I actually agree for once with a lot of what you say here! :)
There's still plenty of applications where either the performance or stability is something that can only be done locally. Though sometimes, it can get a little strange, for example, I trust my money with HSBC's online banking site. But I would never trust my novel with Google Docs and Javascript, locally running Word there is the only safe bet.
On the simple matter of the gmail/twitter/thunderbird thing, I think it comes down to functionality and having a 1 to 1 relationship with the data model. With Twitter, the client provides 1 to 1 and sometimes 1 to 2 relationship using the API with the fact you can still see your history even when the site is fail whaling. It also takes up less 'space' on your desktop as its own little window without a whole browser tab with its chrome and under pinnings wrapped around it. Gmail and $mailclientofchoice on the other hand, is not a 1 to 1. Clients for mail still have folders, flags, outboxes, spam filter configuration screens and when you start it up, it has to go get your mail! What is this, 1999? Gmail has none of these things, as soon as you 'launch' it, your mail is there, there is no outbox, the message is just sent and stuff is tagged or labled, not put in a folder, infact, things can have more than one tag. Trying to map that to a folder model just doesn't work. Until someone actually redesigns local mail clients to actually match the new models established by Gmail and other online e-mail providers, people just won't bother.
Well, yeah, Thunderbird has to go get my mail. Gmail has to go get my mail, too-what do you think its loading screen is for? And both Twitter and its numerous clients have to fetch my tweets. Surprise: data in the cloud has to be downloaded.
Email clients send your email immediately; it only sits in the outbox if it failed. Thunderbird has had tags for ages, Thunderbird 3 makes them more visible, and I'd be very surprised if other popular email clients didn't have tag support.
And email clients have the same advantage that Twitter clients do: they can integrate with your desktop. I get a notification of new email whether I'm looking at my browser or not, whether my browser is even running or not.
I suppose it could also be something to do with the way the protocol works. On the web, my e-mail list doesn't take long to download and therefor I can happily leave my Inbox messy (currently 575 messages, 473 unread). If I used POP, it would want to download everything to my local box, if I wanted it or not. And if I used IMAP it would still want to download 575 headers just to display my inbox in a big list rathered than paged. Thusly the client doesn't suite my model. For notification and quick preview, I can use Trillian connecting to Google Talk to give me a quick look at my inbox status and notification of new mails.
My inbox has 8000 emails, going back four years. It's yet to be a problem.
POP is a joke. IMAP doesn't need to be paged exactly because it isn't limited like the browser. Read your most recent email and let the client do everything else in the background. Then you have it around for reading and searching even if Internet access goes out or the server is down, which has saved my ass several times. It works pretty much like git.
There's still plenty of applications where either the performance or stability is something that can only be done locally. Though sometimes, it can get a little strange, for example, I trust my money with HSBC's online banking site. But I would never trust my novel with Google Docs and Javascript, locally running Word there is the only safe bet.
On the simple matter of the gmail/twitter/thunderbird thing, I think it comes down to functionality and having a 1 to 1 relationship with the data model.
With Twitter, the client provides 1 to 1 and sometimes 1 to 2 relationship using the API with the fact you can still see your history even when the site is fail whaling. It also takes up less 'space' on your desktop as its own little window without a whole browser tab with its chrome and under pinnings wrapped around it.
Gmail and $mailclientofchoice on the other hand, is not a 1 to 1. Clients for mail still have folders, flags, outboxes, spam filter configuration screens and when you start it up, it has to go get your mail! What is this, 1999? Gmail has none of these things, as soon as you 'launch' it, your mail is there, there is no outbox, the message is just sent and stuff is tagged or labled, not put in a folder, infact, things can have more than one tag. Trying to map that to a folder model just doesn't work.
Until someone actually redesigns local mail clients to actually match the new models established by Gmail and other online e-mail providers, people just won't bother.
Reply
Email clients send your email immediately; it only sits in the outbox if it failed. Thunderbird has had tags for ages, Thunderbird 3 makes them more visible, and I'd be very surprised if other popular email clients didn't have tag support.
And email clients have the same advantage that Twitter clients do: they can integrate with your desktop. I get a notification of new email whether I'm looking at my browser or not, whether my browser is even running or not.
Reply
If I used POP, it would want to download everything to my local box, if I wanted it or not. And if I used IMAP it would still want to download 575 headers just to display my inbox in a big list rathered than paged. Thusly the client doesn't suite my model. For notification and quick preview, I can use Trillian connecting to Google Talk to give me a quick look at my inbox status and notification of new mails.
Reply
POP is a joke. IMAP doesn't need to be paged exactly because it isn't limited like the browser. Read your most recent email and let the client do everything else in the background. Then you have it around for reading and searching even if Internet access goes out or the server is down, which has saved my ass several times. It works pretty much like git.
Reply
Leave a comment