3 ms·

I agree completely. I think the thing that's going to "kill google" [1] is that they tend to look at everything as a. A,gorithmic problem. This approach works

I agree completely. I think the thing that's going to "kill google" [1] is that they tend to look at everything as a. A,gorithmic problem. This approach works great for web search, and pretty well for email (gmail really beats everyone else with their thread view of email, which can only be achieved with really smart algorithms), but groupon has showed that there is still value in traditional sales, apple that there is still value in good design, and facebook that there is still value in being cool. [1] And by kill I mean potentially slightly slow profit growth. Google will keep being google for a long time.

stanleydrew 16y ago
"gmail really beats everyone else with their thread view of email, which can only be achieved with really smart algorithms..." Really smart algorithms like comparing subject lines without the Re: or Fwd: and matching to and from email addresses?
- alexgartrell 16y ago
  I would bet something along the lines of my first born that the algorithm is significantly more complicated than you make it out to be. jwz's threading algorithm from the 90's - http://www.jwz.org/doc/threading.html http://www.jwz.org/doc/threading.html But, if it's so easy, I urge you to write and open source it, so that other products might take advantage of it :)
  - BCM43 16y ago
    https://github.com/kanru/notmuch https://github.com/kanru/notmuch This has thread views. It works very well. If you look at the headers of an email you can see: In-Reply-To: and References I'm not sure that this is what it is using, but it is a possibility.
  - stanleydrew 16y ago
    Oh come on. The standard "if it's so easy why don't you just build it yourself and open-source it" is a meaningless reply. There are tons of things in the world that are pretty easy to implement if you are provided the supporting systems and environment. If I had already put in the tons of hard work to build an email client I would write the threading piece the way I suggested above. But I haven't. My point above is that the algorithm probably isn't terribly difficult for someone to implement. I don't think the fact that I will never implement an email threading algorithm precludes me from making that assessment. The algorithm may be more complicated. It may not be. I doubt that it is significantly more complicated. I certainly wouldn't bet my first born on it, but then I doubt that you actually would either.

stanleydrew 16y ago

"gmail really beats everyone else with their thread view of email, which can only be achieved with really smart algorithms..." Really smart algorithms like comparing subject lines without the Re: or Fwd: and matching to and from email addresses?

alexgartrell 16y ago
I would bet something along the lines of my first born that the algorithm is significantly more complicated than you make it out to be. jwz's threading algorithm from the 90's - http://www.jwz.org/doc/threading.html http://www.jwz.org/doc/threading.html But, if it's so easy, I urge you to write and open source it, so that other products might take advantage of it :)
- BCM43 16y ago
  https://github.com/kanru/notmuch https://github.com/kanru/notmuch This has thread views. It works very well. If you look at the headers of an email you can see: In-Reply-To: and References I'm not sure that this is what it is using, but it is a possibility.
- stanleydrew 16y ago
  Oh come on. The standard "if it's so easy why don't you just build it yourself and open-source it" is a meaningless reply. There are tons of things in the world that are pretty easy to implement if you are provided the supporting systems and environment. If I had already put in the tons of hard work to build an email client I would write the threading piece the way I suggested above. But I haven't. My point above is that the algorithm probably isn't terribly difficult for someone to implement. I don't think the fact that I will never implement an email threading algorithm precludes me from making that assessment. The algorithm may be more complicated. It may not be. I doubt that it is significantly more complicated. I certainly wouldn't bet my first born on it, but then I doubt that you actually would either.