Heroku COO on RapGenius, legendary hires, and developers’ woman problem
venturebeat.com
venturebeat.com
I live in Silicon Beach in LA. I've architected two very successful startups in .NET. I have written .NET for sites making 100's of millions like Hulu, Ticketmaster, Fandango.
I wake up in the morning thinking about software and fall asleep thinking about software. I take my craft very seriously. I am not some aberration in the .NET world, as I've worked with hundreds of guys like myself.
Sounds like you don't have a whole lot of respect for the .NET community. I've just lost a little bit more respect for Heroku.
"Why we don’t hire .NET programmers" http://blog.expensify.com/2011/03/25/ceo-friday-why-we-dont-...
You even fit what he's talking about — you're not in Silicon Valley, and you're using a different technology stack. I don't see anything wrong with being you, and I don't think Teich does either.
It is just such a turn off. You work hard and do good work and get ignored for a bunch of people who only think they're inventing new things because they've never seen anything outside of Rubython before. So you don't make mention of doing .NET to avoid getting dismissed from the start and then everyone thinks nobody does anything in .NET because they've never heard of it in the pop-tech press.
But I think I'm just about over it. I've got my own core set of code that I use. It makes me super productive. I know I can do anything I set my mind to. I don't need acceptance from the trendy crowd to get good work done.
And I wouldn't call documentation that indicates a core piece of infrastructure works one way but then functions an entirely different way as a "mis-communication". That sort of CEO-speak is simply meant to minimize the situation to make any further discussion around it look petty.
Respect, indeed.
In the end, the characterization of this matter means little. What's key is how this reflects on their reputation, and right now it's worse than it was previously.
And the reporter bought it.
I don't have a stake in the game, and I have been interested in seeing the back and forth, but this is fluff malarky.
I think he's a bit too touchy-feely and not enough make-it-worky.
No details will be forthcoming.
When RapGenius didn't get the answers they thought they deserved, they went to the court of public opinion, and put Heroku on the defensive. All Heroku can do now is damage control, and we recently saw an example of that which didn't work out very well for Elon Musk.
So they give interviews to Venture Beat (a C-level exec publication) with the COO who talks about trust and community, and the C-level execs feel a little better.
You and I are left to wonder why a random routing dispatcher is better than a deterministic routing dispatcher regardless, of single-thread, multi-thread, evented, non-evented listeners.
Sure, developer success is important, but the issue is with server performance and efficiency. I would rather they address their problems head-on. Perhaps I'd prefer that approach for every company and every problem.
Certain aspects of Heroku, and especially some of the people who work there, are excellent. However, it seems like there are chronic systemic problems with how the organization monitors/checks quality.
[0] https://twitter.com/jtaylor/status/307261409262202880
[1] http://stackoverflow.com/a/15087542/293064
edit: added a missing comma
>"Ultimately, our stances always come from a place of respect. If you can’t say anything nice, don’t say anything at all."
>"I think it’s great when someone provides feedback , positive or negative, but we need to do it in a respectful way. When these things come up — and they will always come up — just treat everyone involved with respect."
>Two years ago, he told me, the word “respect” wasn’t on that list. Today, it defines how Heroku will respond to RapGenius and the rest of its customers.
What? This guy sounds like a politician.
If you accidentally (or deliberately) hide the way a major part of your product works and it just so happens that this means that to get a certain performance from your product, your customers will have to pay a lot more than they think, do not give an interview talking about "respect" (shouldn't that be a given!?), apologise and tell us how you are going to fix it.
What they probably should have said is that at some point if you are running too much stuff on Heroku, you probably should be running your own servers tuned to your own very specific workloads. Of course, that would sort of violate the whole premise of a PAAS in the cloud that can scale to infinity, but the point remains.
If you are paying $10,000+ a month, that buys you a lot of server time elsewhere that would probably be tuned better for current workloads or maybe Heroku needs to offer more turning options.
Either way, it feels an awful lot like Heroku has punted every time they had a chance to run with the ball on this and that's pretty disappointing.
...and if you're using a blocking, single-threaded server/framework your definition of "too much stuff" is a lot less than everyone else's.