected, re-architect it. Don't assume that bugs will ultimately get fixed; so far, Firefox has added features faster than bug fixes. Rather, focus on design principles that minimize the chances that problems will happen and minimize the consequences when they DO happen (and they will!).

I agree with the above. Firefox also has a reputation for being fast and relatively slim and there already is a plugin system for those who want a feature laden browser, so the focus should be on essential improvements that will have an impact on usability for the majority of users. A lot of the proposals here go way beyond what a dedicated browser should be able to do (more or less why Firefox was created from the Mozilla suite in the first place, no?) and would just add a lot of bloat.

I concur with the above two requests. Fix all the memory leaks! Firefox leaks memory like a sieve. I hate having to restart it periodically in order to get it all reclaimed. A web browser shouldn't be using nearly 150 MB of ram. That's 15 MB for each of my open tabs!

(removed further repetitive commentary here)

Customizable extension sets

The ability for each user to create, save, install, and manage "feature (extension) sets".

Rich Internet Application Support

There is an opportunity here to lead the way in supporting the growing number of DHTML- and Ajax-heavy applications and to provide the browser support needed to eliminate some of the drawbacks of these apps. In all cases a security model will be necessary, probably including some automated safeguards and some user confirmations.

Preloaded javascript libraries

  • with AJAX large javascript libraries (e.g. "dojo.js") have to be downloaded which need a lot of time. It would be nice if such libraries could be downloaded local in the background and taken from this local store instead of downloaded from the web when referenced by a page. This would speed up any AJAX page considerably. Which libraries are handled this way may be configured in the "offline" settings.

Local filesystem access

  • Allow pages to request access to the local filesystem or some subset thereof (for example, a user's home directory, desktop, or docs folder) for reading and/or writing.

Here here! This would enable productivity applications on the web far beyond what we can do today. Mozilla has an opportunity to create a standard API for local storage.

Access to global UI elements (menubar, taskbar, panels...)

  • Allow a page to request access to major OS-level global UI elements like a first-class application. Such elements include the Mac menubar, Windows taskbar, GNOME/KDE panels, Mac Dock, etc. Some are more important than others; for example, being able to write menus to the global menubar is far more important on the Mac platform than Dock access. However, Dock access on the Mac is probably more important than taskbar access on Windows because Windows will already have an entry for each browser window (though being able to define that window's icon would be nice so users can distinguish between "regular" browser windows and in-browser apps).

Enhanced clipboard integration

  • Provide APIs to allow non-text apps to manipulate the system clipboard/pasteboard more.

More robust file uploads

  • Provide single- and multi-select access to standard OS Open dialogs for opening/uploading files. Include an upload manager for asynchronous uploads. Include functionality to resume interrupted uploads. For apps that store data, combine with local filesystem access to allow easy setup of sites that keep local and remote data seamlessly synchronized.

An alternative view of the web browser

  • Firefox as a new kind of desktop manager.
  • Active front end.
References

Enabeling new Top-Level-Domains (TLDs) on browser basis

  • Top-Level-Domains for regions, cities and others
  • Top-Level-Domains with IDN characters

The ICANN process for getting approval for a new TLD is quite slow and very stressful for TLD applicant. Making new and by ICANN not yet approved TLDs available could help to foster the development of the DNS. Firefox could easily enable lots of new TLDs and thereby give regional, cultural, linguistic or ethnic communities the opportunity to present themselves by an own identifier. Not to split up the Internet there can be only domains at those TLDs available for which the community indicates to apply at ICANN for an official approval of their TLD.

References

Annotations

I think it would be really great if there would be enhanced possibilities of graphical marks. It's tough to explain. I think of an complicated scientific text. If I print such a text. I highlight some words, cross other chapters out, add some notes. That's the big advantage of printing a page. If I could mark within a browser wouldn't have to print the page. I could work on screen.

Decentralized/distributed browsing

In case a website has been slashdotted, Firefox should still be able to reach to its data by using the offline website data which is stored locally at other Firefox users, who were able to connect to it, their disk. Firefox should act as a kind of a torrent client. Of course, there might be some security issues. Https websites should be denied and the user itself should be able to disable this "torrent functionality".

Offline browsing

I don't know where this would go, but how about a really easy off-line browsing agent? Here's an example of what I do. Every morning I pull up about a dozen web pages in tabs, as a sort of 'newspaper' I read specific articles that catch my eye, or have an interest, like say the North Korea problem. I then go to the next article. (just like a newspaper.) I'd like to be able to click a 'offline this link' (or something like that), where it downloads just that article, as deep as it needs to go, with all the stuff that goes with it) then save it. I can then take this on my laptop and read it at work, in the car or such if/when I don't have internet connections. The off-line browsers I've used want to download the entire site(s), and are so hard to configure I give up on them.

Thunderbird in a tab

I have no idea if this is possible, but if Thunderbird is installed it would be cool to have the option of opening Thunderbird as a tab.

Bookmark behaviour

Two buttons for favorites that open as many tabs as there are links in a special folder: One for daily checked pages (e.g. news paper, special forum a.s.o) and one for pages you check only sometimes (e.g. homepages of former friends or employers, news of your hometown a.s.o.)

Socket connection

I think the possibility to setup and use a socket connection for client-server communications would be a great improvement for Firefox to use it in Web2.0 / intranet-webapplications.
I know it's already possible, but it requires 'signed/secure javascript' through a plugin to deploy the possibilities, but it would (and should) be a lot easier if a could just setup a javascript object, like so: 

 mySocket = new SocketClient("tcp://domain.com/application");
 mySocket.onData = processReponse();
 mySocket.connect();

If you look at the flash player for example, it is used a lot in all kinds of applications, not just gaming or chatting, but it's also penetrating the intranet-applications business.

HTTP Sniffer

Have a HTTP Sniffer, where you see what exactly is transmitted, which POST / GET parameter, cookies, loading times etc. Something like httpwatch, see link. That's why I have to start IE in regular intverals.

References
Commentary
Should this functionality really be in the browser itself, when there are things like Paros Web Proxy?
Sometimes we are required to use a different proxy (e.g. in corporate networks). Besides, when using HTTPS and gzip compression of pages, the external sniffer will not show the contents of the page. It would be better to sniff by means of the browser itself (both for outgoing requests and for incoming pages, when they are about to be encoded or have already been decoded, respectively, by browser), and to be able to manage these requests in the browser and/or extensions (allow, deny, save, redirect etc.).

Security

  • I see many attacks upcoming, which manipulate the DOM-tree inside the browser, altering transactions in banking-systems a.s.o.
  • What i wish for Firefox is, to have ONLY signed extensions!
  • I want a default block of all extensions to manipulate SSL-protected sites.

This security-feature will be a boost for the usage of Firefox, because all banks will promote FF.

Flexible table element selection and copy

For example, select and copy a column (or columns) from a html table

  • Press Ctrl/Cmd when selecting

User dictionaries

I have an existing "user.dic" dictionary, and a couple of specialist dictionaries ("avifauna.dic", "lepidoptera.dic"). They are used by several of my apps, including my e-mail client (Turnpike) and various MS Office apps. I should be able to direct FF to use the former as its dictionary to which new words are added, and to check against all of them when spell-checking.

Also, when the word I'm spell checking is at the bottom of the window, then the suggested replacements should be at the bottom of the context menu.

Various

  • Move all MPL code to GPL

Better Support for Web Apps

See my article on that topic.

Memory management

Try to minimize the memory usage of Firefox. May be we can have one feature (one button click), which enables the browser to run with minimum feature and there by reducing the memory usage.

Retrieved from "https://wiki.allizom.org/index.php?title=Firefox/Feature_Brainstorming:Other_or_Uncategorized&oldid=39077"