513 karma · joined July 17, 2008
- We deployed tarballs to production machines which unpacked to a single directory that included a virtualenv that had both all our code and all the modules we relied on, and anything else that the code needed to function that wasn't installed in a very basic RHEL install.
- We used the --relocatable option in virtualenv to remove all references to absolute paths, which meant we could copy the virtualenv around to various directories and machines and still have it work.
- We had a series of makefiles that would make/update a virtualenv, and which worked on both mac and linux. We would use this both in development and when deploying. For several packages, we had to hand tune this to work on both mac and linux, but for most things, it Just Worked.
- When we deployed to production, we would unpack the tarball to a directory whose name included a version number for the push. We then had a symlink that pointed at the currently running version. This meant all we had to do to rollback a push was flip the symlink to the previous deployment directory, and restart. This rollback included any modules that changed (since they were in the virtualenv). I haven't seen any other way to reliably do this.
The one thing you mention that we didn't have to deal with was mixed 32/64-bit environments. One nasty solution is to have two build machines, one 32-bit and one 64-bit, but there's probably a better way...
Virtualenv allows you to sandbox all your packages in a directory local to your development/deployment environment. You no longer need to install anything on the base system, and you can have multiple virtualenvs side-by-side. In addition, you can copy a virtualenv from your build environment to your production environment by just copying the directory.
This gives you huge wins in versioning packages, testing out different versions alongside each other, and being able to strongly validate that what you tested in your staging environment is the same thing that got deployed to your production systems.
In general, I'm a very strong believer of pushing _everything_ your production environment relies on to a single directory on your production system, rather than installing it on the base system. It's actually rather surprising more people haven't moved to this model, and that there aren't better tools to support it.
I like to think of finishing things as a demon I have to conquer. The last 20% sucks so hard, but if you do it, you'll be rewarded well, because most people suck at the last 20% too, and thus, never finishing anything either. Learning to be someone who ships makes you a rare person indeed.
I still struggle with finishing things, but I've gotten a lot better in the past year. The best advice I can give you is to "show up" every day and do the work, no matter what. Sit down x amount of time every day, butt in seat, and do the next things on your list of things that are necessary to ship. Not what's fun, not what looks interesting, not surfing the web or HN. The next thing standing in the way of shipping. If you can make that a habit, and do it without fail, you can't help but ship.
In Patrick's case, BCC solves a problem that is worth more than $30 to his customers. It turns out, making bingo cards by hand is a pain in the butt.
You asked for some other examples:
- http://audiomixology.com/ - The problem is "I want to learn to be a bartender". The perceived cost is the cost of bartending school, which is hundreds of dollars. This site has rotted since I first spotted it, but I think the price was $79. And the product was basically a set of mp3s with mnemonics to remember drink recipes, an iphone app that present those same menomics (and maybe a flash card system), and a couple ebooks. This guy was definitely making money and had written a few blog posts about his process of developing Audio Mixology on http://muselife.com though I can no longer find them...
- http://onswipe.com - these guys have gotten funding and a lot of press recently. However, I think they started as pretty much just a wordpress plugin that made your wordpress site look better on an ipad. And if that isn't what their mvp was, it still sounds like a legitimate mvp to me:)
That being said, check out some hacker spaces, specifically Noisebridge (in the city), and Hacker Dojo on the peninsula.
If you happen to be around for SHDH(shdh.org) or Hackers and Founders (http://www.hackersandfounders.com/), go. They're both really stellar events, where you'll meet lots of cool people.
Lastly, if you'd like to meet up for a cup of coffee, send me an email (it's in my profile). I'm down in Palo Alto. I've worked at Google and a few startups, and am now working on my own company. I'd be happy to share any advice that might be helpful.
That is, other than the "offical" April Fool's jokes. Googlers LOVE April Fool's day, and there's generally a solid handful of nonpublic April Fool's day jokes on campus.
But there's also just random pranks. Some I'm aware of:
- Everyone's photos on the internal people database (available on an internal website) showed up with either crazy sunglasses or mustaches. Someone, probably a vision researcher, used facial recognition to do in bulk.
- Eric's picture had a photo on the wall behind him, and the photo was very subtly swapped with a picture of Bill Gates.
- There was a long series of pranks having to do with pink plastic flamingos and the T-Rex skeleton in Mountain View that were quite funny, though I forgot the full sequence. It had to do with T-Rex being surrounded by pink flamingos, then someone changed it to flamingos attacking T-Rex, then someone else brought a giant can of "flamingo bait", and all the flamingos were lying dead around T-Rex. It unfolded over a week or two, with it changing every few days.
- I was on a team that pranked each other whenever someone went on vacation. The most memorable of which involved filling a 12'x12'x8' glass walled office from floor to ceiling with pink balloons: http://www.menalto.com/photos/miscellaneous/google/Pranks/20...
It magically appeared one April Fool's day, and has been attributed to some of the SREs in charge of keeping web search running, but credit has never been claimed. There was a rather funny "I am spartacus" thread about the vending machine that appeared on a major email list after one of the facilities or kitchen management asked who was responsible, ostensibly so they could congratulate them.
I no longer work at Google (though I used to work downstairs from the vending machine), so it's awesome to hear it's still up and running and being stocked. I believe, at least for a while, facilities/cafe staff were officially condoning it if not restocking it. Definitely one of the more clever pranks I saw around campus.
The best part though, is that they'll print however many designs you want, evenly distributed across however many cards you order, at no additional cost. This has led to me a/b testing my card designs:)
I order cards 50 at a time, and order 3 or 4 variations. Then, when people ask for my card, I tell them that I'm a/b testing my cards, pull out a couple, and let them choose the design they like the most. It's a great conversation point, a good way to talk about my love for user testing, and I'm gradually honing in on a design people seem to like (I'm on my third batch).
P.s. If you're a hacker (or someone else) that hasn't haven't ordered personal business cards, you should. I can't believe how many I ended up giving out, and how much easier it makes making business connections when I meet people. Well worth the 30 minutes and $27. I wish I'd done it a year sooner.
Two possible outcomes:
- The recruiters quickly learn you have no intention of talking about job opportunities, and stop calling.
- You develop a very well validated product that you can effectively sell to recruiters, a space that is red hot right now.
None of these three points say anything about what you bring to the table that the engineer couldn't find anywhere else (ie, domain knowledge, partnerships, distribution, etc). Once you figure that part out, you won't need the NDA, because it would be a bonehead move to try and build it without you. Not to mention, your likelihood of success will go up dramatically.
Just my 2 cents...
If all you bring to the table is an idea, than to be honest, the developer probably doesn't need you.
I wish we as a valley entrepreneur culture valued this more.
I quit my fulltime job at a startup recently, with the goal of starting my own company. I spent a bunch of time having fairly in depth conversations with potential cofounders. Really sharp people, many of which who had solid, well validated business ideas with significant investor interest.
But I couldn't fully commit. In the end, I realized it was because deep in my heart, I wanted to have a great life now, not at some unspecified (and uncertain) point way in the future. So I decided to focus on building a business that generates solid cash flow and which doesn't require a 60 hour workweek for me to maintain, once I get it up and running.
The process of going around and explaining my decision to everyone who I had been talking with awkward. If you liken the process of finding a cofounder to dating leading up to marriage, then the process of "breaking up" was akin to going around to all of the girls you're dating and telling them you're gay. "We're both attracted to different things", and, "I respect your lifestyle choice" were common phrases:) People were supportive in a "That's nice for you, I'm glad you're following your heart, but I would never choose that" sort of way.
It's a shame that the valley culture is so focused on big, home run hit businesses. We're now in an era of entrepreneurship where very solid cashflow businesses can be created with very little capital, and significantly lower risk than a big homerun hit business. Yet, as a community, we treat them like the uncle whose lifestyle everyone accepts but no one wants to talk about. A lot of the advising, incubators, and accepted best practices all center around taking funding and growing as big as possible as quickly as possible, above all else.
If any "lifestyle" entrepreneurs want to grab a cup of coffee sometime and swap advice and/or encouragement, drop me a line - my contact info is in my profile.
The key to making good money on these sites is differentiating yourself from cheap offshore talent. Figure out why you're different from the guy halfway around the world who is charging $5-10/hour, and show (not tell) the prospective client that. Some ideas:
- You're a native english speaker, and can communicate effectively.
- You understand the problem the client is trying to solve.
- You're in the same timezone, and can respond to questions quickly.
- You have good understanding of the technology, and can design an appropriate solution quickly.
You're not going to compete with these guys on price - you're going to compete with them on quality and reduced risk that the project will be done correctly and on time.
The best way to demonstrate all of this to a prospective client is to write really good bids. Ask lots of questions to start with, as well as a summary of your understanding of the problem, and how you might go about solving it. Consider it a dialog, not a 2 sentence cut and paste marketing blurb.
Lastly, be picky about the projects you choose to bid on. Because you're going to spend more time bidding on them, you can respond to fewer than the shotgun approach some offshore companies take. That's ok - choose projects that look like they care about getting it done right, and aren't necessarily just looking for the cheap shortcut.
- You're passionate about being a software engineer. This could manifest itself as side projects, or simply showing that you were passionate about a work project and went above and beyond.
- You've had broad experience, and understand many areas, not just one narrow area. This is less about specific programming languages as broader technology areas. For instance, have you worked on performance tuning? Can you explain to me how a network works in excruciating detail?
- You have a formal CS education. Some may not care about this, but it's a good indication that you have solid fundamentals, and understand things like big O notation. If you're a new grad, getting good grades is important too.
- Have you worked at another startup? If so, I want to know about it - it's an indicator you understand what working at a startup entails.
I'm also a big fan of Tim Ferriss's idea of limiting the amount of information you consume. The more HN I read, the more talk shows I listen to the radio, the more twitter I read through, the stuff that's competing for attention in my head. The big realization for me was that the cost of doing these things wasn't just the time I spend doing them, but also the brainshare they take up, once I've moved back to doing something productive.
Read that puppy cover to cover, and try your hand at implementing some of the data structures and exercises in it. I studied it in college, and reviewed out of it before I interviewed at Google. Some of the best interview prep I did.
One more tip: when you're practicing, practice on a whiteboard, with no reference material, and only try compiling it on a computer after you're absolutely positive it's correct and have tried a number of test cases. Nothing like simulating real interview situations to help you prepare...
I think the dating metaphor that's been mentioned is apt though. I'd think that even if you hadn't started with the "pregnant" metaphor. Getting someone excited about an idea is always a gradual process, not just when trying to find a cofounder. It's the same when you're trying to sell a product, too.
You need to gradually woo someone. Present the rough idea initially, but then allow them to start taking ownership of the idea themselves by brainstorming with you, and start helping you to shape it. Cofounders (and most early employees) need to feel like they're working on "their" idea, not someone else's. And chances are, your idea still needs a ton of refinement yet, so you actually really need the help.
BTW, my father _still_ can't pronounce Ooyala properly, after 2.5 years, nor can a lot of people I run into in the valley.
Based on that experience, here's a couple rules I go by when naming products/companies I'm involved in:
- It has to be easy to pronounce. - It has to have a completely unambiguous spelling. - An average person should be able to spell it based on hearing it once.
You shouldn't have to say that you work for "Ooyala, spelled o-o-y-a-l-a". That's just embarrassing.