2010年8月9日 星期一

FW: Application News: [Register] Why has Thunderbird turned into a turkey ? [Aug 8 2010]


-------------------------------------------
???: Jamz Yaneza (RD-US)
????: 2010?8?9? ?? 03:16:43
???: Newsbank
??: Application News: [Register] Why has Thunderbird turned into a turkey ? [Aug 8 2010]
????????


Why has Thunderbird turned into a turkey ?
Feathers fly over virus accusations
By Andrew Orlowski
Posted in Applications, 8th August 2010 09:24 GMT
http://www.theregister.co.uk/2010/08/08/thunderbird_3_no/
An opportunistic and little-known software utility outfit has stuck another carving knife into Thunderbird 3.
 
The Performance Protector blog reckons the once-reliable email client has the profile of a virus. Let it loose on your systems, and wave goodbye to network performance and any space CPU cycles.
 
Fair criticism ?
 
Yes and no.
 
PerfProtector highlights two factors for the stress Thunderbird 3 imposes on a PC. One is that version 3, unlike version 2, downloads the full contents of IMAP folders by default.
The other is that it then creates a full text index of the material, but does so very inefficiently <https://developer.mozilla.org/en/Thunderbird/Gloda_indexing> . Gmail seems to provide a perfect storm, as folders are downloaded several times.

Still, other systems such as Mac's Mail.app and the Vista-rized version of Outlook Express (including Live Mail) also produce a full text index, but do so considerably more efficiently in their use of CPU and RAM resources.

Something has gone seriously wrong with Thunderbird. A year ago, the only criticism you ever heard about it was that development moved too slowly. Nothing ever happened, it seemed.

Back in June we pointed out that the version 3.1 beta was noticeably faster <http://www.theregister.co.uk/2010/06/18/thunderbird_3_1_rc_review/> , it that 1GB of RAM is now recommended, with 768MB as a system minimum.

This is quite bizarre. The Mac OS X mail client is hardly the fastest in the world, or the most efficient: but with several Gigabytes of email stored and indexed, and a few dozen virtual folders, it takes up 128MB of RAM.

At least with open source, you can see what the developers are up to - or not.

One lingering culprit has been found - the database file format that dates back to 1994, Mork. A bug filed back in August 1999 <https://bugzilla.mozilla.org/show_bug.cgi?id=11050>  requested de-Morking at least part of the software's reliance on Mork, and 3.0 promised to implement it <https://wiki.mozilla.org/Thunderbird:Thunderbird3> . Alas, the message files are still in Mork format.

A while ago I wrote an old bugger's whinge <http://www.theregister.co.uk/2010/03/08/the_great_email_client_mystery/>  about the state of email clients in general. I realise this is now a minority interest.

Do I need to update it ?

I think I might. The best Windows email client, called The Bat!, won't run under Windows 7 for me - in either 32bit or 64bit. (If anyone from RitLabs would like to get in touch - please do.)

Since that article, I've discovered Claws <http://www.claws-mail.org/>  - a fork of the cross-platform Sylpheed project, which has been around for ever. (The earliest Sylpheed email I received is dated November 2000.)

Claws is available on Windows, Solaris and BSD, too - but on Linux I've found it's a much more efficient alternative to Thunderbird. Using Claws, like Pegasus, reminds you that dealing with IMAP isn't so difficult after all. Like Pegasus, it's one man effort - or two, I believe after the fork.

Back on Windows, LiveMail seems to be more reliable, but not so reliable that I trust it. It doesn't refuse mail so much as refuse to work promptly, as ordered.

Opera is my second choice mail client on all systems - I need to flag IMAP messages (with a flag - not a colour or a pretty haiku), and I need to access multiple signatures files. If you don't need to do either, and haven't tried it, you may be pleasantly surprised. It certainly sets the benchmark for text retrieval that all the others need to reach.

So competition remains pretty thin. This means restoring Thunderbird's reputation is important - we need a functional client. Maybe it's time to slow down the release schedule again, and do the major refactoring needed to repair the software for a long-term future.
Related stories
張貼者： 於

沒有留言:

張貼留言

訂閱： 張貼留言 (Atom)