Lamson Project Ideas(zedshaw.com)
zedshaw.com
Lamson Project Ideas
http://zedshaw.com/blog/2009-05-20.html
6 comments
XMPP solves many of the issues with email only because it requires explicit whitelisting of message senders, and because it's still a niche ecosystem compared to email. Most users don't even try to accomplish cross-domain message sending, much less extensive file transfers or server-side storage and archiving of messages, no matter what the server capabilities might be.
As a set of protocols, SMTP and IMAP are both pretty well-suited for the task they set out to accomplish. Furthermore, the delays you mentioned in message delivery are largely due to intermediate spam filters and store-and-forward queues, not inherent issues with email as a transfer mechanism.
As a set of protocols, SMTP and IMAP are both pretty well-suited for the task they set out to accomplish. Furthermore, the delays you mentioned in message delivery are largely due to intermediate spam filters and store-and-forward queues, not inherent issues with email as a transfer mechanism.
Cross-domain messaging is easy, as long as you either accept all servers, or at least whitelist the ones you want, and have the S2S port forwarded, or better yet, no NAT at all. Same goes for file transfer, but I will admit I've never personally succeeded at a file transfer, but I couldn't get DCC to work either.
One of the big differences between this approach and email is that it would be a prompt-based system, instead of a switch applications and click the link. By prompt-based I mean that it would pop up a window asking, instead of you having to go look for it (and possibly open Outlook, Mail, or some other heavy app thats not running 24/7 like most people leave IM messengers)
One of the big differences between this approach and email is that it would be a prompt-based system, instead of a switch applications and click the link. By prompt-based I mean that it would pop up a window asking, instead of you having to go look for it (and possibly open Outlook, Mail, or some other heavy app thats not running 24/7 like most people leave IM messengers)
I don't know enough about Jabber/XMPP to decide how well it could replace email, but email is plenty fast for me. For almost any website with password reset/confirmation, by the time I click my gmail tab it's already arrived in my inbox.
This one is simple: MS Exchange dominates the corporate world, but it is used like a giant file share by the users. They attach 40MB presentations and send them to hundreds of people without even thinking. This is then actually duplicated to each user for no real reason (at least it did last time I used Exchange)
Must have been a long time ago - Exchange 2003 stopped doing that.
Must have been a long time ago - Exchange 2003 stopped doing that.
Email provides such a wonderful interface. I definitely wish that blogs w/ comments, web fora, etc. (not to mention bug trackers), used interfaces like email/nntp, instead of little web boxes. If I could make blog posts by sending an email, and then receive all the comments as emails, to be sorted into a thread by my email cleint, that would be fantastic.
I'll take the bait and plug posterous here: http://www.posterous.com
Lamson is pretty awesome, been dicking around with it as a possible mailing list solution for projects on massify, but I can think of a kajillion things to do with it.
Care sharing like zed? ;)
Nice and useful project. It shows a lot of potential.
Skimmed through the code and it seems easy enough to build a simple app, even for Python beginners like me.
or just use google groups...
What I'd like to see is mass adoption of Jabber/XMPP, since it handles IM stuff, along with tons of other features. I think it could probably totally replace e-mail, since most servers and clients support offline messaging, so you can queue up messages even when you're not online. Also, imagine how everything would change if you used IM for password resets or signup confirmations, instead of having to wait a few minutes to get an e-mail with your info, it would be nearly instant.
Another cool possibility I see with Jabber mass-adoption is authentication. You could develop (if there isn't already) a way for people to login everywhere with only their jabber id, and then the site could just IM you to ask if it was really you who attempted to sign-in. This is a lot like OpenID, except I think alot faster and probably a lot easier to implement.