Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Helm: Open Source Real Time Discussions

indiegogo.com

Post reply on HN
Print view View original

11–20 of 21 posts

Re: Helm: Open Source Real Time Discussions

#11

Post by toastdriven »

jrussbowman wrote: post #9

Hmm. I already built something similar at chatfor.us I used tornado load balanced by haproxy for the Web servers instead of Django. Backend is mongodb and I use a small tornado driven zeromq to manage passing updates between the front ends real time. That router also does all the db writes. I built it in my spare time (and it shows) just to have a place to take twitter conversation to get around the 140 character lim…

Sounds interesting, though I wouldn't trust MongoDB with my (or anyone else's) data.

The problem is that, while building the core of this is easy-ish, the end result should be something very complete & easy to stand up. Our prototype functions (barely, mostly just due to a lack of time), but it's not polished or complete. It's not smooth, easy to use/administer & it doesn't include all the little details that go into successful software projects.

Re: Helm: Open Source Real Time Discussions

#12

Post by petercooper »

toastdriven wrote: post #10
petercooper wrote: post #7

Earlier quoted context omitted.

However, what we're trying to build is more like an open-source product, essentially an entire service made public so anyone can use it. I think the part of this that got me pondering is that typically a basic prototype or initial version would be open sourced, people would play with it, and then funding or sponsorship would come along. Projects like Redis, WordPress, TravisCI, jQuery, Node.js, Ruby, Phusion Passenge…

Perhaps I'm mistaken, but I see it more as crowd-funding (I feel dirty using that term) a useful service with the added benefit of being able to run it yourself. The goal was to spur more/better discussions (than what private chat or crappy forums currently do). We could certainly open up the code sooner, it just won't be polished enough to use from the get-go. Most of all, it's an experiment for us. Do other people…

Perhaps I'm mistaken, but I see it more as crowd-funding (I feel dirty using that term) a useful service with the added benefit of being able to run it yourself.

That's a better way for me to interpret it actually. I latched on to the "open source" aspect as being of utmost significance, but really it seems you are taking the well trod route of getting funding for your product - much like anyone who applies to YC or seeks investors - and the open source part is a nice big cherry on the top when you're done. Given that, your approach makes a lot more sense vs my initial "we wanna do an open source project, give us $ to start!" interpretation ;-)

Well, good luck! You have certainly picked an area of software that is crying out for better solutions.

Re: Helm: Open Source Real Time Discussions

#14

Post by toastdriven »

dchuk wrote: post #13

you'd think if you were asking for $50,000 you could at least take some time to create a simple mockup of how the software will look and work...

That's in the video (~1m0s mark). The prototype is present, but ugly/unfinished. We could put it up publicly, but it wouldn't represent the final product anyhow.

Re: Helm: Open Source Real Time Discussions

#16

Post by jvoorhis »

petercooper wrote: post #6
izak30 wrote: post #3

Earlier quoted context omitted.

For those not a part of the Django community: Toastdriven is not new, Daniel Lindsley is a staple for us. Haystack and Tastypie on their own are wonderful projects with untold hours of unpaid OSS. This is more of a full-stack project than either of those, but I have a huge amount of faith in their ability to execute, given the time required. I would much rather see them working on OSS than consulting, it's better for…

Just so it's clear, I don't question the sincerity, stature or trustworthiness of these guys. They come across well in their video. My lines of thought are about the project, not them personally. What I find confusing is why a big upfront investment is necessary merely to start the project and, perhaps more importantly, why it wouldn't be a true open source project and only open sourced at the end. Why not open sourc…

Your concerns are valid. It costs nothing but a few minutes to start an open source project. Projects of this scope, however, do benefit from some amount of early design and planning. After reading PHK's recent rant [1], and reflecting on Brooks's idea of "surgical teams" [2], I wonder if some initial work in isolation is such a bad thing.

Or maybe the lines between a startup and an OSS project are just being blurred in some weird new way.

[1] http://queue.acm.org/detail.cfm?id=2349257

[2] http://en.wikipedia.org/wiki/The_Mythical_Man-Month#The_surg...

Re: Helm: Open Source Real Time Discussions

#17

Post by petercooper »

jvoorhis wrote: post #16
petercooper wrote: post #6

Earlier quoted context omitted.

Just so it's clear, I don't question the sincerity, stature or trustworthiness of these guys. They come across well in their video. My lines of thought are about the project, not them personally. What I find confusing is why a big upfront investment is necessary merely to start the project and, perhaps more importantly, why it wouldn't be a true open source project and only open sourced at the end. Why not open sourc…

Your concerns are valid. It costs nothing but a few minutes to start an open source project. Projects of this scope, however, do benefit from some amount of early design and planning. After reading PHK's recent rant [1], and reflecting on Brooks's idea of "surgical teams" [2], I wonder if some initial work in isolation is such a bad thing. Or maybe the lines between a startup and an OSS project are just being blurred…

Or maybe the lines between a startup and an OSS project are just being blurred in some weird new way.

A response from the people running the project leads me to believe this is true. Essentially it seems they're seeking funding for their product but the open sourcing of the project is a big cherry on top (rather than it being purely an "open source project").

Re: Helm: Open Source Real Time Discussions

#18

Post by jmathai »

petercooper wrote: post #2

Interesting times! It seems to be increasingly popular that developers are trying to raise money so they can start open source projects. It's impossible to say how this trend will work long term but I suspect a lot of the open source software I love today wouldn't exist if it had relied upon raising $50k up front.. not only because of the likelihood of raising the cash but because the lesser extrinsic motivation sugg…

I Kickstarted to begin working on The OpenPhoto Project [1]. I had quit my job already so even if it didn't get funded I was going to start the project.

We successfully raised about $25k [2] about 14 months ago. It took that long before we could get any additional money since the service we were offering (similar to what Automattic does with Wordpress, wasn't generating revenue yet).

I spoke with a handful of angels and seed stage investors. Never got much traction with them. Applied at the standard incubators and got a few interviews but no offers.

Oddly, what did stick was a fellowship we received from The Shuttleworth Foundation [3]. It's been an interesting journey at what started with traditional angel and VC funding to something more in line with our vision, a fellowship from a foundation that focuses on "open".

I've got ton of blog posts about the experience so far...

[1] http://theopenphotoproject.org

[2] http://www.kickstarter.com/projects/jmathai/openphoto-a-phot...

[3] http://www.shuttleworthfoundation.org/fellows/jaisen-mathai/

Re: Helm: Open Source Real Time Discussions

#19

Post by dchuk »

toastdriven wrote: post #14
dchuk wrote: post #13

you'd think if you were asking for $50,000 you could at least take some time to create a simple mockup of how the software will look and work...

That's in the video (~1m0s mark). The prototype is present, but ugly/unfinished. We could put it up publicly, but it wouldn't represent the final product anyhow.

wow, I'm an ass, sorry. I honestly stopped the video at like 55 seconds.

Carry on

Re: Helm: Open Source Real Time Discussions

#20

Post by jrussbowman »

toastdriven wrote: post #11
jrussbowman wrote: post #9

Hmm. I already built something similar at chatfor.us I used tornado load balanced by haproxy for the Web servers instead of Django. Backend is mongodb and I use a small tornado driven zeromq to manage passing updates between the front ends real time. That router also does all the db writes. I built it in my spare time (and it shows) just to have a place to take twitter conversation to get around the 140 character lim…

Sounds interesting, though I wouldn't trust MongoDB with my (or anyone else's) data. The problem is that, while building the core of this is easy-ish, the end result should be something very complete & easy to stand up. Our prototype functions (barely, mostly just due to a lack of time), but it's not polished or complete. It's not smooth, easy to use/administer & it doesn't include all the little details that go into…

Gotcha, I guess I was thinking more along the lines of open source project than actual product. Based on the replies above I see where you're going now. Good luck.

Not sure if you're coming back to read this, but I'm curious about why django? Did you choose it because of your comfort level with it or are there advantages django offers for this kind of app? I built a few things with django a while back before I moved on to tornado for everything. Not because django was bad, tornado was just a better fit for what I've been building.

Post reply on HN