Hacker News new | past | comments | ask | show | jobs | submit login
Open-Sourcing PEPS: A modern webmail server (hbbio.tumblr.com)
66 points by hbbio on Nov 7, 2013 | hide | past | web | favorite | 26 comments



Why is nobody working on a better rich client for email?

I use Thunderbird, which is the most full-featured of the three or so desktop email clients for Linux. Thunderbird aggregates my many email addresses in one place and allows me to manage my email mostly in the way I would like.

My biggest problem with Thunderbird is simply that it is not a joy to use. As a tool, it accomplishes the job. Maybe that's all I can expect.

It seems like many people are releasing very cool webmail solutions. But webmail is kind of gross. I have yet to find one that does a great job aggregating several accounts in a seamless way. It requires my webserver to be able to see the folder where I store my mail, which is a big concern (or alternately it requires me to run an intranet-only webserver for the explicit purpose of providing webmail, at which point why not just run a desktop client?). And I am extremely skeptical of any end-to-end encryption strategy using webmail...I understand that an HTTPS connection to the server + secure encryption on the server before transmitting should be sufficient, but it doesn't feel right. I'll be over here in my tinfoil hat if you need me.

Writing a desktop email client for Linux is not the way to get massive loads of HN karma. It's kind of unsexy, and since in the Year of our Lord 2013 it is still unnecessarily difficult to write something that looks nice and runs on multiple platforms, only a few die-hard throwbacks will be grateful (and half of them are running mutt or pine anyway and will sneer at your sissy X windowing system).

OK, so I probably answered the question I opened this comment with, but it was rhetorical anyway.


Couldnt agree more, thunderbird is ok, but its fairly slow and hasnt gotten much improvements usability-wise since netscape mail, its basically the same.

Speedwise its not up to par either I think, searching is quite bad, but this might be issues really with imap rather than anything else. My last efforts have been beefing up our mailserver and also adding full text indexing for searching which helped a bit but there is still a long way to go to get it really instant.


I agree, perhaps I'm missing something but I don't understand why a decent desktop mail program that's a breeze to use seems to be beyond all modern engineering. Thunderbird is "fine" - but I cringe a bit every time I open it up. There's really nothing going on with mail that's so complicated that I have to dedicate 20% CPU time and 250MB of RAM to it. And simple things like just searching never seem to work quite right.

It gets the job done, but I've seen enough great software that I feel like it should be so much better.

But the truth is, email clients have always kind of sucked like this.

What are the alternatives?


I don't know what OS you're running, but on Linux the biggest contenders seem to be KMail (for KDE) and Claws-mail (formerly Sylpheed). I've used the latter before, and it seems to primarily be a lighter-weight Thunderbird clone. I haven't used KMail.

mutt and pine (or alpine) running in a terminal are fairly cromulent alternatives to those who don't mind navigating with the keyboard and working modally, but that's not really how many of us prefer to work with our email.

(I have over ten years of email archives, and in order for me to switch, the new system will need to be able to import and search these archives.)


I've thought it would be nice to have a public server/service that you can pay a nominal fee for (like dropbox) that can also run a handful of web based applications against your data.. Multi-client webmail, and iGoogle, Google-Reader replacements spring formost in my mind... It would be great to have an application platform that's hosted, where the user owns their apps and their data.. and can simply take their ball and host on Amazon, Joyent, or wherever if they like.

That's been in my head for years now.. just not sure how to have something accessible, with a free tier and make money.


mutt user here, in an xterm actually but that's another story...

I'd kill for a decent cross platform mail client. They're all shit. I've literally tried them all. Always end up back on mutt + procmail.

Scarily enough, a version of outlook express that bothers to respect standards, has a decent on disk format (maildir please) and a decent editor would hit the mark.


I miss the Thunderbird of old that looked like a version of Outlook Express. That was my favorite version and while I still use Thunderbird (and, these days, Outlook), I wish I had the old format back. Maybe it's nostalgia or well-worn habits but modern mail clients just don't "feel" right.


What's funny, is the windows Live Mail client (now a separate download) seems to be closer to the old 2.x Thunderbird than the current tbird is.


Definitely. That is actually outlook express's current incarnation and it's pretty good for the money ($0). Also it does newsgroups still.


You present a number of licence choices but I think what you really need to consider is who you believe your potential competitors are.

Are they (small) consultants / consultanties who are likely to install your system for other to use and either provide feedback for environments you don't / won't have access to; or provide patches.

Or are they companies like Google / IBM / Oracle. If the later you should always choose a copyleft licence like AGPL.

Why?

Well all those companies, when they have the chance to release Free Software pick BSD / MIT style licences; because all of them have the engineering resources to fork something and take it in-house and improve on it AND continue to pay the 'stupid tax'.

But if look at something like the Linux kernel, which is GPL'd, all of them have had to contribute back.

You should pick the licence that makes things easiest for you.

Each comes with its own set of communities; I know people who will not contribute to GPL projects, whereas others never contribute to BSD / MIT projects.

Plenty of organisations make money from GPL software (e.g. Redhat) and plenty do from MIT (but I can't think of any atm - late night, wine, etc.)

Summary: Pick what is right for you, and ignore nay-sayers. You will have customers - if the product is good - no matter the licence.


Yesterday's thread about PEPS: https://news.ycombinator.com/item?id=6682226


Seriously, start collecting emails for the release of this. I'm pretty sure I'm not the only one who really wants this.


That screenshot looks an awful lot like Inky (http://inky.com/).



Or twitter for OS X, now that I think of it.

A veritable design plague.


The theme actually reminds me of the AdminFlare[1] theme on WrapBootstrap.

1: https://wrapbootstrap.com/theme/adminflare-WB0246423


Awaiting the project release:

https://github.com/MLstate/PEPS


Just played around with the demo. I love it.


Looks like it's currently slightly overwhelmed.


Indeed. I'd love to have a look, but the link got nuked.


Is this on github?


Being that they have not chose a license to release the code under, it should not be on Github. Github repositories without a license are generally covered by copyright. [1]

[1]: https://help.github.com/articles/open-source-licensing#what-...


To clarify, the default copyright rules apply (which means, more or less, no copying allowed).

Copyright very much applies to MIT/BSD/GPL licensed code as well. In fact, the mechanism of the GPL requiring changed+released versions to include an offer for the (modified) source code depends on copyright.


Not yet. What we will do probably until the end of the year is choose a license, clean a bit the tree and commit.

Edit: Technically speaking, it IS already on github, just not on a public repository :)


cool, thanks!

My entire workflow for keeping track of projects I want to follow centers around GitHub, so I can't wait :P


I have a side project related to this, would you be interested in talking about your workflow? You can email me at hello on umamicollective.com




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: