Mozilla ditches Messaging wing, eyes social webby sweetshop Mozilla has folded its Messaging subsidiary into the open source outfit's Lab division as it slowly backs away from the development of client-based email software. The move, announced yesterday, means that Mozilla's Thunderbird email software failed to survive and thrive as an independent org as did its far more popular …
"The move, announced yesterday, means that Mozilla's Thunderbird email software failed to survive and thrive as an independent org as did its far more popular sibling, the Firefox browser."
English is complicated.
"The move, announced yesterday, means that Mozilla's Thunderbird email software failed to survive and thrive as an independent org unlike its far more popular sibling, the Firefox browser."
The two sentences are equivalent... though perhaps neither is in Simple English.
I disagree with your interpretation.
John failed to do_X as did Paul: means John failed, Paul failed too.
cf. John drank three pints as did Paul: John drank three pints, Paul drank three pints too.
(there should be a comma before 'as did Paul')
John failed to do_X as Paul did: John did not do the same as Paul. (but Paul may have failed to do_X in a different way).
Thunderbird is an excellent email client. It's full featured, easy to use, and it does what it says. With one problem. Spam. I am not exagerrating when I say I get 50 spams for every legit piece of mail. Thunderbird's bayesian filter is utterly useless these days, 3rd party spam blockers are pretty crap and even broad filters based on usual terms like Lagos / Mbacha / Viagra / Rolex etc. only sweep up 50% of the crap.
It's simply too much bother and it's kind of sad. So instead I use Gmail to filter the crap out for me and it does a pretty good job too.
I solved my spam problem (maybe 300 spam emails every day compared to 50 legit) by adding my pop3 email into Gmail / Google Mail and then using thunderbird with imap to access Gmail (you can of course use pop3 instead). I had previously used thunderbird and mailwasher but Gmail spam filtering seems to be excellent rarely missing anything and rarely giving false positives.
I use Thunderbird at the moment and I am not surprised that it is failing to "survive and thrive" as I find each update makes it increasingly irritating to use. In particular the IMAP accounts now seem to be constantly out of sync.
When Opera Mail fixes their "multiple accounts on the same server" bug, I will probably switch.
Just checked their site. Their "features" are exactly the stuff available in TB3. They still don't have a built-in calendar and also require Lightning. And others are straight out of TB's feature list: "Ultra Fast Searches" ... uhhh ... that's global indexed search isn't it? "Tabbed email browsing" ... how's that different from TB's "Tabbed email browsing"? etc.
There seems to be some where they've simply installed an addon from TB by default: e.g. The signatures thingy looks soo close to the Signature Switch addon for TB as to have the exact same image in its button!
Now if Postbox is a "version" of TB which has all the necessary addons pre-installed I'd say yes! It's better. But I've seen similar which goes a whole lot further than PB. E.g. just look at Zimbra Desktop: http://www.zimbra.com/products/desktop.html
BUT! I don't like any one of the 3 ... since they still use that 80's (or rather make it 60's) file system: A flat text file to store all mail. This makes for huge performance issues due to things like fragmentation. Not to mention the 2GB limit per folder. It's only "slightly" better than Outlook's corruption-prone "Database" file with a PST extension, at least the 2GB limit is on folders not the "WHOLE" account. The TB mbox files only have disc fragmentation, but OL's "db" files have internal fragmentation on top of that!
I'd like to see some client which uses single EML files direct. That would enable backups to run smoothly, not to mention off-line syching of emails would be a breeze to something like a Smart Phone. Same should apply about contacts - that mbox is just being silly! And fragmentation would become a non-issue since the OS would handle it already (especially if you're on a Linux).
So actually I'd say to Mozilla, scrap the damned thing and start over. You don't want this Albatross no-one seems to want to fix. The ONLY thing I can say that's good about it is: "It's better than Outlook, only just!"
I agree, but I have a different suggestion. Use the maildir format, created by qmail, and used nowadays by other server mail software, including Dovecot. Again, each message is a single file, with one index file per folder, and flag files - but the flexibility and stability is supposed to be legendary. I believe I have seen a thread somewhere about one of the developers working to implement maildir for the backend storage of Thunderbird.
And I hope the move mentioned in the article will finally kick the Thunderbird and Lightning development into touch - as it's been dragging its feet for a good while. A proper calendar sharing and storage protocol and standard and a calendar server capable of dealing efficiently with large calendar sets (about 10000 appointments would do) would be a true Exchange killer and a really useful development. If someone would start with IMAP as a conceptual idea, and develop something similar, but for calendaring, it would be absolutely great.
I hope they don't dump Thunderbird, I like it... I have multiple accounts set up, including IMAP and the spam filter works great. I guess those that have spam filter problems didn't RTFM... You've got to mark good ones as "not spam" as well as bad ones as "spam", then it's spot on, at least it is for the junk that I get.
the built in thunderbird spam filters work to a certain extent but I had to add mailwasher pro as i was getting so much spam that thunderbird missed but eventually found running my email through gmail and then back to thunderbird is the most effective spam filter.
I'm asking for them to start from scratch, since they seem to refuse to fix this dead-war-horse from the past. By all means, if Mozilla would put a similar emphasis on TB as they are onto FF, then there may just be a future for TB. But this article shows they're going to remove programming resources from it. So it's even more unlikely to get fixed!
If TB can at least incorporate a decent LDAP editor then the contacts sharing would be reasonable, but LDAP's not as robust as CardDav. And for LDAP you need a server back-end whereas in CardDav all you need is a shared folder on one of the PC's on your LAN.
As for spam, it does seem to get better over time as you keep marking messages as either Not-Junk or Junk. But I'm in agreement with using GMail's filters to augment TB's - that's caught nearly everything off the cuff. If you can't use GMail, then try some other stuff such as MailWasher - or even some AV's incorporate this like AVG / Avast! paid versions.
Yep the MailDir is exactly what I was referring to. The EML file is a straight-forward text file containing the (single) RAW email message, then there is an Index file to make stuff like searches quicker. Both concepts are incredibly "simple" programming-wise, a lot more simple than re-inventing a full DBMS system, and a lot less problematic than incorporating a half-@rsed DBMS like SQLite which is used for TB settings.
As for shared calendars and contacts. It shouldn't be too difficult to simply use the current "standard" WebDav method.
http://en.wikipedia.org/wiki/WebDAV
It already has specific branches CalDav, CardDav, as well as GroupDav. At present only Lightning has any way of working with CalDav "out the box". And you need some add-ons to work with Card- or GroupDav for contacts. They should use these already robust systems internally as KDE Kontact & Evolution does ... it's just a file for Pete-sakes, there need not even be a server back-end - though that would become necessary when sharing with numerous clients.
And the -Dav file idea could also still be similar to the EML idea in that a single event / contact could be stored per file. Again making sharing even simpler (just a normal network share would do) and removing the fragmentation problem.
Not to mention, all this would make that stupid & SLOOOOOWWWWW "Compact folders" unnecessary.
BTW, could you post a link to the thread about the MailDir format for TB? I'd like to see the progress on that.
Here is the bug report - it seems there is some work in progress, just slow:
https://bugzilla.mozilla.org/show_bug.cgi?id=58308
As to the suggestions involving CalDav and Webdav - I've tried several server implementations with Lightning as client and they were pretty awful. It's one thing being able to just about get something going, and another thing having a robust solution capable of working with 5000-10000 appointments in a single calendar, share them to several computers, not corrupt data and not hold proceedings forever while it updates. Once you have even several hundred appointments in Lightning with CalDav, every update freezes the interface for 10-20 seconds at a time. Completely unacceptable. Even starting Thunderbird + Lightning when there is a CalDav calendar connected takes ages. Anything that does syncs every now-and-again is not suitable. A fine grained update mechanism, similar to IMAP, able to keep lists updated in the background is ideal.
As to just having a shared iCal (or whatever other type) file on a network share - that's for playing at home. Try implementing anything like that with large data sets and several clients accessing it in a corporate network environment and you will see it's a no go. A proper calendar server, with solid locking mechanisms, with solid procedures of resolving conflicts and dealing with updates (which avoid data loss or corruption at all costs) is needed. After a client lost about 2000 appointments out of the blue, I learned my lesson. Anything else is just child's play.
Also, anything stored in and accessed as a single file is a nightmare to search from a speed point of view. Only a server solution with solid back end (and indexing, if possible) can do fast searches on large appointment sets. And searches through appointment sets are an essential requirement for a proper back-end calendaring solution - actually, for any calendar solution, back end or front end.
Thanks for the links.
As to the CalDav/CardDav being slow, that's due to the method used to communicate between TB and the server. It basically downloads the entire file (i.e. all events / contacts) to read from the server, and re-generates the entire file then uploads. That's just being silly IMO.
If they can use a similar idea to the MailDir format (i.e. a single file for each event / contact but combined with an index file) it should have a similar performance. It's only logical, since it would then only need to download / upload those files which are different between the client & server (instead of the entire thing). The index file is what makes for searches becoming faster.
Definitely true about a simple file share being only for something like a home network. I would never think of sharing such with more than 5-10 people (preferably no more than 2-3). Even then I'd watch out for corruption. But for a home network (especially if you're not a techy type) it's a quick-n-simple solution instead of trying to setup a true groupware server.
It should even be possible to use IMAP folders as calendars / address books - I know the Addressbook Sync addon has the capability of saving & restoring a VCard file to & from an IMAP folder - but again this is the entire book in one file (so it's not a feasible idea for a true backup / sharing solution).
TB is the only realistic and popular alternative to the proprietary email clients, especially, God Help Us, Outlook.
I find it reliable, extensible and flexible, it renders HTML email intelligently, and, importantly for me, I can rely on a portable version which gives me the environment I want regardless of the machine I am saddled with.
The only problem with TB is that the development appears to be lethargic, and addons, on which TB is very dependent, tend to be slow in catching up with TB version. But that is only logistic.
.deb packages The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
