Mozilla Thunderbird Gets Extension SupportWednesday June 18th, 2003New Mozilla Thunderbird builds with support for installing extensions have been released. Read Scott MacGregor's Thunderbird Builds forum posting for more details. A page of information about creating extensions has been published and Scott has already released the first official add-on, Offline Support. More extensions are available from Extension Room, which also hosts XUL-based add-ons for other Mozilla products. Why does the offline extension excite me so? Because I'd never need it, and so I don't have to choose it. Rock on! cool.. been waiting for this for my laptop. http://bugzilla.mozilla.org/show_bug.cgi?id=75866 Viewing message for very short time shouldn't mark it as read http://bugzilla.mozilla.org/show_bug.cgi?id=157372 enhancement: multiple smtp/pop3 accs per set of folders http://bugzilla.mozilla.org/show_bug.cgi?id=187578 From-field when replying to message in "local folders" should be "To"-field of original I love MailNews but those things are PITA. Thnx for listening, I'll shut up now. I understand the aim of keeping the UI light, but if everything is an extension (and most people will want some extensions if core stuff like offline mode qualifies as one) doesn't that just shove all the clutter into the extensions pref panel? <p> Instead of having a collapsible tree of prefs as Mozilla has now, doesn't the extensions area run the risk of becoming a large unordered list of miscellaneous crap? While I agree with the basic premise of your post that the Extension pref panel will likely become cluttered with extra configuration, at least it will be clutter that each individual cares about and, likely as not, understands. To you, offline mode is "core stuff", I don't even know what offline mode does! You will install this extension, I won't. I won't be encumbered with a GUI/functionality I don't need and you will be able to do whatever offline mode allows you to do! That is where the beauty of extensions lies. The product ships with the absolute minimum core set of functionality (i.e. in this case it can send and receive mail) that *everyone* needs, everything else plugs in/out, on demand, later. The thing is, offline mode is core functionality. It's there, implemented in the very heart of the mail/news DLLs whether you see it in the UI or not. All the extension does is expose that functionality to the user. And practically anyone using IMAP will use it in one capacity or another, even if its only because they have junk filtering enabled. Now in theory, perhaps making it an extension is a good idea. But now imagine that mail filters, news, spell checker, SSL, html editing, address book, ldap, junk filter, password manager are extensions too. After all, some people don't need a junk filter so why show the options, or the spell check button etc. ? But now instead of having a reasonably ordered prefs panel you now have all of this stuff dumped in extensions. Furthermore, when someone raises a bug on mail/news instead of being confronted with a single base you are face potentially dozens of mix and match configurations - "I was using Thunderbird 0.1 with Offline 0.02 and SSL 1.3 etc. when the bug occurred". So while I think extensions are a good idea, I wonder if there shouldn't be a core set of things which while notionally extensions are more tightly integrated and shipped with Thunderbird. By all means they could hide options you don't care for, but to have to visit an 'Extension Central' to get basic functionality is a recipe for disaster. I guess we have to sort out what functions to include in core 1.5. Currently, Firebird have quite some extensions that I think should be in the "core" structure. For example, Smooth Scroll and Language Menu. It is very annoying when too many commonly used functions become extensions. But of course, functions like User Agent Toolbar and Googlebar should be extensions. Off topic: I still cannot agree with the design of that search bar in Firebird. I think mozilla's design of search/location bar is a lot better. I'm with you on the search bar! I've stuck with Mozilla over Firebird this far because frankly, I like the Mozilla interface much better. If the interface changes too much between 1.4 and 1.5 (& particularly if they 'dumb down' the menu too much), I can see myself sticking with 1.4 for quite a while... On the other hand, a lot of people discovering the new search box in Firebird are really excited. Especially people switching over from IE. The bad thing is, of course, that now you have 2 fields instead of one before. The good one is that you can have a set of dozens of different search engine and have just to select which one to use. While in Mozilla 1.4, you see only your default search engine. And if you want more, you will have to open the sidebar, which makes an additional field + an additional sidebar - overall worse than the firebird solution if you ask me. I've had the same "problem" of the search field with firebird. Well that an Firebird 0.6 couldn't load up a web page after an initial shut down (Mozilla and IE worked, so it wasn't a problem with the internet connection, but that's what it appeared). Anyway it seemed like Firebird was a 0.6 release and not anywhere near where Mozilla was in release 0.9 stability-wise (tested on a few different computers). I haven't tried the fix that "sgbouwhu" suggests, but I don't feel like d/l this version of Firebird again. Also it would require a manual edit to get to go to google results (which I prefer). The most interesting thing here is that there seems to be two general choices. One, is the dedicated search bar in Firebird which seems to provide the ability to search multiple search engines, presumably because you need to look at vast amount of results. The other choice, sgbouwhu's suggestion doesn't take you to any search results at all out of the box. I don't see why the middle-ground of Moz. 1.4 wouldn't be carried over to Firebird for those users who dont want half a tool bar taken up with a dedicated search bar. Another thing this middle ground would preserve is a year or two of users typing something in the location bar and pressing down to search. I can't say how many times, I've typed stuff in the location bar, had to cut it out and put it in the other search bar. > For example, Smooth Scroll and Language Menu. Excuse me? What makes smooth scroll a core feature? > Excuse me? What makes smooth scroll a core feature? Feature parity with IE, I'd suppose. If your target audience was IE users (ala Mozilla Firebird), then it's something you should probably have. IE users get confused when everything doesn't work just like they're used to. Well, many people may not need or care about it. I care about it because when reading information/novels online, smooth scroll allows me to read while I'm scrolling down.Being able to read while scrolling speeds up searches when you know exactly what you want to find (search function doesn't really work when you're not finding certain words). Also, this saves time when you're reading (since you can read while scrolling), and it's especially true when you're reading novels with a few thousand pages or more. It's just hard to do that without smooth scroll. Further more, you sometimes may need to find where you are reading after a (larger) scroll without smooth scroll. It is a lot easier for your eyes to trace where you're reading when smooth scroll is on. More importantly, I do recognise that using smooth scroll does slower the speed that my eyes get tired. Those are some reasons I like smooth scroll, and I am never a IE user. (Well, I do have to use IE for 1 site I go to.. but not any others..) What I think is that Smooth Scroll is not really a functional feature. It is like an enhancement of "look and feel". Smooth scroll should not be large and heavy (<-- just a guess??). So why not just put it in the core? For example, let's assume OpenOffice have such "extension approach" also. We know OpenOffice's buttons are not beautiful (not that I care, it's just example). If an OpenOffice developer have checked in some more "beautiful" buttons, will they make it an extensions or just put it in core? They'll probably just put it in core. My idea is that it does not hurt to include such light enhancement in "look and feel" into core. > Smooth scroll should not be large and heavy It's not, per se. The UI for enabling and disabling it heavy as UI goes (as in, it's present). One of the goals here is to have a _simple_ pref panel. Unless you think it should be on by default and an extension required to turn it off.... in which case it better get a lot faster first. >>Currently, Firebird have quite some extensions that I think should be in the "core" structure. For example, Smooth Scroll and Language Menu.< Smoothscroll is no extension, it's still in Mozilla Firebird, but got disabled a few weeks ago because it's still buggy on some OS's. you can still enable it via about:config > general.smoothScroll, though. I think too much functionality is currently headed for the extensions tab. I think this has two disadvantages: First, users will initially see a simplistic UI which will likely not fullfill whatever needs they have. Some of them will find their way to the extensions and be happy, some won't and will install something with a better out of the box experience. Second, the quality will suffer. Testing is hard, testing with 50+ plugins is much harder. The simplistic UI thing is also affecting firebird acceptation at the moment. The default configuration is so simplistic that many users never find out how to configure it properly. Out of the box firebird is not likely to satisfy anyone's taste. You need to configure it to make it useful. A partial solution to both problems is to endorse and ship a number of extensions with thunderbird (and firebird). By endorsing I mean that they are included in regular regression testing and live in the mozilla CVS. If the offline feature breaks (important to a large number of users) because of a change in some API, this needs to be discovered during rtesting, not after a release has been put on the ftp server. It might be useful to distinguish between opt-in and opt-out extensions. The latter category should be reserved for plugins that are used by a large number of users. Users can disable it if they don't need it, prefs integrate nicely with the regular prefs. Opt-in extensions need to be located and installed by users and may or may not work properly. With such a mechanism in place it is feasible to move more functionality to extensions without affecting the out of the box experience too much. For example, the junk filtering functionality should be an opt-out extension. The only reason it's included by default is that so many people use & like it so it should be included by default. However, it is advanced functionality so novice users should not be burdened with it and advanced users should be able to remove it (e.g. because they have server side junk mail filtering). I think too much functionality is currently headed for the extensions tab. I think this has two disadvantages: First, users will initially see a simplistic UI which will likely not fullfill whatever needs they have." Where's your evidence that users needs will not be fulfilled? If by "users" you mean "most users" then I think you're wrong. If by "users" you mean a handful power or niche users then maybe so. I also don't have any substantial evidence but I think that the number of users who want or need functionality that doesn't exists (or isn't planned) in Firebird and Thunderbird stock builds is probably a small minority. I happen to be a power user and there aren't any features missing in Firebird (which exists in SeaMonkey) that I require. And how on earth is junk-mail filtering an advanced feature? The people who need it most are the novice users that have no idea how to find or set up a server-side solution. --Asa cdn's TTLO extension for Firebird is a "one stop shop" to restore panels for Firebird. Mose people will never need it, but if someone insists on having a UI, it is available. Something similar for Thunderbird would be helpful. I don't want clutter in the Options tab, but below are lines that I shouldn't need in user.js. user_pref("browser.enable_automatic_image_resizing", false); user_pref("accessibility.typeaheadfind.linksonly", false); user_pref("general.smoothScroll", false); user_pref("keyword.URL", "http://google.com/search?btnG=Google+Search&q="); user_pref("nglayout.initialpaint.delay", 0); user_pref("browser.cache.check_doc_frequency", 1); I think there are two points here that you have absolutely right. First, too much stuff appears to be implemented as 'extensions', including things that could be core features. For example, I notice that (according to the page about extensions) the feature to view HTML as plain text or as simplified HTML is an extension. This should be a core feature - it's a 'privacy'/'user control' feature, like better control over cookies, ad blocking, and popup blocking in Mozilla/Firebird, that should help differentiate Mozilla Mail from Outlook Express. If all the app's features are extensions that you have to download, then normal users (who aren't going to bother finding out about extensions) won't have those features. And there's less reason for them to switch to the app. Imagine if TB achieves a small footprint - great, it now only takes 10 minutes to download, but then you have to spend 45 going through a list of extensions choosing the ones you want. Now, from a programmatic point of view, there's no reason not to have these features as extensions - but many of them should be extensions which ship by default. Fine. That way, people who *don't* want them can *uninstall* them, rather than vice versa. That leaves the second correct point here which is that, given that extensions are used to implement basic core features, using the extensions pref panel for everything is a UI mess. Where do you go for a preference about viewing mail? Well, uh maybe to the Appearance pref panel, no, ok maybe the Mail pref panel (or whatever), uh no, ... oh okay then I just have to go through 17 extensions pref panels to see if I can figure out where the hell that feature comes from... I think somebody needs to strike a careful balance betwen extensions which are clearly differentiated from the rest of the app because they introduce entirely new features (which, fine, can have their own prefs panels under extensions), and extensions which fill in missing core features within parts of the app. These latter should generally have any prefs in the normal prefs structure. A simple guideline would be, does it need just one checkbox or a couple of radio buttons? then it should probably be categorised under application preferences. Does it need a whole panel to itself? then it should go under extension prefs. Whether this is technically possible is, of course, another matter. IMO the second point is less important and the key thing is that when it's time for Mozilla Mail to be released, somebody sits back and takes a careful look at all the extensions - those which ought to be default features and don't add hugely to the download can then be included in the default install, and their preferences incorporated within the normal preferences panel where appropriate. Everything else can then fend for itself. I just hope somebody's going to make that decision and we don't end up with the default install being a bare-bones, does-nothing mail client that has few advantages over the alternatives like Outlook Express. (And some disadvantages. *coughs* my pet bugs *coughs*. I won't plug them *again* though. :) --sam Entirely agree. It already annoys me that when installing the latest version of Mozilla I have to reinstall the latest versions of things like the Spellchecker which should be shipped with Moz (I know that this is a bug, but it has been forever!) I wouldn't have thought it would be too much hassle to have two builds available, a minimal build with just the core (mail|browser) clients, and a standard build containing a comprehensive list of the usual extensions (things like the Offline extension for example and Smooth Scrolling). There could also perhaps be a developers build which would include extensions such as Venkman, the DOM Inspector, Javascript console, Site Navigation Bar, PNH Developer Toolbar etc. It always seems strange to me that developer/beta-tester types who are far more likely to have a fast connection and lots of processing power and memory are the most likely to want their apps to be small and sparsely featured. I think it's just an excuse to play - the sparser the feature set of an app, the more reason to play at extending it to put those features back in ;-) A selection of builds would certainly be great. However, Mozilla.org may not have the resources to provide all the builds that people want themselves (and if they can't do choices, they'll do builds for developers, because they are necessary for continuing development). Other people could certainly make regular builds with other stuff though - it's already happening for Firebird (check the Firebird builds forum) - there are unofficial builds with extras included. Mozilla is really only for developers, anyway, and end users are supposed to be those of the Netscape product that is eventually based on the stable Mozilla release. So, who knows how long from now, if Netscape releases--for example, version 8.0--and it's based on our Birds, then it very well may come with what a lot of end users want. That's its target audience. Mozilla, however, is just supposed to be unsupported and primarily for developers (although, since it's used for Netscape, it's pretty good for "normal" people, too!). Just my thoughts... I see your point and it makes since to leave it up to browser producers (like Netscape) to choose which extension pieces they want to include by default in their Mozilla based browser, but still the potential user confusion of the core prefs and various extension prefs still is an issue to address. Mozilla is just going from one extreme to another. The monolithic Mozilla included everything but the kitchen sink.... now we are faced with a stripped-down Thunderbird and have to use a highly unreliable installer with no "uninstall" option. What an improvement! Why don't we have a fully-featured browser and a fully-featured mail client? The 6 people who want to spend their days hunting down extensions have the time to manually strip down their brower too. I hope you realize Thunderbird is at version 0.1 alpha...hence no installer/uninstaller. When it and Firebird becomes the official focus of Mozilla, there will be just as reliable of an installer as Mozilla has. Perhaps even better, since I have read stuff about allowing users to choose what extensions they want to automatically download + install in the web-installer version. Firebird and Thunderbird are full-featured. Most extensions provide support for hidden features that there is no GUI for. That is why extensions will be better then the current kitchen sink preferences panel in Mozilla. Besides, it isn't that hard to find extensions. Go to one of the extension pages setup by Mozilla fans, they are all listed right there in proper categories. Takes me approx 10 minutes to setup Firebird with the extensions I use when doing a clean install, instead of spending 15 minutes finding an obscure feature in Mozilla's preferences panel. Mozilla is just going from one extreme to another. The monolithic Mozilla included everything but the kitchen sink.... now we are faced with a stripped-down Thunderbird and have to use a highly unreliable installer with no "uninstall" option. What an improvement! Why don't we have a fully-featured browser and a fully-featured mail client? The 6 people who want to spend their days hunting down extensions have the time to manually strip down their brower too. I seem to recall them planning to include a bunch of the most-used extensions into Firebird by default, with the really useful ones enabled by default. So what you're suggesting is, long-term, supposed to be reality. Now, I agree with the protest that we need to have a more readable options dialog. It would be wonderful to have a way to categorise extensions so that their options would appear in predictable places in the dialog, rather than in a little oddball island list with odd names and no sorting. Someone file a bug! It's a great idea, and it shouldn't be impossible to make it work in general. -Billy Sound like the extension need to logically catigorize themselves instead of just being lumped together... I wish someone would write an extention to add a toolbar button to switch between Plain Text and HTML view of the email. That would be really useful. Same for composing letters, 1 in 10 of my e-mails is HTML... and the only way to write HTML letters is to permanently enable them :( and then each time go into the menes and set e-mail as "plain-text" "Same for composing letters, 1 in 10 of my e-mails is HTML... and the only way to write HTML letters is to permanently enable them :( and then each time go into the menes and set e-mail as 'plain-text'" Ever tried holding down shift while pressing Compose (or Reply or whatever)? :-) Alex ./thunderbird-bin: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by /home/ivo/thunderbird/libmozjs.so) ./thunderbird-bin: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by /home/ivo/thunderbird/libxpcom.so) ./thunderbird-bin: /lib/i686/libc.so.6: version `GLIBC_2.3' not found (required by /home/ivo/thunderbird/libnspr4.so) I've revived Forumzilla as a Thunderbird extension. Forumzilla is an RSS feed reader that presents news and blogs in the traditional three-pane mail interface. The new version adds a "News & Blogs" folder to your Thunderbird folder pane. Then, when you subscribe to feeds, Forumzilla checks them upon startup (and every 30 minutes thereafter) and downloads new items into sub-folders named after each blog. You can then read and manage them via the traditional Thunderbird mail interface. For more information, see my post in the Mozillazine forums: http://forums.mozillazine.org/viewtopic.php?p=101140#101140 That's great -- I think I'll be using that extension! One little request: could you allow the "News & Blogs" folder to be user-named? I know it sounds petty, but I just can't _stand_ the word "blog". I may have to read it in _some_ of the sites I go to, but I don't want it on my email client unless it _must_ be. Okay, okay. End of petty little request :-). A more interesting request: how about building the extension to imitate a real email, one that comes into your inbox, rather than into a "fake" box? That way we could take advantage of Thunderbird's own filtering and categorization rather than having to invent our own. There's some good-looking enhancements out there... See http://bugzilla.mozilla.org/show_bug.cgi?id=181866 for one example (using Bayesian filtering to classify mail into any number of folders, not just detecting spam). -Billy Don't foreget to visit Thunderbird Help http://texturizer.net/thunderbird/ Eagerly awaiting the thunderbird release. Phoenix has been treating me while and I like the default theme. I'd really like to see a thunerbitrd extension for setting up regular reminders emails to be sent to a group of people, and for event reminders - eg. I've organised an event within my organisation for 1 months time. I send an email now to tell people it's happening, then I sent one with 1 week to go, then 1 day - just to make sure they're adequetly reminded. I'd really like to see a thunerbitrd extension for setting up regular reminders emails to be sent to a group of people, and for event reminders - eg. I've organised an event within my organisation for 1 months time. I send an email now to tell people it's happening, then I sent one with 1 week to go, then 1 day - just to make sure they're adequetly reminded. |