David Louis Edelman David Louis Edelman

Reinventing E-mail

Mozilla Firefox managed to become the web browser of choice among techies and a legitimate mainstream alternative to Microsoft Internet Explorer in just a few years. So why is the Mozilla Foundation’s Thunderbird e-mail client dying a slow, painful death?

Mozilla Thunderbird logo Recently, Mozilla began talking about spinning off or abandoning Thunderbird altogether. And just the other day, Scott McGregor and David Bienvenu, the two principle developers of Thunderbird, left the project. (Or at least it appears that way; they’re no longer working for the Foundation, but they’re staying on as volunteer “module owners,” whatever that means.) Mozilla has also brought on David Ascher of ActiveState to launch “a new mail and communications software initiative.” What exactly does that mean? Well, it’s not clear. Apparently Mozilla is skimping on paying their PR people too. (Update 10/9/07: See Al Billings’ comment below for some links to Mozilla’s explanations for what’s going on there.)

I’ve tried to use Mozilla Thunderbird before. Every time I configure a new computer, I try to go Thunderbird. The prospect of a vibrant, evolving e-mail client with a zillion plugins is just too good to ignore. But here’s the problem: it just doesn’t fuckin’ work. Every time I try to use Thunderbird for extended periods of time, it crashes on me. Repeatedly. Ungracefully. Perhaps they’ve changed things since version 1.5, but Thunderbird doesn’t recover nicely from crashes the way Firefox does. You lose messages. It’s irritating as hell.

Don’t take my word for it; I’m not the only one who’s had problems with Thunderbird. The commenters on Slashdot aren’t exactly technoidiots — most of them, anyway — but in the Slashdot discussion “Thunderbird in Crisis?” I’ve learned that Thunderbird also:

  • Permanently loses all your e-mail if a folder climbs over 2GB in size
  • Has an import function that’s “more buggy than a New York City apartment in the summer”
  • Only shows three e-mail accounts in your accounts folder, regardless of how many you actually have
  • Doesn’t have its shit together where calendar integration is concerned

Thunderbird’s not the only e-mail client that’s in transition. The once-mighty Eudora software has been discontinued by Qualcomm, and Microsoft has changed its bare-bones e-mail client from Outlook Express to Windows Mail to Windows Live Mail in the space of a year. Yahoo’s webmail has been undergoing a facelift for the past, oh, thirty years, and so far everyone except Walt Mossberg has greeted it with an overwhelming yawn.

You want to know why programmers lose their hair? Ladies and gentlemen, I give you the state of electronic mail!

It’s been approximately 35 years since Ray Tomlinson tacked together two names and an “@” symbol to create the e-mail address. The latest estimates say that there are about 171 billion e-mail messages sent per day (of which 70% are spam). And yet just look at all the things you still can’t do reliably across platforms on e-mail:

  • Confirm that your e-mail has been received
  • Indicate high or low importance on a message
  • Ensure that your message will reach its recipient
  • Authenticate that the message sender is who she says she is
  • Include basic formatting like bolding, italicizing, and underlining
  • Have a reasonable expectation that your message won’t be intercepted by someone else

E-mail standards are still all over the place. When Mac users send mail to Outlook users, often the formatting is stripped out or filled with unintelligible characters. When Outlook users send e-mails to any other client, there’re still burdened with clunky, semi-invisible attachments. Some e-mail clients prevent you from opening attachments; others block CSS styling. Some, like Gmail, strip out just about everything but the plain text.

Hell, you can’t even count on e-mail software being able to display URLs correctly. Half the time, your long URLs end up getting sliced in half when they reach the 72- or 80-character line limit — itself an idiotic relic that should be moldering away in the Incan burial chamber of technology by now. This same limit is responsible for the incomprehensible way most e-mail programs thread long conversations. Certainly you’ve noticed that once you’ve got an e-mail thread longer than three or four messages, the whole thing devolves into an unreadable mess of “>” symbols and line breaks.

Google Gmail screenshotNot only have we not advanced in the field of e-mail at all; we’ve moved backwards. Your e-mail is actually less likely to reach its intended recipient than it was ten years ago, because of overly aggressive spam filters.

As usual, Google’s on the case, and as far as I can tell, they’re the only major player pushing innovation right now. I’ve got a love/hate relationship with Gmail, their flagship mail product (see my previous blog post “Why Is Gmail So Irritating?”), but at least they’re thinking outside the envelope. Once they integrate Gmail with their Google Gears offline API, they’ll have a rather formidable e-mail offering.

Here is Doctor Edelman’s prescription for how to fix e-mail: Don’t. Let it die, and start from scratch.

I’d like to see a bunch of major software vendors get together — Google, Microsoft, Yahoo, AOL, IBM, Apple, etc. — and come up with a new electronic messaging system. I don’t care if it’s open source, but the party should at least be open to independent developers. Make every user pay a nominal fee to an international standards body to register for it, and then charge a very small fee per message (say, a penny) to discourage bulk e-mailing. Have multiple methods of validation behind CAPTCHAs so that at least we’ll slow the spammers down. Only allow a restricted subset of attachments to be sent across it — no executables allowed — and do a virus scan of everything at the source.

Is it possible to eliminate spam on this new e-mail system? Maybe it’s not possible to completely eliminate it. Yet somehow I’ve never received a spam message on my Facebook account, and every friend invitation I’ve received has come from a real, live human being. Certainly if Facebook has figured out a way to reliably authenticate the identities of tens of millions of users, it’s not an impossible task.

I’m not saying anyone has to kill good ol’ SMTP e-mail. After all, Usenet’s still around, isn’t it? We can all maintain our goofy anonymous Hotmail addresses for sending out party invitations and arranging illicit sexual liaisons. But we should have a real e-mail system where the grown-ups can safely and reliably communicate too.

And while we’re fantasizing about the impossible: why not throw a few more things into the mix?

One thing that Google seems to have recognized before anyone else is that communication is communication, no matter how you format it or what protocol it gets transmitted on. We have too many needless choices for our digital communications today. Why should I have to choose whether I want to use STMP, IM, RSS, HTML, SMS, MMS, VoIP, fax, or voice to send my wife a two-sentence message? I should be able to just click an icon on my desktop, type in my message, and click Send. Let our digital overlords determine the best route to convey the message to its recipient; I honestly don’t care. I don’t want to be bothered thinking about communication protocols any more than I feel like running IP traceroutes to determine the shortest number of hops between our machines.

So an ideal communication client would seamlessly handle every major format listed above in one interface. It would treat e-mails and IMs and blog comments and faxes the same. It would archive all your communications in a single database, so you wouldn’t have to check five different app logs to find it later.

This is essentially getting back to the vision that Tim Berners-Lee had for the World Wide Web in the first place: an easy-to-use method of universal communication. No more FTPing files down from an IP address you’ve scribbled on a piece of notebook paper. Just point at the hyperlink and click; all the work of translating the hyperlink to a specific file on a specific web server is completely transparent to the user.

There’s no reason we should have to muck around with multiple communication interfaces anymore. After all, what’s a web page except an e-mail to the world, cc everybody?

Comments RSS Feed

  1. Peter Hood on October 9, 2007 at 4:30 pm  Chain link

    Solution? Buy Turnpike.

  2. Al Billings on October 9, 2007 at 4:52 pm  Chain link

    Well, since I actually WORK at Mozilla, I can probably comment.

    You wrote:

    Mozilla has also brought on David Ascher of ActiveState to launch “a new mail and communications software initiative.” What exactly does that mean? Well, it’s not clear. Apparently Mozilla is skimping on paying their PR people too.

    What that means isn’t unclear in the slightest. There have been many postings by Mitchell, the CEO of the Mozilla Corporation and a board member of the Mozilla Foundation on this. I suggest reading her blog and that of other people as this has all been discussed quite heavily recently. (See http://weblogs.mozillazine.org/mitchell/). You could read David’s blog as well at http://ascher.ca/blog/.

    The Mozilla Foundation is a non-profit entity. It owns the Mozilla Corporation that is a normal company (in the sense of being for profit) with the foundation as its shareholder. MoCo, as we call it, puts out Firefox and Thunderbird. After much public discussion over the last few months, MoFo (the foundation) announced that it was creating a new company, codenamed “MailCo” at the moment, for Thunderbird. This spins off Thunderbird to sink or swim on its own. MoFo is seeding this company with a few million dollars and David Ascher was hired from ActiveState, which has been building an IDE using Mozilla technologies for a while, to be the CEO of MailCo.

    Scott and David Bienvenu have decided that even though they will continue to contribute to Thunderbird (it is an open source project after all), they want to leave Mozilla, presumably to form their own company. This changes nothing about the creation of MailCo or the spinning off of Thunderbird. The stuff on Slashdot is FUD.

    As to whether Thunderbird has changed a lot since 1.5, the answer is “yes.” I’ll be catty and point out that the version number increment by .5 or 1 numbers in software for major changes normally. Now that TB is going off to find its way, I would actually expect quite a bit more development on it. Most of MoCo’s development resources have been spent on Firefox in the past. That is part of why this change was seen to be necessary.

  3. Yaron on October 10, 2007 at 2:59 pm  Chain link

    Err, the link regarding facebook spam in my previous comment actually belong one part up, when responding to the quote dealing with facebook spam. (David, feel free to fix the above and drop this comment, I promise I won’t complain about censorship)

  4. David Louis Edelman on October 10, 2007 at 3:24 pm  Chain link

    Oh, they all have that. Outlook often completely breaks down when its pst file (The file containing ALL the folders together, not just an individual folder) reaches 2GB.

    Oh, don’t get me wrong. I’m not holding Outlook up as the paragon of all e-mail clients. I think it’s the best all-in-one PIM out there (does anybody use that term anymore?), but it’s far from perfect.

    That one has been discussed in the past, and is not a good idea.

    Seems like your objections are largely based on today’s current, craptaculous SMTP mail. If this theoretical new system is just as insecure and easily spoofed as the current one, then charging would never work. But I think if you put an economic incentive in to stopping the spam, that might help.

    I really really really don’t want to pass a CAPTCHA on each and every message I’m sending

    Oh, I quite agree with you there, Yaron. I just meant CAPTCHAs when you sign up for the service, plus maybe a random check once every couple weeks when you sign in.

    What? No, wait. We have that today.
    you can create hyperlinks with FTP locations. It works.

    I think you misunderstood me here. I’m saying that the World Wide Web was the solution to what people did in the old days, FTPing documents back and forth all the time. As for writing down the IP addresses… I guess domain name servers came along well before web browsers, so I was combining two innovations into one here. :-)

    That’s a matter of taste, but I, well, like it.
    Specialized interfaces almost always do better than a one-size-fits-all solution.

    Again, I don’t think we necessarily need to abandon the specialized interfaces for an all-in-one solution; it’d just be nice to have the option. We’re moving there anyway naturally, I think.

  5. Yaron on October 11, 2007 at 8:14 am  Chain link

    I think it’s the best all-in-one PIM out there (does anybody use that term anymore?), but it’s far from perfect.

    I’m pretty sure people do still say PIM, yes. At least, I can’t think of a different and more common term that replaced it. And you’re quite right in your assessment of it.

    If this theoretical new system is just as insecure and easily spoofed as the current one, then charging would never work.

    Not really. Most of the spam, right now, is not sent because SMTP servers accept everything without authentication. Some of them do, sure, but not nearly as much as in the past, and the percentage by now is small.
    The problem is compromised personal computers, “zombie” computers if you want.
    Once a user has their computer compromised by a program made by a spammer, that program could do whatever the regular email/messaging program can do, including using the messaging program’s credentials and authenticated connection.
    The ways to solve this problem are to make the computer/OS more secure, which is an entirely different matter altogether. The security of the messaging protocol, SMTP or otherwise, isn’t really the issue. And if we can make OS secure, preventing all this spyware/adware/”spamware” crap from running, that will drastically reduce spam over SMTP as well.

    But I think if you put an economic incentive in to stopping the spam, that might help.

    Of course it would help. Does help. Some of the (not good enough, obviously) anti-spam measures today exist because someone hoped to make money of them.
    But if you’re talking about increasing the costs of sending spam, I don’t really see it as a viable option. The only way is, like you suggested, to have a price per sent message. But that won’t help, since spammers will just use other people’s message-sending capabilities. And it will extremely and drastically limit the usefulness people get from the ability to freely send messages.

    You’d also get the problem of either creating a single strong world-wide monopoly to control this, not a good idea and probably not possible, or have some level of entry for competition, in which case you’ll get players who are not as good at it or who are less hard on spammers. It’s not a technology problem.

    Again, I don’t think we necessarily need to abandon the specialized interfaces for an all-in-one solution; it’d just be nice to have the option. We’re moving there anyway naturally, I think.

    Yes, that’s going to happen. And probably regardless of whether there would be another new messaging standard or not. More and more programs are trying to integrate other messaging standards than they were developed to support originally.

  6. modafinil cognitive enhancer on January 30, 2016 at 4:19 pm  Chain link

    Narcolepsy is a part of who I am, but it no longer defines
    me. Researching air ambulance companies can be overwhelming based on the amount
    of companies in operation today. Unfortunately as the industry has expanded, insurance has not.

Add a Comment

I don't censor comments; please don't make me have to start. You can use common HTML tags, such as <b>, <i>, <a>, and <blockquote>. Comments with more than one hyperlink automatically go into the moderation queue. Your information will not be rented or sold, ever.